# SceneMixer (scenemixer.com) — full text

> Full-text companion to https://scenemixer.com/llms.txt: the pricing page, the three solution pages and every guide as plain text for AI engines to quote. Prices are the same numbers as the live pages and refresh with them.

# SceneMixer Pricing 2026: Memberships & Credit Packs

URL: https://scenemixer.com/pricing/
SceneMixer pricing: 50 free credits; memberships from $7.99/mo, credit packs from $1.49; video from 10 credits per second. Every unit price on one page.

Pricing

# Memberships, credit packs and unit costs

Updated October 15, 2026 · prices are read from the live configuration and match checkout

SceneMixer pricing: 50 free credits; memberships from $7.99/mo, credit packs from $1.49; video from 10 credits per second. Every unit price on one page.

## Memberships

### Pro

$7.99/mo
≈ $0.011 per credit

- 700 credits per month
- Quarterly $20.99 (≈$7.00/mo)
- Yearly $74.99 (≈$6.25/mo)

### Studio

$19.99/mo
≈ $0.010 per credit

- 2,000 credits per month
- Quarterly $52.99 (≈$17.66/mo)
- Yearly $186.99 (≈$15.58/mo)
Popular

### Enterprise

$44.99/mo
≈ $0.009 per credit

- 5,000 credits per month
- Quarterly $119.99 (≈$40.00/mo)
- Yearly $419.99 (≈$35.00/mo)

### Ultimate

$94.99/mo
≈ $0.008 per credit

- 11,500 credits per month
- Quarterly $252.99 (≈$84.33/mo)
- Yearly $886.99 (≈$73.92/mo)

### Max

$199.99/mo
≈ $0.008 per credit

- 25,000 credits per month
- Quarterly $532.99 (≈$177.66/mo)
- Yearly $1,866.99 (≈$155.58/mo)

Free: 50 credits on sign-up; First drama: 15 concept sheets + first storyboard free; one free 5-second preview per account; up to 15% top-up bonus; buy credit packs as you go, no monthly fee, no card required.

## Credit packs

| Pack | Credits | Bonus | Price | Per credit |
| Starter Package | 100 | — | $1.49 | $0.0149 |
| Basic Package | 500 | +10 | $6.49 | $0.0127 |
| Pro Package | 2,000 | +80 | $22.49 | $0.0108 |
| Flagship Package | 10,000 | +600 | $102.49 | $0.0097 |
| Studio Package | 50,000 | +5,000 | $479.49 | $0.0087 |
| Enterprise Package | 100,000 | +15,000 | $899.49 | $0.0078 |

## Credit cost per feature

Novel / script parsing: 4 credits per 3,333 characters of English text
Extra reference image: 7 credits each
Storyboard script: per episode by script length, 1 credit per 333 characters of English text
Video generation: from 10 credits per second (cheapest tier: Wan 3.0 · 480P; default tier 10), roughly $0.06–$0.47 per second from the cheapest to the priciest tier (at the lowest credit price: largest pack or annual membership, bonus included; small packs and monthly plans run a little higher); all tiers in the table below
Episode compositing: 1 credit per second
AI writing: 4 credits per 3,333 characters of English text

### Video tiers: price per second

| Tier | Credits / s | ≈ $ / s | Max clip (s) |
| Wan 3.0 · 720P | 19 | $0.12 | 30 |
| Wan 3.0 · 480P (default) | 10 | $0.06 | 30 |
| Wan 3.0 · 1080P | 38 | $0.24 | 30 |
| Seedance 2.5 Pro | 75 | $0.47 | 30 |
| Seedance 2.5 Lite | 38 | $0.24 | 30 |
| Seedance 2.5 · 480P | 19 | $0.12 | 30 |
| Seedance 2.0 | 63 | $0.39 | 15 |
| Seedance 2.0 Fast | 25 | $0.16 | 15 |
| Seedance 2.0 · 480P | 13 | $0.08 | 15 |
| MiniMax H3 | 25 | $0.16 | 15 |
| MiniMax H3 Fast | 16 | $0.10 | 15 |
| Kling 4.0 | 38 | $0.24 | 30 |
| Kling 4.0 Flash | 29 | $0.18 | 20 |
Per-second amounts at the lowest credit price: largest pack or annual membership, bonus included; small packs and monthly plans run a little higher; you are charged in credits.

## FAQ


---

# Novel to Video AI: Turn a Novel into a Series | SceneMixer

URL: https://scenemixer.com/novel-to-video/
Turn a novel into an AI short drama series: script analysis, consistent characters, storyboards and per-shot video, composited into full episodes.

Novel to video

# Turn a Novel into a Video Series with AI

Updated November 3, 2026

Quick answer: SceneMixer turns a novel into a watchable short-drama series in one pipeline: upload the novel (Chinese or English) → AI extracts characters, scenes, and an episode-by-episode structure → generates consistent character and scene art → writes storyboard scripts → renders video segment by segment (16:9 by default, 9:16 in one setting) → composites full episodes. Credit packages start at $1.49; a typical episode is generated in minutes, not months.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Why novels are the hardest input — and why it matters

A novel is not a screenplay. It has inner monologue, time skips, and dozens of characters who must look the same in chapter 1 and chapter 40. Generic text-to-video tools generate one clip per prompt and forget everything between prompts. The result: a heroine whose face changes every scene.

SceneMixer treats the novel as an IP to industrialize, not a prompt. The analysis step builds a persistent asset library — every character gets a full-body reference sheet with a locked visual identity, every location gets scene art — and those references are injected into every downstream shot. That is what keeps a 60-episode web novel adaptation coherent.

## How the SceneMixer pipeline works

- Upload your story — paste text or upload .txt / .docx / .md. Chinese and English are both supported, or let the built-in AI writer draft a script for you.

- AI script analysis — one pass extracts characters, scenes, props, and a full episode structure, each episode outlined with conflict and a cliffhanger.

- Consistent reference art — full-body character sheets (9:16) plus scene and prop art in your chosen visual style (realistic, 2D, 3D CG and more). These references anchor every later shot, so faces and outfits stay consistent across episodes.

- Storyboard scripts — every episode is broken into 4–15 second video segments with editable shot descriptions. Editing the script is free.

- Video generation & compositing — segments are generated at 16:9 landscape (the default) or 9:16 vertical, on Wan 3.0 by default, with per-segment version history, then composited into a full episode with one click.

Localization is built in: pick a target region (Western, East Asian, Southeast Asian, Latin American) and casting, character names, scenery, and on-screen text adapt to that market.

## What you get per episode

| Output | Details |

| Episode outline | Conflict-driven summary with an end-of-episode hook (identity reveal, emotional cliffhanger, or reversal) |

| Storyboard script | 4–15s segments with shot language, dialogue, and duration — fully editable, editing is free |

| Video segments | 9:16 or 16:9, generated with consistent character references, with version history per segment |

| Full episode | One-click compositing of all segments into a single video, ready to publish |

## How long a novel can be

Length is where most novel-to-video attempts quietly fail, so it is worth being specific. A single analysis call has an output budget, and when a book exceeds it the model does not raise an error — it simply stops early. In our own offline evaluation a 50-episode script came back as ten episodes, with nothing in the response to say that forty had been dropped. Past roughly 10,000 equivalent characters SceneMixer therefore splits analysis into two calls: one that returns the cast, locations and props, and a second that returns the episode skeleton. Each episode's full outline is then expanded when you open that episode, so a sixty-episode upload does not pay for sixty expansions you may never use.

What happens at each length, for an English-language upload

| What you upload | How it is analysed | Analysis cost |

| Up to about 33,000 characters (≈ 5,500 words) | One call, whole book in view | about 40 credits |

| About 33,000 to 300,000 characters (≈ 5,500–50,000 words) | Two calls — cast first, then the episode skeleton — with per-episode outlines expanded on demand | scales with length, about 360 credits at the ceiling |

| Beyond 300,000 characters | Split the book into volumes and run them as separate projects; the cast can be carried across | per volume |

Analysis is the cheap part either way: even at the ceiling it is a rounding error against the video bill for the season it produces. Chinese-language uploads use the same architecture with a different character budget, because a Chinese character carries far more story than a Latin one.

## What adaptation actually changes

Prose has three things a camera cannot photograph, and the analysis step has to do something with each of them.

- Inner monologue becomes a marked line — inner voice, voice-over or narration — rather than a shot of someone thinking. It is written that way in the storyboard, so the performance is heard while the picture shows something worth watching.

- Time skips become episode boundaries or explicit on-screen captions. Three years passing in half a sentence of prose is a cut, and it has to be declared as one.

- Description becomes reference art, not narration. A paragraph about a character's face turns into their sheet; a paragraph about a room turns into a location plate. Neither is re-described shot by shot afterwards, which is exactly why the face holds.

What adaptation does not do is invent plot. Scene order, dialogue and beats follow the book; what the pipeline adds is coverage — who is in frame, at what shot size, for how long.

## What it costs

SceneMixer is credit-based — you pay for what you generate, not a flat seat license. Credit packages start at $1.49; membership tiers from $7.99/month add monthly credits, higher concurrency, and member-only features such as custom style prompts, uploads, and AI image fine-tuning. Script analysis is priced by text length (4 credits per 3,333 characters, reference images billed separately, with 15 free per account across characters, scenes and props); storyboard scripts are billed per episode at 1 credit per 333 characters, and regenerating one costs the same; video generation starts at 10 credits per second (≈ $0.06/s at the lowest credit price) of output (the rate depends on the model tier you pick), and episode compositing is 1 credit per second — see the full pricing table.

## A novel in one language, a series in fifteen

Pick the dialogue language in Step 1 and the cast performs the adaptation natively: a Chinese web novel can become an Indonesian-dialogue series, an English manuscript a Korean one, with names, streets and on-screen text adapted to that market by the target-region setting. The working documents you read and edit follow your interface language, and on SceneMixer a segment costs the same credits in any language. One finished sample per language is below; the samples page for 15 languages has them all.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### Can I turn a Chinese web novel into an English-language drama?

Yes. SceneMixer separates the script language from the production language: the output (episode outlines, storyboards, on-screen text) follows your interface language, and target-region localization adapts character names, casting, and scenery for Western, East Asian, Southeast Asian, or Latin American audiences.

### How does SceneMixer keep characters consistent across episodes?

Every character gets an AI-generated full-body reference sheet at analysis time. Face and body crops from that sheet are injected into every video segment that features the character, so the same face, hair, and signature look persist across all episodes.

### How long does it take to produce one episode?

Measured across every job that finished successfully on scenemixer.com in the thirty days to 20 September 2026: analysis of a whole book runs about 2 minutes 49 seconds at the median, one episode's storyboard 1 minute 54 seconds, one video segment 3 minutes 20 seconds, and compositing an episode 3 minutes 52 seconds. Segments run in parallel — three at a time on the free plan, up to twenty on the largest — so a typical seven-segment episode is around a quarter of an hour of machine time, and the first episode of a novel is watchable the same hour you upload it.

### How long can the novel be?

SceneMixer takes up to about 300,000 characters of English in one project — roughly 50,000 words. Past about 33,000 characters the analysis automatically switches to a two-call path built for long material, and per-episode outlines are expanded when you open each episode rather than all at once. Longer than that, split the book into volumes and run them as separate projects.

### Can I see a shot before committing to a season?

Yes. A new SceneMixer account carries 50 credits, 15 free reference images, a free episode-1 storyboard on the first project, and one free preview of up to 5 seconds that renders the opening shot of your own episode against your own cast — not a stock demo.

### How much does it cost to adapt a novel?

Script analysis is priced by text length (4 credits per 3,333 characters; reference images are billed separately, with 15 free per account) and storyboard scripts are billed per episode (1 credit per 333 characters). Video is billed per second of output. Credit packages start at $1.49, and Pro membership ($7.99/month) adds 700 monthly credits and member-only features.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# Script to Video AI: Screenplay to Episodes | SceneMixer

URL: https://scenemixer.com/script-to-video/
Turn a screenplay into finished vertical episodes with AI: scene breakdown, consistent cast, editable storyboards and per-second video generation.

Script to video

# Script to Video, End to End

Updated November 3, 2026

Quick answer: Paste or upload a script and SceneMixer produces the finished video: it breaks the script into scenes and 4–15 second shots, builds a consistent cast with reference art, writes storyboard prompts tuned for AI video models (Wan 3.0 by default), generates each segment in 16:9 or 9:16, and composites full episodes. You edit at the storyboard-script level — no timeline software required.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## At a glance

What you bring, what the pipeline makes, and what stays under your hand

| You bring | A screenplay, a treatment or a chapter file — pasted, or uploaded as .txt / .docx / .md. No formatting standard is required; the parser reads prose and screenplay layout alike. |

| It makes | A cast and location list with reference art, an editable shot list per episode, a rendered video per segment, and one composited file per episode. |

| Shot length | 4–15 seconds per generated segment; some model tiers reach 30. Runtime is built by stacking segments. |

| Frame | 16:9 landscape by default, 9:16 vertical available. Set once per project. |

| Spoken dialogue | Performed on screen in the language you choose before analysis — fifteen of them — not dubbed on afterwards. |

| Free to change | Every description and every line of the shot list. Only model calls are billed; editing text never is. |

| Billed | Analysis by script length, storyboards per episode, video per second of output, compositing per second. |

## What makes script-to-video different from text-to-video

Text-to-video gives you one clip per prompt. A script needs structure: who is in the shot, where it happens, what was established in the previous scene. SceneMixer's parser reads the whole script first and produces a production plan — cast list, locations, props, episode splits — before a single frame is generated. Each storyboard segment then references those shared assets, which is why shot 41 still matches shot 3.

## How the SceneMixer pipeline works

- Upload your story — paste text or upload .txt / .docx / .md. Chinese and English are both supported, or let the built-in AI writer draft a script for you.

- AI script analysis — one pass extracts characters, scenes, props, and a full episode structure, each episode outlined with conflict and a cliffhanger.

- Consistent reference art — full-body character sheets (9:16) plus scene and prop art in your chosen visual style (realistic, 2D, 3D CG and more). These references anchor every later shot, so faces and outfits stay consistent across episodes.

- Storyboard scripts — every episode is broken into 4–15 second video segments with editable shot descriptions. Editing the script is free.

- Video generation & compositing — segments are generated at 16:9 landscape (the default) or 9:16 vertical, on Wan 3.0 by default, with per-segment version history, then composited into a full episode with one click.

Localization is built in: pick a target region (Western, East Asian, Southeast Asian, Latin American) and casting, character names, scenery, and on-screen text adapt to that market.

## Built for AI-friendly cinematography

SceneMixer's storyboard writer favors shots AI video models render well — two-person dialogue close-ups, single-character emotional beats, strong lighting — and avoids known failure modes like large crowds or complex choreographed fights. You can edit any segment's script by hand (free) and regenerate just that segment, keeping the rest of the episode untouched.

## What script-to-video will not do

- Edit footage you already shot. Everything here is generated from the shot list. There is no timeline for imported clips, no face swap onto live-action plates, and no dubbing pass over existing video.

- Render one unbroken take of any length. Each generation is capped by the model tier you pick. A five-minute scene is a stack of segments, which is also what makes re-cutting one shot cheap.

- Reproduce a real person. Casting is generated, and voices are matched from a preset library or uploaded by you. There is no path here for recreating someone else's face or voice.

- Guarantee a shot you did not ask for. The storyboard is explicit on purpose: if a prop, a costume state or a location matters, it has to be named in the shot text. Anything left implied is the model's guess.

- Burn in captions. Text meant to be seen in frame is written into the shot; platform subtitles are added after export.

## What it costs

SceneMixer is credit-based — you pay for what you generate, not a flat seat license. Credit packages start at $1.49; membership tiers from $7.99/month add monthly credits, higher concurrency, and member-only features such as custom style prompts, uploads, and AI image fine-tuning. Script analysis is priced by text length (4 credits per 3,333 characters, reference images billed separately, with 15 free per account across characters, scenes and props); storyboard scripts are billed per episode at 1 credit per 333 characters, and regenerating one costs the same; video generation starts at 10 credits per second (≈ $0.06/s at the lowest credit price) of output (the rate depends on the model tier you pick), and episode compositing is 1 credit per second — see the full pricing table.

## How long it runs, and how to test it for nothing

Measured over every job that completed successfully on scenemixer.com in the thirty days to 20 September 2026: reading a whole script takes about 2 minutes 49 seconds at the median, a reference image 16 seconds, one episode's shot list 1 minute 54 seconds, one video segment 3 minutes 20 seconds, and compositing an episode 3 minutes 52 seconds. Segments run in parallel — three at once on the free plan, up to twenty on the largest — so a typical seven-segment episode is around a quarter of an hour of machine time end to end.

You can walk that whole path before paying: every account gets 50 credits at sign-up, 15 free reference images across characters, scenes and props, a free episode-1 shot list on the first project, and one free preview of up to 5 seconds that renders the opening shot of your own episode against your own cast.

## Fifteen dialogue languages, one script

Your script can stay in the language you wrote it in. The spoken dialogue is set per project before parsing and performed natively by the video model, so timing and mouth shapes match the language; nothing is dubbed afterwards, and the price per segment does not change. Below, one finished series per language from the same pipeline; see the samples page for 15 languages for the full set.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### What script formats can I upload?

Plain text, .docx, and .md — or paste directly. Both Chinese and English scripts are supported, and SceneMixer's built-in AI writer can draft or expand a script from a one-line idea.

### Can I control individual shots?

Yes. Every 4–15 second segment has an editable storyboard script (editing is free). You can regenerate a single segment without touching the others, and each segment keeps a version history you can switch between.

### Which video models does SceneMixer use?

Video generation runs on Wan 3.0 by default (480P, with 720P and 1080P one click away); other model tiers are selectable per episode, each with its per-second rate on the pricing page. Text analysis and image generation use a routed multi-model backend with automatic failover.

### How faithful is the shot list to my script?

Scene order, dialogue and beats follow the script; what the pipeline adds is coverage — which shot sizes, who is in frame, how long each shot holds. Lines are carried through rather than paraphrased, and if a shot reads wrong you rewrite that shot's text and re-render only it.

### Does it work with a screenplay in Final Draft or Fountain format?

Export to plain text first. SceneMixer's parser reads screenplay layout — slug lines, character cues, parentheticals — as well as ordinary prose, so a text export of a .fdx keeps its structure. .docx and .md upload directly.

### What happens with a feature-length script instead of a serial?

Set the content type to film in SceneMixer's Step 1 and the pacing rules change with it: longer scenes, different dialogue rhythm, and no episode hooks. Past roughly 10,000 equivalent characters the analysis also switches to a two-call path built for long material, which raises the length ceiling substantially.

### Do I need video editing skills?

No. In SceneMixer, the unit of editing is the storyboard script, not a timeline. When all segments are ready, full-episode compositing is one click, billed by total duration.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Micro Drama Generator for Vertical Series | SceneMixer

URL: https://scenemixer.com/ai-micro-drama/
Make vertical micro drama series with AI: hook-paced episode outlines, a cast that stays consistent, 9:16 video per shot and one-click compositing.

AI micro drama

# The AI Micro Drama Production Pipeline

Updated November 3, 2026

Quick answer: A micro drama (vertical drama) is a serialized show of 60–120 second episodes shot for phones. SceneMixer is a production pipeline purpose-built for the format: it turns a script or novel into episode outlines with cliffhanger pacing, keeps the cast visually consistent across the whole series, generates 9:16 vertical video segment by segment once the project is set to vertical (the default is 16:9), and composites finished episodes — with credit-based pricing that starts at $1.49.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Why micro drama needs a pipeline, not a prompt box

The micro-drama market reached roughly $11B in 2025 and keeps growing — but the format's economics only work if you can produce series, not clips. A binge-worthy vertical drama is 40–80 episodes with the same faces, recurring locations, and a hook every 90 seconds. That is an industrial-consistency problem, and it is exactly what SceneMixer's asset-library architecture solves.

## Format-native by design

- 9:16 in one setting — character sheets are always generated in vertical framing; video is 16:9 by default, and choosing 9:16 in the project settings makes every segment vertical.

- Cliffhanger pacing built into the outline — every episode summary is generated with a conflict arc and an end-of-episode nail: identity exposure, emotional suspension, or a reversal, rotated so hooks don't repeat.

- Segment economics — video is billed per second, so a 90-second episode is a predictable cost, and regenerating one 8-second segment doesn't re-bill the episode.

- Market localization — one story, multiple markets: region setting adapts names, casting, scenery, and on-screen text for Western, East Asian, Southeast Asian, or Latin American audiences.

## How the SceneMixer pipeline works

- Upload your story — paste text or upload .txt / .docx / .md. Chinese and English are both supported, or let the built-in AI writer draft a script for you.

- AI script analysis — one pass extracts characters, scenes, props, and a full episode structure, each episode outlined with conflict and a cliffhanger.

- Consistent reference art — full-body character sheets (9:16) plus scene and prop art in your chosen visual style (realistic, 2D, 3D CG and more). These references anchor every later shot, so faces and outfits stay consistent across episodes.

- Storyboard scripts — every episode is broken into 4–15 second video segments with editable shot descriptions. Editing the script is free.

- Video generation & compositing — segments are generated at 16:9 landscape (the default) or 9:16 vertical, on Wan 3.0 by default, with per-segment version history, then composited into a full episode with one click.

Localization is built in: pick a target region (Western, East Asian, Southeast Asian, Latin American) and casting, character names, scenery, and on-screen text adapt to that market.

## What it costs

SceneMixer is credit-based — you pay for what you generate, not a flat seat license. Credit packages start at $1.49; membership tiers from $7.99/month add monthly credits, higher concurrency, and member-only features such as custom style prompts, uploads, and AI image fine-tuning. Script analysis is priced by text length (4 credits per 3,333 characters, reference images billed separately, with 15 free per account across characters, scenes and props); storyboard scripts are billed per episode at 1 credit per 333 characters, and regenerating one costs the same; video generation starts at 10 credits per second (≈ $0.06/s at the lowest credit price) of output (the rate depends on the model tier you pick), and episode compositing is 1 credit per second — see the full pricing table.

## Micro drama for fifteen markets

A micro drama travels only if the cast speaks the audience's language. Set the dialogue language once in Step 1 and SceneMixer performs the whole series natively, with the target-region setting adapting names, faces and settings to that market. The sample series below, one per language, were all produced this way; the samples page for 15 languages lists every one of them.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### What is an AI micro drama?

A serialized vertical-video show (typically 60–120 seconds per episode, 9:16) where the cast, scenes, and footage are AI-generated. Viewers watch it like ReelShort or DramaBox content; creators produce it at a fraction of live-action cost.

### Can AI keep the same actors across 50 episodes?

That is SceneMixer's core design: characters live in a persistent asset library with locked reference art, and every segment that features a character is generated against those references — same face in episode 1 and episode 50.

### What does a 90-second episode cost to generate?

SceneMixer bills video per second of output, depending on the model tier you choose (Wan 3.0 at 480P by default; faster or higher-resolution tiers cost more per second). Script analysis (4 credits per 3,333 characters; reference images are billed separately, with 15 free per account) and storyboard scripts (1 credit per 333 characters per episode) are billed separately from video. Most creators start with a small credit package (from $1.49) to price out their first episode.

### Where can I publish the episodes?

You own the output — download finished episodes and publish to TikTok, YouTube Shorts, Instagram Reels, or short-drama platforms. SceneMixer's settings include an optional AIGC watermark toggle.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Short Drama Generator: Script In, Episodes Out | SceneMixer

URL: https://scenemixer.com/ai-short-drama-generator/
Generate a multi-episode AI short drama from one script: cast locked to reference art, shot-by-shot storyboards, 16:9 or 9:16 video per segment, one-click episodes.

AI short drama generator

# AI Short Drama Generator

Updated November 3, 2026

Quick answer: SceneMixer's short drama generator takes one script — pasted, uploaded as .txt / .docx / .md, or drafted by the built-in AI writer — and hands back finished episodes: a cast and location list with locked reference art, an editable shot-by-shot storyboard per episode, video rendered segment by segment at 9:16 or 16:9, and a one-click composite per episode. You are billed per unit of work rather than per seat, and credit packages start at $1.49.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## At a glance

What goes in, what comes out, and what you control

| Input | Paste the script, upload .txt / .docx / .md, or let the built-in AI writer draft one from a premise. |

| Output | An editable storyboard plus rendered video for every segment, and a composited file per episode. |

| Aspect ratio | 16:9 landscape by default; 9:16 vertical available. Set once per project. |

| Segment length | Typically 4–15 seconds per generated segment; some model tiers reach 30. |

| Episode length | However many segments the outline calls for — length comes from stacking, not from one long shot. |

| Cast consistency | Character sheets are generated once, then attached as references to every segment that character appears in. |

| Dialogue | Spoken on screen in the language you pick in Step 1; voices matched from a preset library. |

| Visual control | 21 render styles across live-action, 2D, 3D and stop-motion, plus 55 art-direction and 80 wardrobe presets; members can write their own. |

| Markets | A target-region setting adapts names, casting, scenery and on-screen text for Western, East Asian, Southeast Asian and Latin American audiences. |

| Billing | Credits, per unit of work: analysis by text length, storyboard per episode, video per second, compositing per second. |

## Five generation stages, and where you step in

- Script analysis — one SceneMixer pass reads the whole script and returns the cast, locations, props and an episode-by-episode outline, each episode carrying a conflict arc and an end-of-episode hook. Yours to change: every description, plus the project name and the three style axes, before anything is drawn.

- Reference art — full-body 9:16 character sheets, location plates and prop shots in the render style you picked. Yours to change: regenerate any sheet, upload your own image in its place, or add a state variant such as an outfit change, a ruined room or a drawn blade.

- Storyboard scripts — one episode at a time, cut into 4–15 second segments with shot descriptions, dialogue and per-shot durations. Yours to change: the text of any segment, free; asking the model for a fresh pass costs the same as the first one.

- Per-segment video — each segment is generated against the reference art for the characters and places it names. Yours to change: regenerate a single segment without re-billing the episode; every take is kept as a version you can switch back to.

- Compositing — the segments are joined into one episode file. Yours to change: swap a segment and re-composite.

The same five stages run whether you start from a finished script (script to video) or from a novel that still has to be adapted (novel to video); the micro drama page covers why the vertical format needs a pipeline in the first place.

## What stops the cast from drifting

Drift is the failure mode that kills generated series: by episode 5 the lead has a different jaw, and the audience stops believing it is the same show. Three mechanisms work against it.

- SceneMixer generates the cast once, not per shot. Character sheets live in the project, so no later shot re-invents a face from a text description.

- References are attached, not described. Face and body crops from the sheet ride along with every segment that names the character, so the model is matching a picture instead of interpreting adjectives.

- Links are by id, not by name. Storyboards reference assets by identifier, so renaming a character halfway through a project does not silently break the reference chain.

The mechanics are laid out in the character-consistency guide.

## What this generator will not do

- Clone a voice you do not own. Voices are matched from a preset library, and you may upload a sample of your own; there is no path here for reproducing someone else's voice.

- Burn in a subtitle track. Text that belongs inside the frame is written into the shot; platform captions are added after export.

- Re-edit or dub footage you upload. The generator renders new video from the storyboard. SceneMixer is not a post house for existing clips, and it does not swap faces onto live-action footage.

- Produce one continuous shot of unlimited length. Each generation is bounded by the model tier's cap; episode length comes from stacking segments.

## What a generated episode costs

Nothing here is a seat licence — you buy credits and spend them per unit of work. Script analysis is 4 credits per 3,333 characters of English text (reference images are billed separately, with 15 free per account and 7 credits each after that on the default image model); each episode's storyboard is 1 credit per 333 characters of that episode; video starts at 10 credits per second of output (≈ $0.06/s at the lowest credit price), and compositing is 1 credit per second. A 90-second episode therefore lands around 990 credits of video plus compositing on the cheapest tier. Every account gets 15 free reference images — characters, scenes or props, in any mix, across any project — and in your first project the storyboard for episode 1 is free as well. Packages start at $1.49; memberships from $7.99/month add monthly credits, higher concurrency, custom style prompts and generated video without the AIGC badge. The pricing page carries the full table.

## How long it takes, measured

The numbers below are not estimates. They are the median and ninth-decile run times of every job that finished successfully on scenemixer.com in the thirty days to 20 September 2026 — machine time for the job itself, with queue wait excluded.

Median generation time per stage, measured on scenemixer.com across every successful job in the 30 days to 20 September 2026. Queue wait is excluded.

| Stage | Typical | 9 in 10 finish within |

| Script analysis — whole script to cast, locations and episode outline | 2 min 49 s | 9 min 8 s |

| One reference image — character sheet, location plate or prop | 16 s | 27 s |

| One episode's storyboard | 1 min 54 s | 5 min 53 s |

| One video segment | 3 min 20 s | 7 min 32 s |

| Compositing one episode | 3 min 52 s | 12 min 51 s |

They stack differently than they read, because segments generate in parallel. A typical episode here is seven segments of about fourteen seconds each; on the free plan three jobs run at once, so those seven segments are three waves — roughly ten minutes — with the storyboard before them and the composite after. On SceneMixer, call it a quarter of an hour of machine time per episode, and less than that on a paid plan, where the concurrency ceiling rises from three to twenty.

## See one shot before you spend anything

Every account gets one free preview of up to 5 seconds. It is not a stock demo reel: it renders the opening shot of your episode, with your own character sheets attached as references, on the 480P tier — the same path a paid segment takes, minus the bill. The rest of the starting allowance works the same way: 50 credits at sign-up, 15 reference images per account, and your first project's episode-1 storyboard. That is enough to carry a real script from upload to a rendered shot and judge the result before deciding whether the series is worth paying for.

## Generated natively in fifteen languages

Pick the dialogue language in SceneMixer's Step 1 and the whole series is performed in it, with the target-region setting moving names, faces and settings to that market at the same time. The sample series below were each produced that way, one per language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### Can it generate a whole series, or only one video?

A series. One upload is analysed into an episode-by-episode outline, and every episode gets its own storyboard and its own set of video segments, all drawing on the same cast and locations. Tools that return a single clip make you start from scratch for episode 2; in SceneMixer, episode 2 inherits the project.

### How does it keep the same actor across episodes?

The cast is generated once, at the start, as full-body reference sheets. Every segment that features a character is generated against that character's sheet, with face and body crops attached as references — so the face in episode 1 is the face in episode 40. Since September 2026 those references are resolved by asset id rather than by name, so renaming a character mid-project does not break the link.

### What length and aspect ratio does it output?

Video is 16:9 landscape by default, with 9:16 vertical available; you set it once per project. SceneMixer's segments are typically 4 to 15 seconds, and some model tiers go to 30. Episode length comes from stacking segments, not from one long generation, so a 60 to 120 second episode is a handful of segments composited together.

### Can I edit what the AI wrote?

Yes, at every stage. Character, scene and prop descriptions are editable before any art is drawn; storyboard text is editable for free and re-renders only the segments you touched; each segment keeps a version history you can switch back to. On SceneMixer, only model calls cost credits — editing text never does.

### Does it generate the spoken performance too?

Yes. The video models used for generation produce picture and audio together, so characters speak on screen in the dialogue language you chose in Step 1. SceneMixer matches character voices from a preset voice library rather than designing them from a text description, and you can upload your own sample for a character instead. There is no tool here for cloning a voice you do not own.

### How long does a full series take?

Machine time scales with the number of segments, not the number of episodes — roughly a quarter of an hour per episode at the free plan's three-job concurrency, and proportionally less as that ceiling rises with the plan. What actually takes your time is review: reading the outline, correcting a character description, rewriting a line. None of that is billed, and most people spend longer choosing takes than waiting for them.

### What if I do not have a script yet?

Start from a premise. SceneMixer's built-in AI writer drafts a script from a one-line idea and a target length, in the language your interface is set to, and the draft drops straight into the same pipeline. It is a billed model call like any other, and you can rewrite the draft freely before analysis runs on it.

### Do I need to know anything about filmmaking?

No — the storyboard is written so that you do not have to. SceneMixer picks shot sizes, keeps two-hander dialogue in the framings video models render reliably, and steers around the ones they fail at, such as large crowds and long choreographed fights. If you do know the vocabulary, every shot is plain text you can rewrite, and only the segments you touch are re-rendered.

### Can I publish what it generates?

You own the output. Download finished episodes and publish them to TikTok, YouTube Shorts, Instagram Reels or a short-drama platform. Generated video carries a small AIGC badge by default; paid members can turn it off in settings.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Episode Generator: One Script, a Whole Season | SceneMixer

URL: https://scenemixer.com/ai-episode-generator/
Split one script into a numbered episode list: each episode gets its own outline, cliffhanger, storyboard, video and composite — and its own bill.

AI episode generator

# AI Episode Generator

Updated November 3, 2026

Quick answer: most AI video tools treat a single generation as the unit of work. SceneMixer treats an episode as the unit. One upload is split into a numbered episode list, each episode carrying its own outline, conflict arc and end-of-episode hook; from there every episode is storyboarded, generated and composited separately — so episode 3 can ship while episode 4 is still being written, and re-cutting one episode leaves the rest untouched.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## How one upload becomes a season

- The script is read once, whole. A single analysis pass returns the cast, the locations, the props and the episode list — not a chapter at a time, so episode 12 already knows who episode 2 introduced.

- Long scripts switch to two calls. Past roughly 10,000 equivalent characters, analysis splits into one call for the cast and locations and a second for the episode skeleton. The reason is blunt: a single call silently truncates long books. In our own offline evaluation a 50-episode script came back as 10 episodes with no error raised — the model simply stopped.

- Each episode gets an outline, then a full one on demand. The skeleton pass writes a three-to-five sentence outline per episode; the full treatment is expanded when you open that episode, so a 60-episode upload is not paying for 60 expansions you may never use.

- Storyboard, video, composite — per episode. Every stage after analysis is scoped to one episode and billed to one episode.

## What an episode carries

The fields generated for each episode

| Title | Generated from the episode's own content, editable. |

| Outline | A conflict arc, not a summary of events — with an emotional high point and an end-of-episode hook. |

| Hook type | Identity exposure, an unresolved emotional beat, or a reversal — rotated so consecutive episodes do not land the same way. |

| Cast and locations | Drawn from the project-wide asset library; new faces are added to it rather than invented per episode. |

| Storyboard | 4–15 second segments with shot descriptions, dialogue and per-shot durations. Generated per episode, billed per episode. |

| Composite | One file per episode, assembled from that episode's segments. |

## Per-episode economics

Because the episode is the billing unit, the cost of a season is the cost of an episode times the number of episodes — and a mistake costs you one episode, not a season. SceneMixer charges analysis once for the whole upload at 4 credits per 3,333 characters of English text. Each episode's storyboard is then 1 credit per 333 characters of that episode, first pass and re-run at the same rate. Video starts at 10 credits per second of output (≈ $0.06/s at the lowest credit price) and compositing is 1 credit per second, so a 90-second episode runs about 990 credits of video and assembly on the cheapest tier. On a first project, storyboard generation is free up to episode 1. Full table on the pricing page.

## A ten-episode season, itemised

A worked example with real rates rather than a range. Take a 30,000-character English script, the ten 90-second episodes it typically becomes, and the cheapest video tier (Wan 3.0 · 480P). Every figure below is pulled live from the same price list SceneMixer bills against.

30,000 characters of English script → 10 episodes × 90 seconds, on the cheapest video tier

| Line item | How it is charged | Credits |

| Script analysis | Once for the whole upload, by length | 36 |

| Storyboards | Per episode, by that episode's length × 10 | 90 |

| Video | Per second of output × 900 seconds | 9,000 |

| Compositing | Per second of finished episode × 900 seconds | 900 |

| Total | Fifteen minutes of finished series | 10,026 credits ≈ $62 |

Two things the table deliberately does not net out. Reference art is billed on top, at 7 credits an image on the default model — with 15 free per account, most first seasons pay for a handful. And the free allowance cuts the bill further: the episode-1 storyboard of your first project is free, as are 50 credits at sign-up. Move up the tiers and the video line is what moves: the same 900 seconds costs about seven and a half times more on the most expensive tier than on the cheapest ($0.47 a second against $0.06), and nothing else on the table changes.

## How long a season takes to come out the other side

Measured across every job that completed successfully on scenemixer.com in the thirty days to 20 September 2026: analysis of a whole script runs about 2 minutes 49 seconds at the median, one episode's storyboard 1 minute 54 seconds, one video segment 3 minutes 20 seconds, and compositing an episode 3 minutes 52 seconds. Because SceneMixer runs segments in parallel — three at a time on the free plan, up to twenty on the largest — an episode is about a quarter of an hour of machine time, and the episodes of a season can be queued while you are still reading episode one. None of that clock is billed; you are billed for output, not for waiting.

## What happens when you regenerate an episode

Worth knowing before you click it: asking for a fresh storyboard on an episode replaces that episode's segments outright. Any video already generated for the old segments is no longer attached to the episode, and any per-segment choices you made by hand are recomputed from the new storyboard. SceneMixer leaves other episodes untouched, and the project-wide cast and locations are untouched. If you only want to change a few shots, edit the storyboard text instead — that is free and re-renders only the segments you actually changed.

## How long should an episode be?

Vertical series usually run 60–120 seconds an episode, and that is what the outline generator paces for. Length is built from segments rather than requested as a number: a 90-second episode is roughly eight to twelve generated segments stacked end to end. If your source material is a feature-length script rather than a serial, set the content type to film in Step 1 and the pacing rules change with it. The micro drama page covers the format's economics; novel to video covers the adaptation step that comes before episode one.

## Every episode, in fifteen languages

The dialogue language is set once in Step 1 and the whole run is performed in it, with the target-region setting moving names, faces and streets to that market at the same time. Each sample below was produced that way, one per language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### How many episodes will it generate?

As many as the material supports — SceneMixer takes the episode count from the script's own structure, not from a number you pick. A 30,000-character novel typically lands around ten episodes of 90 seconds; a serial with explicit chapter breaks is split along them.

### Can I generate episode 5 without generating 1 through 4?

Yes. Storyboards and video are requested per episode, in any order. SceneMixer's only project-wide step is the initial analysis, which covers every episode at once.

### Do later episodes remember earlier ones?

SceneMixer keeps the cast, locations and props project-wide, so episode 40 uses the same asset library as episode 1, and the outline for every episode is written in one pass with the whole story in view. What is not shared is the video: each segment is generated independently against the reference art.

### What happens to a very long script?

Past roughly 10,000 equivalent characters SceneMixer runs the analysis as two calls instead of one — a pass for the cast and locations, then a pass for the episode skeleton — and each episode's full outline is expanded when you open it. The ceiling for the two-call path is far higher than for the single-call one.

### Can I see a shot before I pay for a season?

Yes. On SceneMixer, every account gets one free preview of up to 5 seconds, and it is your own material — the opening shot of your own episode, rendered against your own character sheets on the 480P tier. Together with the 15 free reference images and the free episode-1 storyboard, that is a full pass through the pipeline before any decision about the rest of the season.

### What happens if I abandon a season halfway through?

You are out the credits you spent on the episodes you generated, and nothing else. SceneMixer has no seat licence, no minimum term, and no charge for episodes that only exist as an outline — the skeleton pass writes them, but the expensive stages are requested one episode at a time. Unused credits stay on the account.

### Is regenerating an episode's storyboard cheaper the second time?

No — it is the same price as the first pass, because it is the same model call over the same text. On SceneMixer, editing the storyboard by hand is free, so it is usually the better move for small changes.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Storyboard Generator: Shots, Timing, Dialogue | SceneMixer

URL: https://scenemixer.com/ai-storyboard-generator/
Generate an editable shot-by-shot storyboard: numbered shots with durations, framing, tagged dialogue and ambient sound — then render each segment.

AI storyboard generator

# AI Storyboard Generator

Updated November 3, 2026

Quick answer: SceneMixer generates a storyboard as a working document, not a mood board. Each episode comes back as a list of 4–15 second segments; inside a segment, numbered shots carry their own duration, framing change, on-screen cast, dialogue with speaker tags and delivery notes, and a line of ambient sound. You can rewrite any of it for free, and only the segments you touched are re-rendered.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a generated storyboard actually contains

Per segment and per shot

| Segment | The unit that becomes one video generation. Typically 4–15 seconds; some model tiers reach 30. |

| Starting frame state | Where everyone and everything is at frame one — so a segment does not re-stage the room it inherited. |

| Location, time, weather | Named per segment, so the same location is lit and dressed the same way across an episode. |

| Props in play | Only the props the shot text actually mentions; anything listed but never used is dropped before the segment is rendered. |

| Numbered shots | Each with its own duration in seconds, framing change, and action text. The segment's length is the sum of its shots, not an independent number. |

| On-screen cast | Marked per shot. A character who is only being looked at or spoken to, and is not in frame, is deliberately not named — naming them is what makes a video model paste their face into a single-person shot. |

| Dialogue | Written inline on its own line, with the speaker tagged and a short delivery note in brackets (off-screen, on the phone, over the radio). |

| Ambient sound | One line per segment, describing what the scene sounds like. |

## Timing is derived, not guessed

The most common way a generated storyboard fails is a shot that gives an actor 1.5 seconds to say a nine-word line. SceneMixer computes durations from the spoken length of the dialogue: English is paced at three words per second, Chinese at roughly five characters per second, and dialogue shots get a floor so a short line still has room to land. The same formula is written into the generation prompt and re-checked after the model answers, so a shot that comes back too short for its own dialogue is corrected before it reaches you.

## A working document, not a preview

- Editing is free. Rewriting shot text, dialogue or durations costs nothing — only model calls consume credits.

- Edits are surgical. Changing one segment re-renders that segment. The rest of the episode, and every segment you did not touch, stays byte-for-byte as it was.

- SceneMixer keeps every take. Each render is a version on the segment; switch back to an earlier one at any time.

- Asking the model for a fresh pass costs the same as the first one — it is the same call over the same text — so for small fixes, edit rather than regenerate.

## From storyboard to video

Each segment is rendered against the reference art for the characters and places it names: the character sheet crops ride along as image references, the location plate anchors the space, and the shot text supplies everything the references cannot carry — movement, lighting for that beat, weather, what is being held. Since September 2026 those references are resolved by asset identifier rather than by name, so renaming a character halfway through a project does not quietly break the link between a shot and the face it was supposed to use. The consistent-character page covers that side in detail.

## What it costs

Storyboards are billed per episode by that episode's share of the script: 1 credit per 333 characters of English text. On a first project, generation is free up to episode 1. On SceneMixer, rendering the segments afterwards starts at 10 credits per second of output (≈ $0.06/s at the lowest credit price). Editing, re-reading and version switching are free. See the pricing page for the full table, or the generator overview for how the storyboard fits between analysis and video.

## Every episode, in fifteen languages

The dialogue language is set once in Step 1 and the whole run is performed in it, with the target-region setting moving names, faces and streets to that market at the same time. Each sample below was produced that way, one per language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### Is the storyboard images or text?

Text, with numbered shots — and then real video for each segment. SceneMixer has no intermediate frame-by-frame image grid to approve; the reference art for characters, locations and props is what you approve visually, and the storyboard is what you approve as writing.

### Can I write my own storyboard instead of generating one?

You can rewrite any generated storyboard freely, including replacing a segment's text wholesale. What the pipeline needs is the structure — numbered shots with durations and tagged dialogue — because the renderer reads those fields.

### Why does it mark which characters are on screen?

Because naming an off-screen character inside a single-person shot is the single most reliable way to make a video model put that person's face in frame. Frame-by-frame testing showed eight shots that named an absent opponent producing five face swaps, against zero across twelve shots that wrote a direction instead. So SceneMixer's storyboard marks who is actually in frame, and writes directions for everyone else.

### Can I change how long a shot is?

Yes, durations are editable. The segment's length is the sum of its shots, so lengthening one shot lengthens the segment — and the segment still has to fit inside the model tier's cap.

### Does editing the storyboard re-bill the episode?

No. On SceneMixer, editing text is free, and re-rendering is charged per second of the segments you actually re-render.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Video Generator With Consistent Characters | SceneMixer

URL: https://scenemixer.com/ai-consistent-character-video-generator/
Generate video where a character looks the same in every shot: locked reference sheets, face crops attached per segment, references linked by id.

Consistent characters

# AI Video Generator With Consistent Characters

Updated November 3, 2026

Quick answer: character drift is what stops most AI video from becoming a series: by episode 5 the lead has a different jaw and the audience stops believing it is the same show. SceneMixer attacks it with four mechanisms — a cast generated once as reference sheets, face and body crops attached to every segment that character appears in, references linked by identifier rather than by name, and an explicit rule against naming people who are not in frame. Deliberate changes get their own reference art instead of a prompt.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Why characters drift in the first place

A text-to-video model renders each generation independently. Describe a woman in her thirties with a scar through one eyebrow in shot 1 and again in shot 30, and you get two women who match the description and not each other — adjectives underdetermine a face. Every fix below is the same idea in a different place: stop describing the character, and hand the model a picture of them.

## Four locks

What each mechanism does

| Reference sheet | Each character is generated once as a full-body sheet in 9:16 on a plain background under directional key light — a casting photo, not a mood piece, because whatever light is baked into the sheet gets inherited by every shot that uses it. |

| Attached crops | Face and body crops from that sheet are sent with every segment the character appears in. The model matches a picture instead of interpreting a description. |

| Links by id | Storyboards point at assets by identifier, not by name. Two characters with the same name, or a rename halfway through a project, no longer break the chain silently. |

| Off-screen rule | A character who is only being looked at, spoken to or reached for — and is not in frame — is written as a direction, never named. Naming an absent character is what makes a model paste their face into a single-person shot. |

A fifth, smaller one: SceneMixer gives each person in a shot one or two distinguishing facial notes pulled from their own description, which takes the edge off the classic failure where two characters in the same uniform trade faces mid-shot.

## When the character is supposed to change

Consistency cannot mean "never changes" — a character who is set on fire in episode 6 has to look burned in episode 7. The dividing line is drawn by how a video model treats a reference image: it copies the face, the clothing, the form and the structure almost literally, and it applies light, weather and time of day as effects layered on top.

- Give it a state variant when what changes is appearance, wardrobe, form or structure: a full change of outfit, a transformation, burns, a collapsed room, a blade out of its sheath, a snapped object. SceneMixer gives the variant its own reference image, generated from the base art.

- Write it in the shot text when what changes is light, weather, time of day, position, what someone is holding, or what has landed on them. Rain does not need a second version of the street.

Getting this backwards is expensive in both directions: a variant for "holding the cup" is an image you paid for and will never see, and a prompt for "after the fire" is a request the model will politely ignore in favour of the clean room in its reference.

## What is still not solved

- Long coats swallow the face. A full-body 9:16 sheet of a character in a long coat or robe spends most of its pixels on fabric, and the head ends up small enough that fine identity cues are lost. It is a framing limit, not a prompting one.

- Two characters dressed identically in one shot can still trade faces. Distinguishing notes help; matching uniforms remain the hardest case.

- Crowds. SceneMixer's consistency guarantees cover the cast in your asset library. Background extras are generated fresh each time and should be treated as scenery.

The layered explanation, with what each layer can and cannot catch, is in the character-consistency guide.

## What it costs

Reference sheets are their own line item, separate from script analysis (4 credits per 3,333 characters of English text): each account gets 15 reference images free, and extra or regenerated images after that are 7 credits each on the default image model. State variants carry their own small free allowance on a first project and are billed as images after that. Video itself starts at 10 credits per second of output on SceneMixer (≈ $0.06/s at the lowest credit price) whether or not references are attached — consistency is not a surcharge. Details on the pricing page.

## Prove it on your own cast, for nothing

Consistency is the one claim in this category you should never take on trust, and you do not have to. A new account carries 50 credits, 15 free reference images across characters, scenes and props, a free episode-1 storyboard on your first project, and one free preview of up to 5 seconds that renders the opening shot of your own episode against your own sheets. That is enough for a real test:

- Paste a scene in which the same two people appear several times, minutes apart.

- Let the analysis build the cast, then read the character descriptions and fix anything wrong before art is drawn — the sheets are generated from that text.

- Generate the sheets. A reference image takes about 16 seconds at the median, and 9 in 10 finish inside 27 seconds, so a five-person cast is a couple of minutes.

- Spend the free preview on a shot where a character is doing something, not standing still, and compare the face against the sheet.

For calibration: the median project on this platform has five named characters, and nine in ten have fewer than ten. If your story needs thirty, the same mechanism still applies — the cost is the sheets, not the consistency.

## Every episode, in fifteen languages

The dialogue language is set once in Step 1 and the whole run is performed in it, with the target-region setting moving names, faces and streets to that market at the same time. Each sample below was produced that way, one per language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### How consistent is it, really?

The same face across an entire series is the design target and what SceneMixer's reference-sheet architecture delivers for the cast in your asset library. It is not a guarantee on every frame: identically dressed characters sharing a shot, and characters whose sheet is mostly coat, remain the hard cases. Everything else — episode count, elapsed time, how many shots a character appears in — does not degrade it, because every shot is matched against the same picture.

### Can I upload my own character image instead of generating one?

Yes. Any reference sheet can be replaced with an image you upload, and SceneMixer treats it exactly like a generated one. That is the route for a character you have already designed elsewhere.

### Does it work with a real person's face?

Technically the upload path does not care what the image is, but publishing AI video of a real person's likeness is a legal problem in most markets and our compliance guide treats it as one. We do not provide a face-swap tool, and we do not recommend it.

### What happens if I rename a character mid-project?

Nothing breaks. SceneMixer's storyboards reference assets by identifier, and the rename is propagated through the stored text at the same time. Before September 2026 references were matched by name, and that is exactly the failure this change removed.

### Does keeping a character consistent cost extra per shot?

No. On SceneMixer, attaching reference art to a segment does not change the per-second rate — you pay once for each reference sheet, then nothing further no matter how many shots use it. Consistency is an architecture here, not an upsell.

### Do I have to re-describe the character in every shot?

No, and you should not. The shot text names the character; the reference art carries what they look like. Repeating a physical description in the shot text pulls the model back towards the words and away from the picture, which is precisely the drift this design removes.

### Can one character have two looks in the same episode?

Yes — that is what state variants are for. Each look is its own reference image, and SceneMixer's storyboard decides per segment which one applies, down to the individual segment rather than the whole episode.


---

# AI Scene Generator: Location Plates for a Series | SceneMixer

URL: https://scenemixer.com/ai-scene-generator/
Generate reusable location and prop art for a series: depth-layered plates with a named light source, several camera angles, neutral by design.

AI scene generator

# AI Scene Generator

Updated November 3, 2026

Quick answer: a scene generator for a series has a different job from one that makes a nice picture. SceneMixer generates each location once as a plate — three layers of depth, a named light source, no grading and no people — and then every shot set in that location is generated against it. That is what makes the coffee shop in episode 2 the same coffee shop in episode 30, and it is why the plates deliberately look neutral rather than cinematic.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A plate, not a poster

The temptation with location art is to make it beautiful: dramatic side light, fog, a shaft of sun through the window. That is exactly wrong for reference art. Whatever light is baked into a plate is inherited by every shot that uses it, and then fights the per-shot lighting the storyboard asks for — you end up with a 3 a.m. scene lit like a noon scene because the plate said so. So plates are generated to a narrower brief:

- SceneMixer builds three layers of depth — foreground, middle, background — so the model has real spatial structure to work with rather than a flat wall.

- The light source is named, not its effect. "A bare bulb above the counter" rather than "warm light on the countertop". Describe the effect and the model paints a patch of colour exactly where you said it lands.

- No mood grading, no silhouettes, no fog, no light shafts, and no people. Anything that competes with the space for attention is noise in a reference.

- Several angles per location. A location can carry multiple sub-plates for different camera positions, so cutting around a room does not mean re-inventing it.

## Three style axes, three jobs

SceneMixer controls the look with three independent settings, chosen once in Step 1. They are kept apart on purpose: overlapping them is how a project ends up with every character dressed the same and every room graded the same.

What each axis controls, and where it is deliberately excluded

| Render style (21 presets) | The medium and its texture — live-action, 2D, 3D, stop-motion. Applies to character art, location art, prop art and video alike. |

| Art direction (55 presets) | What gets built: spatial language, materials, colour discipline, how densely a room is dressed. Applies to locations, props and video — and is excluded from character sheets, which stay neutral casting photos. |

| Wardrobe (80 presets) | What people wear: silhouette, cut, fabric, construction. Injected only at script analysis and as a fallback for characters whose clothing is unspecified — never into location or prop art. |

Members can write their own descriptor for any axis instead of picking a preset. Whichever you choose, the rule for writing one is the same as everywhere else in this pipeline: describe the cause, not the result. "Concrete, steel and unfinished plaster" is a brief; "dramatic and impressive" is not.

## Props get the same treatment

Props are generated on a plain background with lighting used only to reveal material — is it brushed or polished, cast or forged. An object that lives in a sheath or a case is drawn with the contents out and the container beside it, because a plate of a sheathed sword teaches the model that this object is a sheathed sword, and the moment a character draws it on screen the blade turns back into a scabbard.

## Reuse, limits and cost

Location and prop art is generated once per project and reused by every episode — that reuse is the whole point, and it is also why the art is worth reviewing before you generate video. On SceneMixer, a project holds up to 100 characters, 100 locations and 100 props by default, within an overall budget of 200 assets and shots. Location and prop art is billed separately from script analysis (4 credits per 3,333 characters of English text): each account gets 15 reference images free, and images after that are 7 credits each on the default image model. Full table on the pricing page; the generator overview shows where scene art sits in the run.

## Every episode, in fifteen languages

The dialogue language is set once in Step 1 and the whole run is performed in it, with the target-region setting moving names, faces and streets to that market at the same time. Each sample below was produced that way, one per language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### Can I generate a scene without a script?

SceneMixer generates scene art as part of a project, from the locations the script actually uses, because its purpose is to be reused by that project's shots. If you want a single standalone image rather than a series, a general image generator is the better tool.

### Why do the location plates look flat compared with the finished video?

By design. In SceneMixer, a plate is a reference, so it carries structure and a light source and nothing else; the drama is added per shot by the storyboard, which can light the same room differently at noon and at midnight. A plate that already looks like a finished frame would override that.

### Can one location have several camera angles?

Yes. A location can hold multiple sub-plates for different camera positions, generated together, so cutting around a space stays coherent.

### Can I upload a photo of a real place?

Yes — any plate can be replaced with an image you upload, and SceneMixer treats it as the reference for that location.

### What is the difference between art direction and render style?

Render style is how it is drawn: live-action, 2D, 3D, stop-motion. Art direction is what is built: the materials, the colour discipline, how full a room is. You can put a spare brutalist interior in a 2D cel style or in photoreal live-action, and the two settings do not interfere.


---

# AI Script Generator for Short Drama | SceneMixer

URL: https://scenemixer.com/ai-script-generator/
Write a short-drama script from a premise in your own language, then run it straight down the pipeline: analysis, reference art, storyboards, video.

AI script generator

# AI Script Generator

Updated November 3, 2026

Quick answer: if you have a premise but not a script, the built-in writer drafts one inside the project: you give it an idea, a length and a tone, and it returns a full serial script in the language your interface is in — not only English. The draft lands in the project as the source text, so the next click is script analysis rather than a copy-paste into a different tool.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## It writes in your language, not just English

Two separate settings decide what language things come out in, and conflating them is the mistake most pipelines make.

- Writing language — the language of the script itself. It follows your interface, across all sixteen supported languages. A Korean interface gets a Korean script.

- Production language — the language of the working documents the pipeline generates from it: the episode outlines, the shot descriptions, the asset notes. That stays English (or Chinese on the mainland edition), because those documents are read by the next model in the chain.

The result is that a Korean script behaves exactly like a Korean script you pasted in yourself: English production documents, translated back into Korean for you to read on screen, and dialogue spoken in whichever language you choose in Step 1. Until September 2026 the writer answered in English regardless of the interface, which made it useless outside one market; that is fixed.

## Pick a length that the pipeline can actually carry

What you set before it writes

| Premise | One line to a few paragraphs. The more specific the conflict, the less generic the draft. |

| Length | Chosen from length tiers rather than typed as a number, and any tier that would not fit the analysis ceiling is hidden rather than offered and then rejected. |

| Episodes | Budgeted at roughly 720 words per episode for English, so a length tier is also a rough episode count. |

| Tone and genre | Carried into the draft; the visual style axes are set later, in Step 1, and do not constrain the writing. |

SceneMixer counts English length in words rather than characters, deliberately. Counting an English script by characters — the unit that works for Chinese — cuts the draft to about a quarter of a usable length, which is what our own international testing showed before the unit was changed.

## Then it keeps going

The draft is written into the project as its source text. From there the run is identical to one that started from an uploaded script: analysis extracts the cast, locations, props and the episode list; reference art is generated; each episode gets a storyboard; segments render; episodes composite. Nothing is exported and re-imported, and you can edit the draft at any point before analysis — or throw it away and paste your own. See script to video for the path from a finished screenplay, or the generator overview for the whole run.

## What it costs

SceneMixer bills writing by the length you asked for and by the language you are writing in. Alphabetic scripts — Latin, Cyrillic, Greek, Arabic, Hebrew and the rest — are billed at 30 per cent of their raw character count, because the same story takes far more characters in English than in Chinese; Chinese, Japanese and Korean are billed at full count. Everything after the draft is priced as normal: analysis at 4 credits per 3,333 characters of English text, storyboards at 1 credit per 333 characters per episode, video from 10 credits per second of output. Credit packages start at $1.49; see the pricing page.

## What it is not

- Not a chat you co-write in. You brief it and get a draft. Revision happens by editing the draft, or by asking for another one.

- Not a requirement. The writer exists for people who arrive with an idea rather than a manuscript. Uploading your own script skips it entirely and costs nothing.

- Not a rights launderer. A draft written from someone else's copyrighted premise is still someone else's premise; our compliance checklist covers what that means before you publish.

## Every episode, in fifteen languages

The dialogue language is set once in Step 1 and the whole run is performed in it, with the target-region setting moving names, faces and streets to that market at the same time. Each sample below was produced that way, one per language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### What languages can it write in?

All sixteen interface languages, including simplified and traditional Chinese. The script follows your interface language; SceneMixer's internal working documents stay in English, and what you read on screen is translated back.

### Can I edit the script it writes?

Yes, freely, before analysis — it is just the project's source text at that point. You can also replace it entirely with your own.

### How long a script can it write?

On SceneMixer, length is picked from tiers rather than typed in, and tiers that would exceed what the analysis stage can take are not offered. English drafts are budgeted at roughly 720 words per episode.

### Is the writer free?

No. SceneMixer's writer is a model call, billed by target length and by the language you write in, with alphabetic scripts billed at 30 per cent of their character count. Uploading your own script instead costs nothing.

### Does the dialogue language have to match the script language?

No. On SceneMixer, dialogue language is set separately in Step 1, so a script written in German can be performed in Spanish. The cast, names and settings adapt to the market you target.


---

# AI Character Voices for Short Drama: Matched, Not Cloned | SceneMixer

URL: https://scenemixer.com/ai-character-voices/
How characters speak in a generated drama: the line is performed by the video model, the voice comes from a preset library or your own sample, and no voice is cloned.

AI character voices

# AI Character Voices

Updated November 3, 2026

Quick answer: in SceneMixer the spoken line is not dubbed on afterwards — the video model generates picture and speech together, so a character's mouth and their line come from the same pass. What you control is which voice: each character carries either a written description of timbre and delivery, or a sample matched from a preset voice library, or a sample you upload. What we do not offer is a way to clone a voice you do not own.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Where the voice actually comes from

Most AI video workflows treat audio as a second job: generate silent footage, synthesize lines, then fight the lip sync. The video models here generate audio and picture in one pass, which removes the fight but changes what a “voice setting” is. It is not a text-to-speech engine you feed a script into — it is a reference the generation pass is conditioned on.

Three ways a character can carry a voice

| Written description | Age, timbre, weight, delivery — generated from the script along with everything else about the character, and performed by the video model. |

| Matched preset | A voice picked from a library of system voices, each with a short audio sample you can listen to. One model call proposes a match per character by gender, language and age; you can re-match or pick another by hand. |

| Your own sample | On a paid plan, upload audio for a character and it becomes that character's reference for every segment they speak in. |

Whichever is attached is used consistently: the same character speaks with the same voice across every segment and every episode, for the same reason their face stays the same — the reference travels with them.

## What we will not do, and why

SceneMixer has no tool for cloning a voice from a recording of someone else. It is the deliberate kind of gap, not a roadmap item: publishing a synthetic performance of a real person's voice is a likeness problem in most markets we serve, and a product that makes it one click is a product that makes it routine. Our own compliance checklist treats likeness — face and voice alike — as something you must have the right to use before you publish.

Two things that follow from it, worth saying plainly:

- Uploading your own voice is fine. That is your likeness, and the upload path exists for exactly that.

- Designing a voice from a text description is not the mechanism here. That chain was removed in August 2026 in favour of matching real sampled voices, because a voice designed from adjectives drifts between takes in the same way a face described in adjectives does.

## Language is a separate setting

The dialogue language is chosen in Step 1 and applies to the whole project: the lines are written in that language at the analysis stage, copied verbatim into the storyboard, and performed in it. This is not a dubbing pass over an English original — there is no English original. The cast, names and settings can move to the matching market at the same time through the target-region setting. SceneMixer supports fifteen languages, and the case against dubbing goes into why generating the line beats re-timing it.

## What it costs

Voice matching is a free model call — it does not consume credits and does not count against your concurrent-task limit. Uploading your own sample costs no credits, though on the current plan settings uploading is a membership feature. The audio itself is part of video generation, which starts at 10 credits per second of output (≈ $0.06/s at the lowest credit price); there is no separate per-word or per-character charge for speech. See the pricing page, or the generator overview for where voice sits in the run.

## What you still cannot do

- Re-voice a clip you upload. SceneMixer generates speech with the picture; there is no pass that replaces the audio on existing footage.

- Direct a line by ear. Delivery is steered by the written notes in the storyboard — a bracketed hint on each line — not by a waveform editor. If a reading is wrong, you change the note and regenerate the segment.

- Burn in subtitles. Captions for a platform are added after export; the pipeline writes text into the frame only when the story calls for it.

## Every episode, in fifteen languages

The dialogue language is set once in Step 1 and the whole run is performed in it, with the target-region setting moving names, faces and streets to that market at the same time. Each sample below was produced that way, one per language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### Can SceneMixer clone a voice?

No. Voices are matched from a library of sampled preset voices, or supplied by you as an upload of your own audio. There is no path for reproducing a voice from a recording of someone else, and that is intentional rather than unfinished.

### Is the speech dubbed on after the video?

No. SceneMixer's video models generate picture and audio in the same pass, so the performance and the mouth movement come from one generation. That is also why changing a line means regenerating that segment rather than re-cutting an audio track.

### Can I upload my own voice?

Yes, on a paid SceneMixer plan. Upload a sample for a character and it becomes that character's voice reference for every segment they speak in.

### Do all the characters have to speak the same language?

SceneMixer's dialogue language is a project-level setting, so yes within one project — but the language is chosen independently of the language you wrote the script in, and independently of your interface language. One script can be produced for several markets by running it with different settings.

### Does the voice stay the same across episodes?

Yes, for the same reason the face does: whatever voice is attached to a character is attached at the project level and travels into every segment that character speaks in.


---

# AI Ad Video Generator: One Photo to a Finished Spot | SceneMixer

URL: https://scenemixer.com/ai-ad-video-generator/
Upload a product, storefront or app photo and get a finished 15, 30 or 60 second ad: script, cast, storyboard, generated video and an end card — at one quoted price.

AI ad video generator

# AI Ad Video Generator

Updated November 3, 2026

Quick answer: upload one to four photos of what you are selling and SceneMixer writes the spot and shoots it: it reads the photo for free first — category, product name, brand name off the signage, selling points — then writes a creative script with spoken lines, builds the cast and locations, storyboards 15, 30 or 60 seconds, generates every segment with sound, and finishes on an end card carrying your brand, tagline and call to action. You are quoted one price before it starts, and charged no more than the quote.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | One to four photos of what you sell plus a one-line description; 15, 30 or 60 seconds, one quoted ceiling price |

## At a glance

What the ad generator takes, what it hands back, and what you decide

| You upload | Up to four photos of the same subject — a product, a dish, a garment, a car, a storefront or a property, or a screenshot of your app or site. Different angles and details of one thing, not four different things. |

| You type | Optional. A sentence about what you are selling helps, and when the photo and the sentence disagree, the sentence wins. |

| Length | 15, 30 or 60 seconds — one, two or four generated segments. |

| Frame and resolution | 16:9 or 9:16; 480P, 720P or 1080P. |

| Tone | Eight recipes: Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life. |

| It writes | A creative script with two to four spoken lines, a cast, locations, a storyboard, the generated video and an end card carrying your brand, product name, tagline and call to action. |

| Sound | Speech, ambience and music are generated together with the picture. No library track is mixed in afterwards. |

| Price | One quoted figure before you start, and it is a ceiling — you are charged for the stages that actually ran, never more than the quote. |

## It reads the photo before you write a word

Press the button with nothing typed and the first thing that happens is free: SceneMixer reads the photo, and you are told what came back — whether the subject is recognisable at all, what category it falls into, a product name, a description, the selling points worth filming, a suggested tone, and any brand name legible on the packaging or the signage, transcribed exactly as printed rather than guessed at. If the read fails, you find out there and then, at no cost, with a button to swap the photo.

Two things that read decides, and they are the only reasons the category matters at all:

- Whether the subject is redrawn. A bottle or a jar is worth re-rendering as a clean studio still. A shop front, a car or a house is not: cut it out on white and the signage, the depth and the sense of scale all go with it, so those photos are used as they are.

- Whether the product is a prop or a place. If you are advertising a venue, every scene has to happen inside it — otherwise the model quietly invents a different room and shoots the ad there.

Photographs of shops and restaurants almost always have customers and staff in them. Rather than refusing them, the pipeline removes the people first and keeps the signage, the fittings and the light — because a video model handed a reference image copies the faces in it almost literally, and those are not faces you have permission to put in an ad. Screenshots are the one case that is turned away: editing a screenshot blurs the interface.

## Two ways to run it

The same pipeline, with the steering wheel in a different place

| Leave it to the AI | One quoted price covers the whole run — cleanup, script, cast, storyboard, video and the finished cut. A progress page shows each stage, and you can cancel before the video stage starts. This is the default. |

| Take the wheel | You pay for the cleanup and the written script, then land in the normal editor with the ad already set up: rewrite the script, re-cast, edit any shot, regenerate a single segment and composite when you are happy. Every later step is billed like ordinary generation. |

Underneath, an ad is an ordinary SceneMixer project with the content type set to advertising — the same script analysis, the same reference art, the same shot list, the same per-segment rendering described on the short drama page. What is specific to ads is the creative brief, the shorter structure, the end card and the compliance filter.

## The music is generated with the picture, not mixed on afterwards

The video model used for ad segments produces sound and image in one pass, so it has already given you dialogue, room tone and, very often, a score that fits what is on screen. Laying a stock track over that gives you two pieces of music arguing. Instead, the brief carries a written description of the music — instrumentation, tempo, rhythm, where it should build — and the same description goes to every segment of the same ad, because each segment is an independent generation with no memory of the others. Matching descriptions are the only thing holding the score together across a cut.

The end card is silent by design, with the body audio fading into it. SceneMixer adds only three things after generation: the end card, a loudness pass, and the AIGC badge.

## What it will not do

- Make absolute claims. Superlatives and regulated phrasing — best in the world, number one in sales, state-approved, guaranteed returns — are stripped from the script automatically before anything is filmed, and from the tagline and end card separately. It is a filter on wording, not legal advice: the claims you make are yours.

- Put readable text in the frame. Video models cannot render reliable lettering, so the ad is written without on-screen text and your brand name is delivered where it will actually be legible — the end card. If a shop sign happens to appear in shot, treat whatever it says as decoration.

- Use a real person. People in your photos are removed rather than reproduced, and the cast is generated. There is no face swap and no voice cloning here.

- Edit footage you shot. Everything is generated from the storyboard. Uploads are references, not source clips.

- Run long. 60 seconds is the ceiling, built from four 15-second segments.

## What it costs

SceneMixer shows you a single figure before you commit, and it is computed from the same rates the rest of the platform bills against: reading the photo is free, the cleanup or redraw is one image, the script is one model call, reference art is billed per image with 15 free per account and 7 credits each after that, video is 10 credits per second of output on the 480P tier (≈ $0.06/s at the lowest credit price) and more on the sharper ones, and assembly is 1 credit per second. The quote is a ceiling rather than a deposit: stages are charged as they run, and a 15-second spot that needed fewer reference images than the estimate allowed for simply costs less. If a stage fails, the run stops there rather than spending the rest of the quote, and you can retry that stage without repeating the ones before it. New accounts start with 50 credits; packages start at $1.49. The pricing page has the underlying table.

## Ads in fifteen languages

On SceneMixer, the spoken lines, the end card and the casting all follow your interface language, and the market moves with it — names, faces and streets adapt to where the ad is meant to run. It is the same dialogue engine the series below were produced with, one per language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Brief and script | Upload one to four product photos and a one-line description: the photo is read for free, then the script, cast and storyboard are written | A chat model writes the spot; you describe the product and brand to every tool again |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### Do I need a script, a storyboard or a brief?

No. A photo is the minimum. A sentence about what you are selling improves the result and settles any ambiguity in the photo — “our shop” next to a picture of a counter is the difference between an ad about a venue and an ad about a coffee machine. SceneMixer writes everything downstream, including the spoken lines, and you can rewrite any of it in the advanced mode.

### Can I upload more than one photo?

Up to four, and they should be the same subject from different angles or in different detail — the first one is treated as the hero and is what any redraw and the end card are built from. SceneMixer turns extra photos of a venue into additional location plates, so the ad is set in your actual space rather than a generated approximation of it.

### What if there are people in my photo?

SceneMixer removes them before the photo is used as a reference, and the signage, fittings and lighting are kept. This is not squeamishness: reference images are copied almost literally by video models, so a customer in your storefront photo would otherwise end up performing in your advertisement.

### Will my brand name appear in the video?

On the end card, yes, along with the product name, the tagline and the call to action, in the language of the ad. Inside the generated footage, no — AI video models cannot spell reliably, so SceneMixer writes the ad without on-screen text rather than gambling on it.

### Which languages can the ad be in?

The spoken lines and the end card follow your interface language across fifteen dialogue languages, with the cast and setting adapted to that market. SceneMixer's end card falls back to a language it can typeset when a script needs shaping or right-to-left layout.

### Can I stop it halfway?

Yes, up until the video stage begins — that is the point where the expensive calls start, and it is where the cancel button stops working. If a stage fails, SceneMixer stops the run there, the credits for it come back, and you can retry that stage without repeating the ones before it.


---

# AI Shot List Generator: Script to Numbered Shots | SceneMixer

URL: https://scenemixer.com/ai-shot-list-generator/
Generate a numbered shot list from a script: shot size, duration, who is in frame, dialogue with delivery notes and ambient sound — editable, and rendered from the list.

AI shot list generator

# AI Shot List Generator

Updated November 3, 2026

Quick answer: SceneMixer turns a script into a numbered shot list and then renders from it. Each segment of 4–15 seconds (up to 30 on the long-take tiers) carries a header — location, time, weather, set dressing, props, opening frame — and inside it the individual shots, each with a duration, a shot size, an explicit list of who is in frame, the dialogue with delivery notes, and the ambient sound. The list is free to edit, and editing one segment re-renders only that segment.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A shot list you can render, not a document you then have to shoot

Most AI shot-list tools hand back a table. You still have to go and make every shot in it, in some other tool, and hope the result matches what the table said. Here the shot list is the input: each segment is generated from its own text, against the reference art for the characters and places that text names. Change a line in the list and only that segment is re-rendered. That one difference decides everything else on this page — the list has to be precise enough to be executed, not just readable.

## What a segment carries

The unit is a segment of 4–15 seconds (up to 30 on the long-take tiers), which is also one generation. Inside it sit the individual shots. Every segment opens with a header that fixes the facts a camera would need before rolling:

The header written for every segment

| Location | A reference to a location in the project's asset library, not a fresh description — so two segments in the same room are in the same room. |

| Time and weather | Stated once per segment. Lighting is derived from these plus the location's own plate, rather than being re-invented per shot. |

| Set dressing and props | What is standing in the space, and which props are in play. Props listed here are the ones whose reference art travels with the segment. |

| Opening frame state | The first frame, described as a still: shot size, who is in it, what posture they are in. Anything not visible in that first frame does not belong here. |

| Background people | Present only when the scene genuinely has a crowd. Declared once, in the header, so it does not flicker in and out between shots of the same place. |

## What a shot carries

- A number and a duration in seconds. The shot durations of a segment sum to the segment's length; there is no slack to hide in.

- A shot size. Wide, medium, close, over-the-shoulder, insert — chosen for what the moment needs, and biased towards framings AI video models actually render well.

- Who is in frame. Explicitly, per shot. This is the list that decides whose reference art gets attached, so it is treated as the authority rather than as prose.

- The action. Written as cause rather than result: what the light source is and where it is, not where the highlight lands.

- The dialogue. Each line with its speaker and a short delivery note. SceneMixer derives the length of a dialogue shot from the line — English is timed at three words per second — so a shot is never too short for the sentence inside it.

- Ambient sound. What the scene sounds like, which the video model generates along with the picture.

## Three rules that exist because of how video models fail

These are not style preferences. Each was added after watching generated footage break in the same way twice.

- Nobody off screen gets named. If a character is only being looked at, spoken to or reached for and is not in the frame, the shot says “towards the left of frame”, never their name. Naming an absent character is the single most reliable way to make a model paste their face into a shot that is supposed to have one person in it.

- Single-person shots say so. A shot with one character in it carries an explicit “only this person in frame” note, and characters who are present but not speaking are marked as such, so the model does not animate a mouth that should be closed.

- Depth words are camera decisions. Writing that someone is “in the background, out of focus” behind the speaker has already chosen a camera position — one that often does not exist for a normal two-person mid shot. The list is written so those phrasings only appear where the framing supports them.

After the model writes a list, a pass of deterministic corrections runs over it before it is saved: shot durations reconciled against the segment, the opening shot size matched to shot one, mouths closed on non-speakers, weapon and contact language made specific, leftover drafting notes removed. A second pass asks the same model to rewrite only the specific lines that failed a check, never the whole list.

## How long the lists are, in practice

Across projects on this platform the median episode is seven segments, and the median segment is fourteen seconds — so a typical episode's shot list is about a hundred seconds of screen time, carrying however many individual shots those seven segments need. On SceneMixer, generating one episode's list takes about 1 minute 54 seconds at the median, measured over every list that generated successfully on scenemixer.com in the thirty days to 20 September 2026.

## Editing is free; regenerating is not

Every line of the list is editable text, and editing costs nothing — the only billed events are model calls. That asymmetry is deliberate, because it changes how you work: fixing a shot by hand and re-rendering just that segment is cheap, while asking the model for a fresh list of the whole episode costs exactly what the first one did and throws away any hand edits in that episode. For small corrections, edit. For a genuinely different approach to the episode, regenerate.

## What it costs

Shot lists are billed per episode, by the length of the script that episode covers: 1 credit per 333 characters of English text, at the same rate for a first pass and a re-run. Script analysis, which produces the cast and the episode breakdown the list is written against, is 4 credits per 3,333 characters and runs once for the whole upload. On a first project the shot list for episode 1 is free, and every SceneMixer account starts with 50 credits. Rendering the shots is separate, from 10 credits per second of output. Full table on the pricing page.

## One shot list, fifteen dialogue languages

Dialogue timing is computed per language, so a shot is long enough for the line it holds whichever language that line is in. The dialogue language is chosen before analysis and performed natively rather than dubbed on afterwards; each series below was produced that way.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### Can I export the shot list as a PDF or a spreadsheet?

Not today. The list lives in the editor because that is where it is executed — it is the production document and the render input at the same time. You can copy the text of any segment, and the whole list is plain text rather than a proprietary structure.

### Can I write the shot list myself?

Yes. Every segment's text is editable, and you can insert, duplicate or delete segments. What SceneMixer supplies either way is the connection to the asset library: naming a character in a shot is what attaches their reference art to the render.

### Does it decide camera moves as well as shot sizes?

SceneMixer writes movement where movement serves the beat — a push in, a slow orbit, a handheld follow — and leaves it out where a locked frame is stronger. It deliberately avoids the elaborate moves that current video models turn into warping, and it does not use the technical-jargon vocabulary that image models from an earlier generation responded to.

### How does it decide how long a shot should be?

Dialogue shots are timed from the line itself at three words per second for English, with a floor so that very short lines still hold. Non-dialogue shots take their length from what has to happen in them. The shot durations of a segment are then reconciled to the segment's total before the list is saved, and a segment that overruns is split at a shot boundary rather than compressed.

### Can I get a shot list without generating video?

Yes. SceneMixer charges analysis and the shot list separately from rendering, and nothing obliges you to render. People do use the pipeline this way — a script in, a structured shot list out, the footage shot conventionally.


---

# AI Character Sheet Generator: Reference Art That Holds | SceneMixer

URL: https://scenemixer.com/ai-character-sheet-generator/
Generate character reference sheets built for reuse: full-body 9:16, plain background, directional light, identity first — then attached to every shot that uses them.

AI character sheet generator

# AI Character Sheet Generator

Updated November 3, 2026

Quick answer: SceneMixer generates each character once as a full-body reference sheet — 9:16, plain background, directional key light, identity features written before any technical instruction — and then attaches crops of that sheet to every shot the character appears in. Sheets are built as casting photos rather than artwork, because whatever is baked into a sheet is inherited by every scene that uses it. Every account gets 15 free.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A casting photo, not a poster

A character sheet here has one job: to be the picture every later shot is matched against. That makes it a reference image, and reference images are judged by different rules than artwork. Whatever is baked into the sheet is inherited by every shot that uses it — so a dramatic sidelight on the sheet becomes a dramatic sidelight in a scene set at noon, fighting the lighting the shot actually asked for. The sheet is therefore shot the way a casting photo is: full body, 9:16, plain background, one clear read of the face and the silhouette, and nothing competing for attention.

## Three constraints that came out of looking at bad sheets

What the sheet prompt does, and the failure each rule was written against

| Directional light, never flat | “Flat even lighting, no shadows” is the standard recipe for a plastic face — flat light erases skin texture. Sheets are lit with a keyed main source plus fill and a rim, and the prompt asks for the shadow under the nose, the transition across the cheekbone and a catchlight in the eye. |

| Identity first, technique after | The analysis step writes very specific faces — bone structure, a mole, a scar. Bury that in the middle of a long technical prompt and the image model dilutes it back towards a generically attractive face. The order is fixed: who this person is, then an instruction not to smooth those features away, then the technical block, then framing. |

| Face over wardrobe | The prompt states outright that this is an actor's casting photo and not a clothing catalogue, and forbids pulling the camera back to show off an outfit at the cost of head size. A neutral-expression instruction was removed for the same reason: it erases exactly the presence that makes a character recognisable. |

A fourth constraint is about length rather than content. The default image model attends less and less to a prompt the further down it reads, so the rules that matter most are placed immediately after the identity block and the whole prompt is kept short enough to stay inside the part of it the model actually reads. Adding a long passage near the top is not free: it pushes the skin and lighting rules out of the window.

## What gets a sheet, and what does not

- SceneMixer gives named characters one, generated once for the project and reused by every shot they appear in.

- Characters who are only ever heard — a voice on the phone, a narrator — are flagged during analysis and skipped, because art nobody will see is art you paid for.

- Non-human characters are drawn as a species plate rather than a person: real anatomy, four feet on the ground where the animal has four, no human hands, no clothing, a three-quarter front view. Anthropomorphising happens by default otherwise, and it is very hard to undo downstream.

- A second look — a full change of outfit, a transformation, burns, a uniform — is its own image, generated from the base sheet. When the description hands back two wardrobes in one field, they are separated automatically rather than drawn on top of each other.

- Your own art can replace any sheet. Upload it and the rest of the pipeline treats it exactly like a generated one.

## The description is the sheet

SceneMixer generates sheets from the character descriptions that script analysis produced, which means the cheapest moment to fix a face is before any image exists. Read the descriptions, correct them, then generate. Two habits pay for themselves: describe causes rather than results — the shape of the jaw, not “handsome” — and give each character one visual anchor that is impossible to confuse, a scar, an earring, a streak of grey. The same principle runs through the whole pipeline and is set out on the character consistency page.

Some result-language slips through from analysis anyway — phrasing about makeup that describes the outcome rather than the cause — and is rewritten before it reaches the image model, because an image model handed a result tends to paint it literally as a patch of colour.

## What it still gets wrong

- Long coats swallow the face. A full-body 9:16 frame of someone in a floor-length coat spends most of its pixels on fabric, and the head ends up too small to carry fine identity cues. It is a framing limit of the format, not a prompting failure.

- Two characters in the same uniform are still the hardest case for later shots, sheets or no sheets.

- Extras are not cast. Sheets cover the people in your asset library. SceneMixer generates background figures fresh each time, so treat them as scenery.

## What it costs, and how fast

Sheets are billed as images, separately from script analysis: every account gets 15 free reference images — characters, scenes or props, in any mix, across any project — and images after that are 7 credits each on the default image model. Regenerating a sheet costs the same as the first one. Generation is quick: a reference image takes about 16 seconds at the median and 9 in 10 finish inside 27 seconds, measured across every image that generated successfully on scenemixer.com in the thirty days to 20 September 2026 — so a typical five-character cast is a couple of minutes, and the free allowance covers a first cast with room to spare. The pricing page has the rest.

## One cast, fifteen languages

A sheet is language-agnostic — it is a picture. What changes with the dialogue language is the casting brief, so a series set for a different market is cast for that market rather than dubbed into it. One finished series per language below.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### How many angles does a sheet have?

One full-body pose, plus face and body crops taken from it that travel with each shot. SceneMixer's sheet is not a multi-angle turnaround grid. The reason is practical: what later shots consume is a clear, well-lit read of the face and silhouette, and a grid of small panels gives a video model less of that per pixel, not more.

### Can I use my own character art instead?

Yes — upload an image in place of any generated sheet and SceneMixer treats it identically. That is the route for a character you have already designed elsewhere, or for a look you could not get the model to produce.

### Does it work for animals and creatures?

Yes, and they are drawn differently on purpose: a species plate with real anatomy rather than a person in animal costume. Mark the character as non-human during analysis and SceneMixer changes the whole prompt — no human hands, no clothing, no upright stance unless the species has one.

### Why is the background plain?

Because anything else gets inherited. A sheet shot in a location bakes that location's light and colour into every scene the character later appears in. Plain background and neutral, directional lighting keep the sheet usable everywhere, which is the entire point of generating it once.

### What happens if two characters have the same name?

Nothing breaks. In SceneMixer, shots reference assets by identifier rather than by name, so duplicates and mid-project renames do not silently attach the wrong sheet. New assets are also prevented from taking a name already in use within the project.

### Can I generate a sheet without a script?

Sheets are generated inside a project on SceneMixer, from the cast that analysis found, because their purpose is to be reused by that project's shots. You can add a character by hand and give it a description, which is the closest thing to a standalone generator — but a sheet with no shots to anchor is just an image.


---

# AI Movie Generator: Feature Scripts, Shot by Shot | SceneMixer

URL: https://scenemixer.com/ai-movie-generator/
Generate a film from a screenplay: film-mode pacing with no forced hooks, scripts up to about 50,000 words, a cast that holds across sequences, honest per-second costs.

AI movie generator

# AI Movie Generator

Updated November 3, 2026

Quick answer: set the content type to film and SceneMixer reads a screenplay up to roughly 50,000 words, divides it into sequences without the short-form length cap or the forced end-of-episode hook, allows subtext and quiet endings in the dialogue, then generates it sequence by sequence against a cast that was drawn once. Runtime is built by stacking 4–15 second segments, so a scene can be re-cut one shot at a time — and every finished second is a paid generation, which is worth doing the arithmetic on before you start.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What changes when you set the type to film

Content type is one of the four settings you pick before analysis runs, and it is not cosmetic — it rewrites the rules the analysis and the shot list are written under. Short drama assumes a phone, a viewer with a thumb on the scroll bar, and a hook every ninety seconds. Film assumes someone who has already decided to sit down.

The three content types and what each one changes

| Type | Length of a unit | Pacing and dialogue |

| Short drama | Up to 240 seconds | Conflict inside the first ten seconds, dense dialogue, an explicit hook to close every unit — identity reveal, reversal, unresolved beat. |

| TV series | Up to 300 seconds | Fifteen to twenty seconds of establishment before the conflict; a closing beat that can be an emotional suspension or an unfinished decision rather than a twist. |

| Film | No hard cap | Length follows the dialogue and the atmosphere shots it needs; subtext, silence and half-spoken lines are allowed; suspense can be a shot of a door rather than a cliffhanger. |

The practical difference shows up first in dialogue. Short-form grammar forces characters to say what they mean, because a scrolling viewer cannot be asked to infer. Film mode switches that off: a line can land sideways, a scene can end on a look. Set it before you run analysis — changing it later needs a re-parse to take full effect, since it is the analysis pass that carries the rule.

## Length is stacked, not requested

Every generated segment in SceneMixer is 4–15 seconds, with some model tiers reaching 30. Runtime comes from putting those segments end to end, not from asking for a long take, and that constraint shapes what this tool is good at. It means a scene can be re-cut a shot at a time for the price of one shot. It also means a feature is a very large number of segments, each of which is a paid generation.

The script is still divided into units — in film mode they are sequences rather than episodes, without the length cap and without the forced hook — and every stage after analysis is scoped and billed to one unit. You can finish and watch sequence one before deciding whether sequence two is worth generating.

## How long a script it can read

SceneMixer reads up to roughly 300,000 characters of English in a single project, about 50,000 words — comfortably a feature screenplay, and a short novel besides. Past about 33,000 characters the analysis switches to a two-call path: one call for the cast, locations and props, a second for the structural skeleton, with each unit's full outline expanded when you open it. The reason is unglamorous — a single call has an output budget, and a script that exceeds it comes back truncated with no error raised. A fifty-unit script returning as ten units, silently, is the failure that path exists to prevent. Longer than the ceiling, split the script by act and run the parts as separate projects.

## Be honest about the bill before you start

This is the part most tools in this category leave vague, so here is the arithmetic. SceneMixer charges video per second of finished output: 10 credits a second on the cheapest tier (≈ $0.06/s at the lowest credit price) and up to $0.47/s on the sharpest, plus 1 credit a second to assemble. Analysis and shot lists are rounding errors next to that.

So: a ten-minute short film is six hundred seconds of that rate. A ninety-minute feature is nine times more, and a regenerated take is charged again. At the cheapest tier a short film is an affordable weekend; a full feature at the sharpest tier is a real production budget, just a much smaller one than a camera crew. The honest recommendation is to treat this as a tool for short films, proofs of concept, pilot sequences and previsualisation first — and to generate a feature only once you have watched a sequence of it and know the look holds.

## What it will not do

- One continuous take of any length. The cap is per generation, per model tier. Long takes are the one film grammar this pipeline cannot give you.

- Edit or grade footage you shot. Everything is generated from the shot list. SceneMixer has no timeline for imported clips.

- Cast a real person. The cast is generated, or uploaded by you; voices come from a preset library or your own sample. No face swap, no voice cloning.

- Crowd scenes and long fight choreography. The shot list deliberately steers around them because current video models fail at them; background figures are scenery, not cast.

For what it does instead — a cast that survives ninety minutes of screen time, a shot list you can edit line by line, and a segment you can re-render on its own — see consistent characters and the shot list.

## Fifteen dialogue languages

The spoken language is set once, before analysis, and performed natively by the video model — no dubbing pass, and no change in the per-second price. Each series below came out of the same SceneMixer pipeline in a different language.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### Can it really make a feature-length film?

SceneMixer can read a feature screenplay and generate it sequence by sequence, and nothing in the pipeline stops you at ninety minutes. Whether you should is a budget question rather than a capability one: every finished second is a paid generation, so a feature costs roughly nine times a ten-minute short at the same tier. Most people here are making shorts, pilots and previsualisation, and that is what we would suggest starting with.

### What is the difference between film mode and short drama mode?

Length caps, dialogue grammar and the way a unit ends. Short drama forces explicit dialogue, a conflict in the first ten seconds and a hook at the close; film mode allows subtext and quiet endings and drops the per-unit cap. It is set before analysis, and changing it afterwards requires re-running analysis for it to fully apply.

### Can I generate one scene rather than the whole script?

Yes. SceneMixer analyses the script once, and everything after that is requested per unit, in any order. Generating sequence four without sequences one to three is normal practice — it is how you test whether the look holds before committing.

### Does the cast stay consistent over ninety minutes?

The mechanism does not care about runtime. SceneMixer generates each character once as a reference sheet and attaches crops of it to every shot they appear in, so shot 900 is matched against the same picture as shot 3. The hard cases are unchanged by length: two characters in matching costume sharing a frame, and characters whose sheet is mostly coat.

### What aspect ratio and resolution?

16:9 by default, 9:16 available, set once per project. On SceneMixer, resolution follows the model tier you pick per unit, from 480P up to 1080P, and the per-second price moves with it — which is why it is worth generating a test sequence at the cheapest tier before committing a whole film to the sharpest one.

### Can I upload a Final Draft script?

Export it to plain text first. SceneMixer accepts .txt, .docx and .md uploads, or paste directly. Screenplay layout is read as well as prose, so a text export keeps its structure.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Dialogue Generator: Lines Your Cast Actually Speaks | SceneMixer

URL: https://scenemixer.com/ai-dialogue-generator/
Generate spoken dialogue for an AI drama: lines per character with delivery notes, timed to real speech rate, performed natively in 15 languages — not subtitles.

AI dialogue generator

# AI Dialogue Generator

Updated November 3, 2026

Quick answer: SceneMixer writes dialogue into the script analysis, assigns each line to a shot with a speaker and a short delivery note, and the video model performs it — picture and audio generated together, so the mouth forms that sentence in the language you chose. Shot lengths are computed from the line at three words per second for English, so a shot is never too short for the sentence inside it. Editing a line is free.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Dialogue is not a caption track

Most tools that say “AI dialogue” mean subtitles: text you paste over a clip. Here the line is spoken. The video model generates picture and audio in one pass, so the character's mouth is forming that sentence, in that language, with that pacing — there is no separate voice file glued on afterwards and no lip-sync pass to go wrong.

That has one consequence worth understanding before you write anything: a line's length decides a shot's length. Say it out loud and time it, because the pipeline does exactly that.

## At a glance

What the dialogue layer produces and what it costs you to change

| Written by | Script analysis writes the lines into each episode's outline; the shot list then assigns them to specific shots with a speaker and a delivery note. |

| Language | Chosen once per project, before analysis. Fifteen to pick from. The cast performs in that language; the interface and working documents follow your own language independently. |

| Delivery notes | A short parenthetical per line — how it is said, not what is said. It reaches the video model as performance direction. |

| Off-screen lines | Marked explicitly: voice-over, narration, inner voice, a voice on the phone. Marked lines do not open anyone's mouth on screen. |

| Shot length | Derived from the line. English is timed at three words per second; Chinese at five characters per second. A short line still gets a floor so the shot does not feel clipped. |

| Voices | Matched from a preset voice library, or you upload a sample for a character. No voice cloning of anyone you do not have. |

| Editing | Free. Rewrite any line in the shot list; only the segments you touched re-render. |

## Why the timing rule matters more than it sounds

A line of eighteen English words needs about six seconds of screen time. Ask a video model for a four-second shot and give it eighteen words and one of two things happens: the delivery is rushed into gibberish, or the model quietly drops the back half of the sentence. Neither failure announces itself — you only notice when you watch the cut.

So SceneMixer computes the floor from the line itself and writes that number into the shot. If a segment's shots add up to more than the segment can hold, the segment is split at a shot boundary rather than compressed. You can override any duration by hand; the arithmetic is there so the default is never wrong in the direction that ruins a take.

## Delivery notes: direction, not stage business

Each line carries a short note in brackets — the kind a director says across a table, not a novelist's adverb. “Flat, already decided.” “Trying not to look at the door.” These reach the model as performance instruction and change the read noticeably. Two rules the pipeline enforces so they stay useful:

- The note never repeats the line. A note that paraphrases the sentence adds nothing and eats the model's attention.

- Off-screen markers come first. If the line is a voice-over, a voice on the phone or an inner thought, that is stated before any other direction — because it decides whether a mouth moves at all, and getting it wrong is the difference between a phone call and a character talking to themselves on camera.

When the model returns lines without notes — which it does unpredictably — SceneMixer runs a separate small pass that fills only the missing ones, without touching the lines themselves.

## One project, one dialogue language

The dialogue language is a project setting, not a per-line one, and it is chosen before analysis because it changes what gets written, not just how it is rendered. A Chinese web novel set to English dialogue produces English lines with English rhythm — not translated Chinese sentences. Names and settings move with it if you also set a target region.

What does not change is the price: a segment costs the same credits in any language, and the working documents you read are translated for free. The detail is in the dialogue-language guide; native dialogue vs dubbing covers why this is not a dubbing pass.

## What it will not do

- Clone a voice. Voices come from a preset library or your own uploaded sample. SceneMixer has no path for reproducing someone else's voice.

- Lip-sync footage you upload. Speech is generated with the picture. This is not a tool for putting new words in an existing clip's mouth.

- Draw text inside the generated picture. Spoken dialogue is audio; subtitles, if you want them, go over the finished video (on accounts that have bought credits) or come as an SRT file.

- Hold a four-person conversation in one shot. Two-handers are reliable; past that, coverage is cut into shots, which is what a real edit would do anyway.

## What it costs

Dialogue is not a separate line item — it is written by the two paid stages you already run. Script analysis is 4 credits per 3,333 characters of English text; each episode's shot list is 1 credit per 333 characters of that episode. Rewriting a line is free; re-rendering the segment that contains it costs the segment's video, from 10 credits per second (≈ $0.06/s at the lowest credit price). Voice matching costs nothing. New accounts on SceneMixer start with 50 credits and one free preview of up to 5 seconds — enough to hear a line performed before paying for anything. Full table on the pricing page.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### Can I write the dialogue myself instead of having it generated?

Yes. SceneMixer's shot list is editable text — replace any line with your own and only that segment re-renders. People also paste a finished screenplay, in which case analysis carries the existing lines through rather than inventing new ones.

### How do I control how a line is said?

With the delivery note in brackets next to it. Keep it short and about manner rather than meaning: “quiet, not asking” changes the read; “she says she is leaving” does not, because the line already says that.

### Will the same character sound the same across episodes?

Yes — the voice is a property of the character in the project, matched once and reused, the same way the face is. If you upload your own sample for a character, that upload is a persistent choice and SceneMixer never overwrites it with later matching.

### What happens to inner monologue from a novel?

SceneMixer turns it into a marked line — inner voice or narration — so it is heard while the picture shows something worth watching, rather than becoming a shot of somebody thinking.

### Can the interface be in one language and the dialogue in another?

Yes, and that is the common case: SceneMixer's interface and working documents follow your own language, the spoken lines follow the project setting. An English-speaking producer can ship a Korean-language series and still read every document in English.


---

# AI Script Breakdown: Cast, Locations, Props, Episodes | SceneMixer

URL: https://scenemixer.com/ai-script-breakdown/
Turn a script into a production breakdown: named cast with drawable descriptions, locations, props and a numbered episode structure, each entry quoting its source line.

AI script breakdown

# AI Script Breakdown

Updated November 3, 2026

Quick answer: one pass over the whole script returns the document a production is built from — named characters with descriptions specific enough to draw, the locations the script actually uses, the props the plot touches, and a numbered episode structure where each entry carries a conflict arc, an emotional peak, a closing hook and a verbatim quote of the source text it came from. Everything downstream points at these entries, so this is the cheapest stage and the most expensive one to get wrong.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a breakdown is for

Before anyone shoots anything, somebody reads the script and writes down what it will take: who is in it, where it happens, what has to be on the table, and how it divides into shootable units. That document is a breakdown, and on a conventional production it is what the schedule and the budget are built from.

Generated video needs the same document for a different reason. Every shot is rendered independently, so anything that has to stay the same across shots has to exist as a shared object first — one that later shots point at instead of re-describing. In SceneMixer, the breakdown is where those objects are created.

## What one pass produces

One analysis pass over the whole script

| Characters | A named entry each, with appearance written specifically enough to draw from — face shape, build, a distinguishing visual anchor — plus wardrobe and a voice description. Median project: five named characters. |

| Locations | Every place the script actually uses, described as a space rather than a mood: depth layers, materials, where the light comes from. |

| Props | Objects the plot touches — the letter, the knife, the phone. Set dressing that nobody handles stays in the location description. |

| Episode structure | A numbered list. Each entry gets a conflict arc, an emotional high point and a closing hook — not a summary of events. |

| Summary | Title, logline, era, region, one-line premise. Editable before anything is drawn. |

| Anchors | Each episode entry quotes the source text it came from, verbatim. Used to check the breakdown against the script rather than trusting it. |

## Long scripts are read in two passes, on purpose

A single model call has an output budget. When a script exceeds it, the model does not raise an error — it simply stops early and returns fewer episodes than the material contains. In our own offline evaluation a fifty-episode script came back as ten, with nothing in the response to indicate the other forty were dropped. That is the worst kind of failure: a plausible, complete-looking document that is quietly wrong.

So past roughly 10,000 equivalent characters the breakdown runs as two calls — one returns the cast, locations and props; the second returns the episode skeleton with its anchors. Each episode's full outline is then expanded when you open that episode, so a sixty-episode upload does not pay for sixty expansions you may never read. The ceiling for the two-pass path is about 300,000 characters of English, roughly 50,000 words.

## Read it before you spend anything

The breakdown is the cheapest thing in the pipeline and the most expensive thing to get wrong, because everything downstream points at it. A face described vaguely produces a generic reference sheet, and that sheet is then attached to forty shots. Fifteen minutes spent here is worth more than any amount of regenerating later.

- Check the episode count against the script. Each entry quotes its source line — if the last quoted line is halfway through your file, the breakdown is short.

- Fix character descriptions now, while they are still text. Add the anchor a stranger would use to pick that person out of a crowd.

- Delete props that are set dressing. In SceneMixer, anything left in the prop list gets reference art and gets attached to shots.

- Check who is missing. A character who speaks once and is never touched or named again may not have made the cast list; add them if a later shot needs their face.

## What it will not do

- Schedule or budget a live shoot. There are no day-out-of-days, call sheets or crew counts here. This is a breakdown for generated video.

- Invent plot. Scene order, dialogue and beats follow the script. What the pass adds is structure and specificity, not story.

- Read a PDF or .fdx directly. Plain text, .docx and .md, or paste. Screenplay layout is understood; export first.

- Guarantee every prop. The pass is good at objects the plot handles and unreliable at objects mentioned once in passing. Check the list.

## What it costs

SceneMixer charges analysis once for the whole upload, by length: 4 credits per 3,333 characters of English text. A 30,000-character script is therefore about 36 credits — the cheapest stage by an order of magnitude, and the one everything else depends on. Reference art is billed separately per image, with 15 free per account and 7 credits each after that; shot lists are 1 credit per 333 characters per episode. Timing: reading a whole script takes about 2 minutes 49 seconds at the median, measured across every analysis that completed successfully on scenemixer.com in the thirty days to 20 September 2026. Full table on the pricing page.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### How is this different from asking a chatbot to break down my script?

SceneMixer wires its output into production. Characters, locations and props become objects the rest of the pipeline points at by identifier — the reference art is generated from these entries, and every later shot attaches the art of the entities it names. A chat transcript is a document; this is the project's spine.

### Can I edit the breakdown?

Most of it, and you should before generating art. Names and descriptions of characters, locations and props are editable text; episode outlines are not, so you change an episode in its storyboard. Editing costs nothing in SceneMixer.

### What if it splits my script into the wrong number of episodes?

Check the source quote on the last entry first — if it stops early, the script exceeded a single pass and you want the two-pass path, which SceneMixer triggers automatically past about 33,000 English characters. If the count is simply not what you wanted, set the content type in Step 1 (short drama, series or film) — that changes the length cap each unit is paced for.

### Does it work on a novel rather than a screenplay?

Yes, and that is the harder input it was built for: prose has inner monologue, time skips and description that a camera cannot photograph. Novel to video covers what happens to each of those.

### How long can the script be?

SceneMixer takes about 300,000 characters of English — roughly 50,000 words — in one project. Past that, split by act or volume and run them as separate projects; the cast can be carried across.


---

# Vertical AI Video Generator: 9:16 Built In, Not Cropped | SceneMixer

URL: https://scenemixer.com/vertical-video-ai-generator/
Generate 9:16 vertical drama end to end: the frame is set before any art, so sheets, plates and every shot are composed tall — faces high, no letterboxed landscape.

Vertical AI video generator

# Vertical AI Video Generator

Updated November 3, 2026

Quick answer: pick 9:16 in Step 1 and the whole project is composed for a tall frame — location plates, prop art and every rendered segment, not a landscape shot letterboxed after the fact. The shot list changes with it: faces sit high, two-handers become over-the-shoulder singles, and the camera never pulls back far enough to shrink a head for the sake of showing wardrobe. Vertical costs the same as landscape.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Vertical is a format, not a crop

You can letterbox a landscape shot into a 9:16 frame and call it vertical. It will look like exactly that. A frame three times taller than it is wide changes where faces sit, how much room a two-hander needs, how far the camera can pull back before a head becomes a dot — and generated video is unusually sensitive to all three, because every shot is composed from scratch rather than cropped from coverage.

So the aspect ratio is a SceneMixer project setting chosen before any art is generated, and it propagates: character sheets, location plates, prop shots and every rendered segment are all composed for the frame you picked.

## At a glance

What changes when you pick 9:16

| Where it is set | Step 1, once per project. 16:9 landscape is the default; 9:16 vertical is the other option. |

| Character sheets | Always 9:16, in both projects — a full-body reference is taller than it is wide regardless of the delivery frame. |

| Locations and props | Follow the project frame, so a vertical project gets vertical plates that later shots are composed against. |

| Segment length | 4–15 seconds; some model tiers reach 30. Runtime is stacked, not requested. |

| Episode length | Short drama is paced for up to 240 seconds per episode. Vertical series usually run 60–120. |

| Resolution | 480P by default, 720P and 1080P a click away. The per-second price moves with the tier, nothing else does. |

| Text in frame | None by design. Video models cannot spell reliably, so captions are added after export on whatever platform you publish to. |

## What the shot list does differently in a tall frame

- It keeps faces high. In a vertical frame the eye lands in the upper third; a composition centred for 16:9 puts the head in the middle and wastes the top.

- It avoids wide two-shots. Two people side by side in 9:16 either both become small or one leaves the frame. The list uses over-the-shoulder and alternating singles instead — which is what vertical drama does on real sets too.

- It does not pull back to show wardrobe. A full-length figure in a tall frame shrinks the head until identity is gone. Clothing is established once on the reference sheet, not re-established in every shot.

- It states who is in frame, per shot. In a narrow frame the difference between “two people” and “one person and a voice” is most of the composition, so it is written explicitly rather than implied.

## Changing your mind later

In SceneMixer, the frame is not locked forever, but it is not free to change either: the reference art was composed for the old ratio. Switching a project from landscape to vertical leaves you with location plates framed for the wrong shape — usable, but they will fight the new compositions. If you know the series is for phones, pick 9:16 before generating art; if you are unsure, generate one segment in each and look, which is what the free allowance is for.

## What it will not do

- Reframe an existing landscape video. There is no crop-and-track tool here. Vertical output comes from generating in a vertical frame.

- Add a safe-zone overlay. Platform UI overlays differ and change; the pipeline keeps the subject clear of the extreme edges but does not render a template.

- Produce one continuous vertical take of any length. Segment caps are per model tier. Runtime comes from stacking.

- Burn in captions or a hook card. Those go on after export, where you can change them per platform without re-rendering.

## What it costs

On SceneMixer, vertical costs exactly what landscape costs — the frame is not a surcharge. Video starts at 10 credits per second of output on the 480P tier (≈ $0.06/s at the lowest credit price), up to $0.47/s on the sharpest; assembly is 1 credit per second. A 90-second vertical episode is therefore about 990 credits of video and assembly at the cheapest tier. Analysis and shot lists are billed by text length and are the same either way. New accounts get 50 credits, 15 free reference images and one free preview of up to 5 seconds. The micro drama page covers the format's economics; full rates on the pricing page.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### Is 9:16 the default?

No — in SceneMixer, 16:9 landscape is the default, and 9:16 is one setting away in Step 1. It is worth setting deliberately before any reference art is generated, because the art is composed for the frame.

### Can I get both a vertical and a landscape cut of the same series?

Not from one render. SceneMixer generates each segment in the project's frame; producing the other shape means generating those segments again in a project set to that ratio. There is no reframing pass.

### How long should a vertical episode be?

Sixty to a hundred and twenty seconds is the working range for short drama, and that is what the outline paces for. A 90-second episode is typically seven segments of about fourteen seconds stacked together.

### What resolution should I publish at?

480P is the default and is genuinely fine on a phone; 720P is the usual upgrade when the series is working. The per-second price is the only thing that changes — generate one segment at each and compare before committing an episode.

### Where do captions and hook text go?

On after export. SceneMixer renders no text inside the generated picture, because current video models cannot produce reliable lettering. Accounts that have bought credits can add dialogue subtitles over the finished video or download an SRT file; hook text goes on at the platform, where you can change it without re-rendering a shot.


---

# AI Hook Generator for Short Drama: Endings That Earn Episode Two

URL: https://scenemixer.com/ai-hook-generator/
Hooks are a structural constraint, not caption copy: every episode ends unresolved, and the kind of unresolved beat rotates so a series never runs the same trick twice.

AI hook generator

# AI Hook Generator

Updated November 3, 2026

Quick answer: retention is decided by where an episode stops, so the hook is built into the episode breakdown rather than bolted on as text. Every episode is required to end on an unresolved beat, and the kind rotates — identity exposed, feeling suspended, understanding overturned — so eight episodes do not run the same trick eight times. Opening shots start on the problem, and no text is ever burned into the frame.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A hook is a structural decision, not a caption

Most “hook generators” write you a line of text to paste over the first frame. That is the last five percent of the job. The thing that actually decides whether someone watches episode two is where the episode stops — and that is a decision made while the script is being divided into episodes, long before any video exists.

So hooks here are not a separate tool. They are a constraint the episode breakdown has to satisfy: every episode ends on an unresolved beat, and the kind of unresolved beat rotates so the series does not feel like the same trick eight times.

## The three kinds, and why they rotate

The end-of-episode beats the breakdown cycles through

| Kind | What is left open | Reads as |

| Identity exposed | Someone now knows something about who another person really is — but we cut before we see what they do with it. | Dread. Works best mid-season. |

| Feeling suspended | A confession, an accusation or an offer lands, and the reply is the next episode's first line. | Ache. Works anywhere, wears out fastest. |

| Understanding overturned | The audience learns the situation was not what they were told; the characters have not caught up. | Vertigo. Best saved for act breaks. |

Rotation matters more than any single hook's strength. Three “feeling suspended” endings in a row train the viewer to expect a non-answer, and the fourth one stops working. SceneMixer's breakdown is told to alternate for exactly this reason.

## The opening seconds, done as craft rather than text

The first shot of an episode carries the same weight as the last, and the shot list treats it as a distinct problem: it establishes the frame, who is in it, and what is wrong — in that order, in one shot, without a line of exposition. Concretely, what the list is made to avoid:

- Establishing shots that establish nothing. A wide of a building tells the audience where they are and nothing about why they should stay. The opening shot starts on the problem.

- Text on screen. No burned-in hook copy, ever — not a stylistic choice but a technical one: video models cannot spell reliably. You add caption text after export, per platform, without re-rendering.

- A cold open that contradicts the previous ending. Each episode's opening state has to be consistent with where the last one stopped; when a long segment is split, the second half carries an explicit continuation marker so the first frame is written as a real static frame rather than copied from the wrong moment.

## Where you can see and change it

SceneMixer's episode list after analysis gives you each episode's outline — three to five sentences including how it ends. Outlines themselves cannot be edited on the site, so the cheapest place to change a hook is the episode's storyboard: it is text, editing it is free, and nothing has been rendered yet.

Once a shot list exists you can still edit the script text directly, and the editor keeps the machine-readable structure in sync as you type. What you cannot do is bolt a hook onto finished video — that means regenerating the segment, which costs what a segment costs.

## What it will not do

- Write you ten interchangeable hook lines. SceneMixer has no hook-copy generator, because copy is not where retention comes from and the platform is where copy belongs.

- Promise a retention number. Nobody can. The structure is the part that is under your control; the performance is not.

- Burn text into the frame. By design — see above.

- Restructure an episode you have already shot. Moving the break point means re-rendering the segments after it.

## What it costs

Hooks cost nothing extra: the episode breakdown that produces them is part of the analysis pass, billed by script length at 4 credits per thousand characters (3,333 characters per unit for alphabetic scripts). Per-episode shot lists are 1 credit per hundred characters of that episode's share, and editing the storyboard text is free. Only rendering costs money: from 10 credits per second of output. New accounts on SceneMixer get 50 credits and one free preview of up to 5 seconds. See the episode generator page for how episodes are divided, and the pricing page for full rates.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### Will it write the text I put on the first frame?

No, and deliberately. SceneMixer never draws hook text into the video — video models cannot produce reliable lettering — so caption and hook copy go on at the platform, where you can change them per post without re-rendering a shot.

### Can I choose which kind of ending an episode gets?

Yes, at the storyboard. Outlines cannot be edited on the site, but the episode's shot list is plain text and editing it is free, so rewrite the closing shots to the ending you want before rendering.

### Does every episode really need a cliffhanger?

For short vertical drama, effectively yes — that is the format's engine. For a feature-length project the breakdown paces differently, because a film does not need to reacquire its audience every ninety seconds.

### What if the hook it writes is weak?

Rewrite the closing shots in that episode's storyboard. It is the cheapest correction available: text in, text out, no rendering involved.

### Can I add a hook to an episode I have already rendered?

You can change the script, but the segments after the change have to be rendered again — finished video is not editable text. This is why the outline stage is the place to get endings right.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Episode Outline Generator: Structure Before You Render

URL: https://scenemixer.com/ai-episode-outline-generator/
Episode outlines are the instruction shot lists obey: conflict, emotional peak, unresolved ending and a source anchor checked against your script — free to rewrite.

AI episode outline generator

# AI Episode Outline Generator

Updated November 3, 2026

Quick answer: each episode gets a three-to-five-sentence outline that the shot list is generated from — so editing an outline changes the shots, and editing text is free. Every outline must carry a conflict, an implied emotional peak, an unresolved ending, and a fragment of your original script quoted character for character and checked against it. Long scripts get the full outline list first and shot-level detail per episode, because a single pass silently truncates.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The outline is the contract everything downstream obeys

An episode outline here is three to five sentences, and it is not a summary written after the fact — it is the instruction the shot list is generated from. Which makes it the most valuable text in the project. It cannot be edited on the site; what you can rewrite for free is the shot list generated from it, before anything is rendered.

Each outline has to carry four things, and the breakdown is checked against them rather than trusted to volunteer them.

## At a glance

What an episode outline is required to contain

| A conflict | Not a list of events. “She goes to the office, then to dinner” is a schedule; an outline has something at stake in it. |

| An emotional peak | Implied, not stated — the outline has to make clear which moment the episode is built toward. |

| An unresolved ending | For short drama, every episode. The kind rotates so the series does not repeat itself; see the hook page. |

| A verbatim source anchor | A quoted fragment of your original script, copied character for character, so you can see exactly which passage became this episode. |

The anchor is the part people do not expect and end up relying on most. It is checked mechanically: the quoted fragment has to actually appear in your text. If a model invented an episode out of nothing, the anchor is where that shows.

## How many episodes, and who decides

SceneMixer's breakdown decides, from the material — chapter and scene boundaries in the source, where conflicts resolve, and how much runtime a beat can carry. A novel with long interior chapters yields fewer, denser episodes than a dialogue-driven web serial of the same length. Short drama paces for up to 240 seconds per episode; most vertical series settle at 60–120.

What the breakdown will not do is pad. If your script contains four episodes of material, asking for twelve gets you twelve thin ones, so it does not try.

## Long scripts get outlines first, detail on demand

Above roughly ten thousand characters of equivalent length the analysis runs in two passes, and the reason is worth knowing because it explains a failure you may have seen elsewhere: a single pass is limited by how much it can write, not how much it can read. Handed a fifty-episode script, one pass will quietly return ten episodes and no error — it silently drops the rest.

So the second pass produces the episode skeleton only: outline, anchor, no shot-level detail. Full detail is expanded per episode, in the background for the first few and on demand for the rest — open an episode you have not touched and it is expanded then. You get a complete, honest episode list for a long script in one sitting instead of a truncated one that looks complete.

## Changing an episode after the read

- Outlines are read-only; shots are not. SceneMixer does not let you edit an outline on the site. The shot list for an episode is generated from it when you ask, and every shot in it is plain text you can rewrite for free.

- Regenerating a shot list replaces it. Segment-level manual choices for that episode are recalculated. Episode-level settings are not affected.

- Regenerating costs the same as the first time. SceneMixer bills per-episode shot lists by that episode's share of the script, first pass or fifth.

## What it will not do

- Hit an episode count you name. The count comes from the material. You can merge or cut after the fact.

- Invent a season out of a logline. Analysis reads a script; writing one from a premise is a separate pass with its own price.

- Keep your manual per-shot edits through a regeneration. Regenerating a shot list rebuilds it.

- Reorder episodes. Structure follows the source; reordering means editing the source and analysing again.

## What it costs

Outlines come out of the analysis pass — 4 credits per thousand characters, with 3,333 characters counted as one unit for alphabetic scripts, so a 40,000-character English script is about 36 credits. Per-episode shot lists are 1 credit per hundred characters of that episode's share. Your first SceneMixer project gets episode one's shot list free, plus 15 free reference images. Editing text never costs anything. Full rates on the pricing page; see script breakdown for what else the analysis pass produces.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### Can I tell it how many episodes I want?

Not as a target. SceneMixer takes the count from where the material actually breaks — conflicts resolving, scenes closing, how much runtime a beat can carry. You can merge or drop episodes afterwards; padding a thin script into more episodes only produces thinner ones.

### What is the source anchor for?

Traceability. Each outline quotes a fragment of your original text character for character, and SceneMixer checks that quote against your script mechanically — so you can see which passage became which episode, and an invented episode has nowhere to hide.

### Why does my long script only show outlines at first?

Because that is the honest way to handle length. A single analysis pass is bounded by output, not input, and given a very long script it silently returns a fraction of the episodes. On SceneMixer, skeleton-first means you get the whole list, with detail filled in per episode as you open them.

### Does regenerating an episode's shot list cost again?

Yes, at the same rate as the first time — it is the same amount of work. Editing the shot list's text is free, which is why it is worth fixing an episode in its storyboard before rendering video.

### Can I write outlines myself instead?

Not on the site: outlines come from analysis and cannot be edited. Write the episode the way you want it into your script before the read, or rewrite its shots in the storyboard afterwards. SceneMixer does not let you skip analysis entirely, because the same pass also produces the character, location and prop lists the shot lists reference.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Multi-Character Scene Generator: Only Who Is Actually There

URL: https://scenemixer.com/ai-multi-character-scene-generator/
Every shot states who is in frame, and that list alone attaches faces — characters merely mentioned become directions, and background crowds are declared or absent.

AI multi-character scenes

# AI Multi-Character Scene Generator

Updated November 3, 2026

Quick answer: each shot carries an explicit list of who is in frame, and that list is the only thing that attaches character art — so someone merely looked at or spoken to becomes a direction rather than a name, and never arrives on screen with an invented face. Only the named speaker's mouth moves. Background crowds are a declared state of the scene or they are absent. Past four named people in one frame, the list covers the scene in singles instead.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The hard part is who is in frame, not how many

Generated video has a specific failure with more than one person in a shot: it will put in a face that should not be there. Someone mentioned in the text — looked at, shouted to, thought about — arrives on screen, wearing a face invented on the spot, because nothing told the model the difference between “present” and “referred to”.

So the shot list states it per shot. Each shot carries an explicit list of which characters are in frame, and that list — not the prose, not the previous shot's ending state — is the only thing that decides whose reference art gets attached.

## At a glance

How a multi-character shot is specified

| Who is in frame | A per-shot list. The union across a segment's shots decides which reference sheets are sent with it. |

| Who is off screen | Written as a direction — “towards frame left”, “eyeline to camera” — not as a name. No face is attached. |

| Single-person shots | Marked explicitly, so the renderer is told the frame holds one person and not to invent company. |

| Who is speaking | Named per line; everyone else in frame is marked lips closed, so only the speaker's mouth moves. |

| Background crowd | A declared state of the scene, written once in the segment header, or absent — never implied by ambient sound. |

| Practical ceiling | Four named people in one frame is the point past which quality falls off. The shot list works around it rather than through it. |

## Why naming an off-screen character costs you the shot

This is measurable, not theoretical. Shots that named an off-screen antagonist produced the wrong face or an extra person in five out of eight renders. The same shots rewritten to give a direction instead of a name produced zero such failures in twelve renders. So the pipeline rewrites them: in a single-person shot, a character who exists only as an eyeline or a direction of movement is converted to a spatial phrase and removed from the frame list.

The same logic covers the dead: a character referred to only as a body in this segment is written as a plain noun, not as a character reference, so nobody attaches their reference sheet and stands them up.

## Two people facing each other

There are exactly two ways to shoot it and the list is required to pick one: a profile two-shot, or an over-the-shoulder. What it may not write is “A facing B, with B behind in the background, out of focus” — for B to be in A’s depth of field, A has to have their back to B, and that camera position does not exist in a normal medium shot. Asked for it anyway, the renderer puts one person face-on to camera and the other standing behind them. SceneMixer sets the axis in the first shot where both are actually in frame together.

## Crowds are declared or absent

A room of three hundred guests is a visible state of the scene, so it is written as one — a line in the segment header saying who the background people are, where, and what they are doing. Consistent across every segment in that scene. Without that line the shot may not refer to a crowd at all, because “murmur of guests” in the ambient audio is not a visual instruction, and a renderer told only that will give you guests in one shot and an empty hall in the next.

When a crowd is declared, the frame instruction changes with it: the crowd stays in the depth of field, and SceneMixer's in-frame list still counts only named characters.

## What it will not do

- Put eight named characters in one shot reliably. Past four, identity degrades. Ensemble scenes are covered in singles and over-the-shoulders — the way they are shot on real sets.

- Keep a character whose in-frame flag is missing. If the model forgot to list someone who should be visible, they are left out rather than guessed in. Missing a person is recoverable; a stranger's face in your drama is not.

- Choreograph continuous group action. Crowds moving in formation, sustained multi-party fights — those are outside what current models hold together.

- Read a crowd out of sound design. Declared or absent; nothing in between.

## What it costs

Nothing extra — these are properties of the shot list, which is billed by script length at 1 credit per hundred characters of the episode's share. Reference sheets are 7 credits each beyond the 15 free ones in your first project, and SceneMixer's rendering starts at 10 credits per second of output. See character consistency for how faces stay the same across shots, and the pricing page for rates.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### How many characters can share a shot?

Four named people is the practical ceiling before identity starts to degrade, and SceneMixer's shot list is built to stay under it — ensemble scenes are covered in singles and over-the-shoulder shots rather than wide group frames.

### Why did a character I mentioned show up on screen?

That is exactly the failure SceneMixer's per-shot frame list prevents. A character who is only looked at or spoken to is written as a direction, not a name, and is not in the frame list — so no reference art is attached and no face is invented.

### Can two characters talk face to face?

Yes, as a profile two-shot or an over-the-shoulder. What does not work is one facing the other with the other blurred in the background: that combination of framing and depth is physically inconsistent, and asking for it produces a person standing behind someone talking to camera.

### How do background extras work?

They are declared once in the segment header — who they are, where, what they are doing — and stay consistent across that scene. If nothing is declared, the shot does not reference a crowd. SceneMixer does not treat ambient crowd noise alone as a visual instruction.

### Who moves their mouth?

Only the named speaker of each line. Everyone else in frame is explicitly marked with their lips closed, which is what stops a listener from silently mouthing along with the dialogue.


---

# AI Video With Talking Characters: Speech Generated, Not Synced

URL: https://scenemixer.com/ai-video-with-talking-characters/
Picture and speech come out of one pass — the line verbatim, a named speaker bound to a voice, delivery as direction, listeners lips closed, and a timing floor per line.

Talking characters

# AI Video With Talking Characters

Updated November 3, 2026

Quick answer: speech is generated with the picture in a single pass, not lip-synced onto a silent render. Each shot carries the line verbatim, the named speaker bound to a stable library voice, a short delivery direction, and an explicit note that everyone else in frame keeps their lips closed. Dialogue shots have a timing floor — three words a second in English — so a line is never crushed into a shot too short for it.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The voice comes out of the same generation as the picture

There are two ways to get a character to talk on screen. One is to render silent video and then drive the mouth with an audio track — the lip-sync route, which needs a face crop, a driving audio file, and produces a face that moves correctly and a body that does not act. The other is to generate picture and sound together, so the performance and the mouth are one decision.

SceneMixer's pipeline does the second. The default video tier is audio-visual: the model is given the line, who says it, how it is said, and a voice reference, and returns a shot with the line spoken in it. Nothing is stitched afterwards.

## At a glance

What is sent with a speaking shot

| The line | Verbatim, once. Spoken exactly as written — not paraphrased, not re-timed. |

| Who says it | Named, and bound to that character's voice reference so the same person sounds the same across episodes. |

| How it is said | A short direction in brackets — under her breath, still smiling, forcing it flat. Direction, not subtitle. |

| Who else is in frame | Marked lips closed, so listeners do not mouth along. |

| Off-screen speech | Marked first in the bracket — voice-over, on the phone, over the intercom — and then everyone visible is described with their mouth shut. |

| Timing | A shot carrying dialogue is never shorter than the line needs: three words per second in English, five characters per second in Chinese. |

## Voices are matched, not synthesised per character

In SceneMixer, each character is assigned a voice from a library of pre-built system voices — you hear a three-to-eight-second sample before choosing, can re-match, can pick a different one from the selector, or can upload your own recording. The reason it is matching rather than per-character voice design is cost and stability: designed voices drift between generations and cost per character, while a library voice is the same voice in episode one and episode twelve.

If a character has no voice assigned, the shot falls back to a written description of the voice — age, texture, register — which the model interprets. That is the default for new projects; assigning voices is a deliberate step, not something that happens silently to every character.

## The timing rule, and why it exists

A four-second shot cannot hold sixteen words of English. Either the delivery is rushed into unintelligibility or the line is cut. So SceneMixer's dialogue shots have a floor derived from the line: word count divided by three for English, character count divided by five for Chinese, and short lines get a pause added on top rather than being crushed to their theoretical minimum. Shots that come back too short for their line are flagged and corrected before anything is rendered — which is cheaper than discovering it in the output.

## Language

Dialogue language is a project setting, chosen once, and it is applied where the lines are first written rather than translated afterwards — so the performance is authored in that language instead of being a dub. Fifteen interface languages are available and dialogue is supported across them; a couple of languages occasionally mispronounce individual words on the default model tier, and the selector says so where it is true. The character voices page covers this in detail.

## What it will not do

- Clone a real person's voice. Not offered. You can upload your own recording as a reference for your own characters.

- Lip-sync an existing video or photo you upload. SceneMixer's speech comes from generation, not from driving an existing frame.

- Hit an exact runtime for a line. The floor is enforced; the precise delivery length is the model's.

- Burn in subtitles. No text is rendered in frame. Captions go on after export.

- Guarantee perfect pronunciation of every proper noun. Invented names in particular are worth previewing before you commit an episode.

## What it costs

Speech is included in the video price — there is no separate audio charge. Video starts at 10 credits per second of output (≈ $0.06/s at the lowest credit price) and runs to $0.47/s on the sharpest tier; assembly is 1 credit per second. Voice matching from the library costs nothing. New accounts get 50 credits and one free preview of up to 5 seconds — enough to hear a character speak before deciding. Rates on the pricing page.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### Is this lip-sync?

No. SceneMixer generates picture and speech together in one pass, so the mouth and the performance are the same decision. There is no separate audio track being matched to a silent render, and no way to drive a photo or an uploaded video.

### Can I use my own voice?

You can upload a recording to SceneMixer as a character's voice reference. Cloning someone else's voice is not offered.

### Do listeners in the shot move their mouths?

They should not — SceneMixer explicitly marks every non-speaking character in frame with their lips closed, which is the instruction that stops a listener mouthing along with the line.

### What happens with voice-over and phone calls?

SceneMixer writes the off-screen marker first in the delivery note, and then describes everyone visible in the shot with their mouth shut — so the line is heard without anyone appearing to say it.

### Why is my dialogue shot longer than I wrote?

Because the line needs the time. SceneMixer gives shots carrying dialogue a floor of roughly three words per second in English, five characters per second in Chinese, with a pause added for very short lines. A shot too short for its line gets corrected before rendering rather than after.


---

# AI Costume Consistency: One Reference, Every Shot | SceneMixer

URL: https://scenemixer.com/ai-costume-and-outfit-consistency/
Wardrobe is established once on a reference sheet and inherited by every shot. A full outfit change becomes its own image; rain, dust and a loosened collar are just text.

Costume consistency

# AI Costume & Outfit Consistency

Updated November 3, 2026

Quick answer: clothing is established once on a character's reference sheet and inherited by every shot that refers to them — shots never re-list the outfit, because doing so competes with the face and pulls the camera back until the head is a dot. A real change of garment becomes its own reference image; weather, dirt and a loosened collar are written as text, because reference-conditioned models copy what is in the picture and accept effects layered over it.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Wardrobe is established once, then referred to

The instinct is to describe the clothes in every shot. It is the wrong move, and expensively so: a paragraph of wardrobe in a shot prompt competes with the face for the model's attention, and in a tall frame it pulls the camera back far enough to shrink the head until identity is gone. A pale overcoat filling seventy percent of the frame with a tiny face on top is a real failure mode, not a hypothetical.

So SceneMixer keeps clothing on the character's reference sheet, and shots refer to the person rather than re-describing the outfit. What the sheet establishes, every shot inherits.

## At a glance

Where wardrobe is decided

| The reference sheet | One outfit, described in cause terms — cut, fabric, construction, closure — and rendered as a casting photo on white. |

| The costume axis | A project-level register (period, tailoring tradition, fabric and craft) injected once during analysis, so the whole cast is dressed in one world. |

| A second outfit | Becomes its own reference image — a state of the character, not a sentence in a shot. |

| Small changes | Rain, dust, a loosened collar, something held: written as text in the shot. No second image needed. |

| Shot prompts | Refer to the character. They do not re-list the clothing, and they do not pull back to display it. |

## The line between a state and a sentence

This is the distinction that decides whether you need a second image, and it comes from how reference-conditioned video models behave: they copy what is in the picture — face, garments, accessories, structure — almost verbatim, and they accept as text anything that sits on top of the picture as a global effect — weather, time of day, light, dirt.

- Make a reference image when the garment changes: a complete change of outfit, a transformation, armour on instead of robes, a uniform instead of civilian clothes, burned or torn beyond recognition.

- Write it as text when the garment is the same and something is happening to it: soaked, snow-covered, bloodstained, sleeves rolled, coat carried instead of worn.

- Two forms that can appear in the same frame are not states — they are two characters, or a character and a prop.

Getting this line wrong in the direction of too many images is the common error: early on, most auto-detected “states” were things like holding a cup or a lamp being lit. Those are text. SceneMixer now caps how many states it will propose per character precisely because over-generating them costs money and buys nothing.

## Two outfits written into one description

Scripts routinely write a character's wardrobe as an arc — early episodes in one thing, later in another, plus a wedding look. Handed all three as a single description, the render draws a figure wearing half of each. So the description is split structurally before any image is generated: the first outfit stays on the base sheet, and the rest become separate state images. Part labels (footwear, accessories, fabric) are recognised as parts of one outfit rather than as a second one.

## Which image a given shot uses

Per shot, in priority: a state chosen for that specific segment, then a per-episode override image, then a state assigned for that episode, then the base sheet. Only states that have actually been rendered count — if a state exists on paper but has no image, the shot falls back to the base sheet and carries the difference as text instead, so the change is still described even when there is no picture of it.

SceneMixer assigns states automatically: after a shot list is generated, a pass decides which segments use which state and writes it down. It is done per segment rather than per episode on purpose — a flashback outfit applied to a whole episode would put the character in the wrong clothes for every shot in it.

## What it will not do

- Dress a character from a photo of a real garment. You can upload a reference image for a character, but there is no garment-transfer feature.

- Change an outfit in a rendered shot. Wardrobe is fixed at generation. Changing it means rendering again.

- Keep a state you never rendered. States without images fall back to the base sheet plus text.

- Preserve manual per-segment state picks through a shot-list regeneration. Those are recalculated; episode-level choices survive.

- Hold up under a long coat in a vertical full-length frame. Known limit: the garment wins and the face shrinks. Use closer framing.

## What it costs

Each state image on SceneMixer is a generated image: 7 credits. Your first project includes 15 free reference images across characters, locations and props, and state images have their own separate free allowance in that first project — so trying one costs nothing. Text-only changes cost nothing at all, which is the main reason the state-versus-sentence line is worth learning. See character sheets for how the base image is built, and the pricing page for rates.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### Why is the outfit not described in every shot?

Because re-describing it competes with the face for attention and pushes the camera back to display the clothes, shrinking the head until the character is unrecognisable. SceneMixer establishes wardrobe once on the reference sheet, and every shot that refers to that character inherits it.

### When do I need a second image for a costume?

When the garment itself changes — a full change of outfit, armour, a uniform, something burned or torn beyond recognition. If the same garment is just wet, dusty, bloodied or unbuttoned, write it in the shot text; reference-conditioned models accept those as effects layered over the picture.

### My script describes a character's wardrobe changing over the season. What happens?

SceneMixer splits the description before anything is drawn: the first outfit stays on the base sheet and the later ones become their own state images. Handed all of them at once, a single render draws a figure wearing parts of each.

### Which image gets used in a given shot?

On SceneMixer, a state chosen for that segment wins, then a per-episode override, then a state assigned to the episode, then the base sheet. States that have no rendered image do not count — the shot uses the base sheet and describes the difference in words.

### Can I change the outfit in a shot I already rendered?

No. SceneMixer fixes wardrobe when the shot is generated, so a change means generating that segment again at the usual per-second price.


---

# AI Prop Generator: Objects That Stay the Same Object

URL: https://scenemixer.com/ai-prop-and-item-generator/
Props get reference images with a stated job per shot, attach only from the shot's own prop list, and get a second image only when the object itself changes shape.

AI prop generator

# AI Prop & Item Generator

Updated November 3, 2026

Quick answer: objects the story acts on get their own reference image on a clean background, and each shot that uses one states the image's job — this exact object, in this hand, now. Images attach only from the shot's own prop list, so a mention in the set description never drags an object into seven shots. Drawn or broken gets a second image; held, lit or dirty is handled in text.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A prop is a reference image with a job

Objects that matter in a story — the letter, the blade, the phone with the message on it — need to look the same every time they appear, and generated video has no memory between shots. So each one gets a reference image, and each shot that uses it says what role that image is playing. Without the role, the model treats a picture of a dagger as a mood board rather than as “this exact dagger, in this hand, now”.

## At a glance

How prop art is made and used

| Extracted | During script analysis, alongside characters and locations. Objects the story acts on, not set dressing. |

| Framing | Follows the project aspect ratio. Clean white background; light used to reveal material, not to build mood. |

| Which shots get it | Only shots whose prop field names it. An object mentioned in the location description or the opening state does not attach an image. |

| Role in the shot | Stated explicitly — this is the reference for the object being held, not a style suggestion. |

| Second form | Drawn or broken gets its own image. Held, put down, pointed, lit does not. |

| Cap | One auto-proposed state per prop. Over-generating states costs money and changes nothing. |

## Why the prop field is the only thing that counts

A dashboard panel once appeared in seven consecutive shots of an episode because the phrase sat in the location description and every shot inherited it. SceneMixer's fix was to stop treating mentions as references: images attach only from the shot's own prop list, and a prop that no line of the shot text actually mentions is dropped from that list before rendering. Set dressing describes the world; props are the objects the camera is asked to show.

## Sheathed, drawn, broken

Objects with two physical configurations used to be drawn as one picture showing both — a blade out of its scabbard with the scabbard beside it. That is a good casting photo and a bad reference: the render kept snapping the drawn blade back into its sheath mid-shot. Now the base image is the object in its carried state, and drawn is a separate state image, chosen per segment like a costume change.

The same rule covers breakage. A cracked, snapped or burned version of an object is a different object structurally, so it gets its own image; an object that is merely dirty, wet or lit is the same object and is handled in text.

## Describing an object so it can be drawn

The same rule that governs every visual prompt in SceneMixer applies: describe causes, not results. “Ornate”, “precious” and “masterfully crafted” give a model nothing to draw. Material, construction, wear, closure, proportion do.

- Instead of “an exquisite antique locket” — write a hinged oval case, tarnished silver, worn smooth at the thumb catch, hairline scratches across the front.

- Instead of “dramatic lighting on the blade” — write a single hard source from the upper left, so the forging pattern reads along the edge.

- Do not write what the light should land on. Written as a landing point, the render paints a bright patch exactly there.

## Editing and replacing

Every prop can be renamed, re-described, regenerated or replaced with your own upload. On SceneMixer, regenerating costs one image. Renaming is safe — shot text is rewritten to follow the new name, so references do not break — and uploading your own picture is the right move when the object is real and you have a photo of it.

## What it will not do

- Render readable text on an object. Labels, letters and screens with legible writing are outside what current models do; a letter can be a letter, but not a letter you can read.

- Keep a prop in a shot that does not list it. By design.

- Turn a photo into a 3D asset. SceneMixer treats uploads as reference images, not models.

- Track an object's exact position between shots. Continuity is carried by the written opening and closing state of each segment, not by simulation.

## What it costs

Prop images are ordinary generated images: 7 credits each beyond the 15 free reference images in your first project, which cover characters, locations and props together. State images have their own separate free allowance in that first project. Attaching props to shots costs nothing, and so does renaming or re-describing. See location plates for the same approach applied to sets, and the pricing page for rates.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### What counts as a prop rather than set dressing?

An object the story acts on — picked up, handed over, drawn, read, broken. Set dressing describes the room and stays in the location plate; SceneMixer gives props their own reference image and attaches them per shot.

### Why did an object appear in shots it should not be in?

That happens when a mention leaks in from the location description. SceneMixer now attaches images only from each shot's own prop list, and props no line of the shot actually mentions are dropped before rendering.

### How do I show a sword being drawn?

SceneMixer's base image is the sword as carried; drawn is a separate state image, selected for the segments where it applies. Drawing both configurations in one picture used to make the render snap the blade back into its scabbard mid-shot.

### Can I upload a photo of a real object?

Yes, as that prop's reference image. It is used the same way a generated one is. It stays a reference picture, not a 3D asset.

### Can a prop carry readable writing?

Not reliably — no current video model produces dependable lettering, which is also why SceneMixer renders no text in frame anywhere in the pipeline. A note can be a note; it cannot be a note the audience reads.


---

# AI Animated Series Generator: Drawn Style, Same Pipeline

URL: https://scenemixer.com/ai-anime-series-generator/
Set the medium once and the whole series is drawn in it — sheets, plates, every shot — with style, art direction, costume and structure as four independent axes.

AI animated series

# AI Animated Series Generator

Updated November 3, 2026

Quick answer: an animated series uses the same pipeline as a live-action-style one, with the visual style axis set to a drawn medium. Style, art direction, costume register and content type are four independent choices made once per project, so a drawn medium does not drag the world back to a default look. Identity still comes from a reference sheet attached to every shot, and the quality instructions adapt so a flat-shaded series is never told to show skin pores.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Style is one axis of four, and it only governs how things are drawn

An animated series is not a different pipeline. It is the same one — script to episodes, characters and locations to reference art, shot lists to rendered segments — with the rendering medium set to a drawn style instead of a photographic one. That setting is one of four independent choices made once per project, and keeping them independent is what stops them fighting each other.

The four project-level style choices

| Axis | Governs | Example of what belongs here |

| Visual style | The medium and its surface — how everything is drawn or rendered. | Flat cel shading; painted backgrounds with drawn figures; 3D with an illustrated finish. |

| Art direction | What the world is built out of: space, materials, colour discipline, how dense the dressing is. | Sparse interiors, cool concrete and glass, one accent colour. |

| Costume register | What people wear: period, cut, fabric, construction. | Contemporary tailoring; layered historical dress; institutional uniforms. |

| Content type | Structural rules — episode length ceiling, performance grammar, dialogue pacing. | Short drama, long-form series, or feature. |

SceneMixer ships twenty-one visual styles, thirty-seven art directions and forty-six costume registers as presets; paid members can write their own. The practical consequence of the separation: you can put a drawn medium over a modern art direction and contemporary clothes, and nothing pulls the world back toward a default look.

## What changes in a drawn style, and what does not

- Character sheets are still casting photos. Same white background, same identity-first structure, same rule against flat even lighting — rendered in your chosen medium. A reference sheet is a reference, not a key visual.

- The quality tail adapts. Flat-shaded mediums get a neutral finish instruction rather than the photographic one, so a cel-shaded series is never simultaneously told to show skin pores and not to.

- Non-human characters get a different base. A creature explicitly marked as non-humanoid is drawn as a field-guide plate — real anatomy, on all fours, no human hands, no clothes — rather than as a person in costume.

- Timing does not change. Dialogue still needs its seconds; segments are still 4–15 seconds with some tiers to 30; runtime is still stacked from segments.

- Text still never appears in frame. No title cards, no signage you can read, no burned-in captions.

## Consistency is the whole game, and it comes from art not prompts

Drawn styles are less forgiving than photographic ones about identity, because a stylised face has fewer distinguishing features to begin with. The mechanism is the same: each character is generated once as a reference sheet and that image is attached to every shot they appear in, with a couple of face-distinguishing notes pulled from their description to help when two characters share a silhouette. Shots never re-describe the outfit, for the same reason as in live-action style — it competes with the face.

One consequence worth planning for: the style axes are expanded at generation time, not snapshotted. If you change an art direction preset after generating half a season, regenerating a location plate later gives you a different world than the one you already have. Pick the axes before generating art.

## Where it is weakest

- Sustained stylised action. Continuous fight choreography and crowds moving in formation are outside what current models hold together, drawn or otherwise.

- Frame-exact effects. SceneMixer has no keyframing and no compositing layer. Effects happen inside a generated shot or not at all.

- Character count. Past four named people in one frame, identity degrades — arguably faster in flat styles. Ensembles are covered in singles.

- An existing IP look. Presets describe mediums and techniques, not studios. Custom style text is yours to write, and what you write is your responsibility.

## What it costs

Exactly what live-action style costs — the medium is not a surcharge. Analysis is 4 credits per thousand characters, per-episode shot lists 1 credit per hundred characters, reference images 7 credits each beyond the 15 free ones in your first project, and video from 10 credits per second of output. A 90-second episode is about 990 credits of video and assembly at the cheapest tier. Custom style text on SceneMixer requires a paid tier; the presets do not. Rates on the pricing page.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### Is there a separate animation pipeline?

No — SceneMixer runs the same pipeline with the visual style axis set to a drawn medium. Script analysis, episode breakdown, reference art, shot lists and rendering all work identically.

### How do I keep a stylised character on model across episodes?

The same way as a photographic one: SceneMixer generates the character once as a reference sheet and sends that image with every shot they are in, along with a couple of face-distinguishing notes. Shots refer to the character rather than re-describing their look.

### Can I write my own style description?

Yes, on SceneMixer's paid tier. Write it in cause terms — medium, surface, line, palette discipline — rather than as praise; words like “masterpiece” and “8K” are vocabulary from an older generation of models and do nothing now.

### Can I match a specific studio's look?

SceneMixer's presets describe mediums and techniques rather than studios, deliberately. Custom style text is yours to write, and what you put in it is your responsibility.

### What about non-human characters?

In SceneMixer, a character explicitly marked as non-humanoid uses a different base prompt — real anatomy, four legs on the ground, no human hands or clothing — instead of being drawn as a person in a costume.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Video Cost Breakdown: Five Line Items, One Worked Example

URL: https://scenemixer.com/ai-video-cost-calculator/
What an AI short drama actually costs, itemised: analysis, reference images, shot lists, video per second and assembly — with an 8-episode series priced out end to end.

Cost breakdown

# AI Video Cost Calculator

Updated November 3, 2026

Quick answer: five operations cost money — analysis, reference images, shot lists, video per second and assembly — and editing text is free. In a worked 8-episode vertical series from a 40,000-character script, video is about nine tenths of the bill and all the text work is about three percent. So the only levers that move the total much are how many seconds you render and at which tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Five things cost money, and only one of them is video

Everything in the pipeline is priced per unit of work, and the bill for a project is the sum of five line items. Nothing is a subscription-gated feature with a hidden meter; nothing is billed by seat or by month unless you choose a plan for the discount.

The five billable operations

| Operation | Billed by | Rate |

| Script analysis | Script length, once per project | 4 credits per 1,000 characters (3,333 characters = one unit for alphabetic scripts) |

| Reference images | Per image | 7 credits |

| Shot lists | Per episode, by that episode's share of the script | 1 credit per 100 characters |

| Video | Per second of output, by model tier | from 10 credits/second |

| Assembly | Per second of finished episode | 1 credit/second |

Editing text — outlines, shot scripts, character descriptions, names — is free. Re-generating anything costs the same as generating it the first time, because it is the same work.

## A real project, itemised

An 8-episode vertical series from a 40,000-character English script, 90 seconds an episode, on the cheapest video tier:

Worked example: 8 × 90-second episodes

| Analysis | 36 credits |

| Shot lists, 8 episodes | 90 credits |

| Video, 720 seconds | 9,000 credits |

| Assembly, 720 seconds | 900 credits |

| Total | 10,026 credits ≈ $62 |

Two things to read off that table. Video is roughly nine tenths of it, so the only lever that matters much is how many seconds you render and at which tier. And the text work — the part that decides whether the series is any good — is about three percent, which is why iterating on outlines and shot lists before rendering is not just good practice but good arithmetic.

## The levers, in order of how much they move the bill

- Seconds rendered. Linear and dominant. A 60-second episode costs two thirds of a 90-second one.

- Video tier. SceneMixer's per-second rate spans $0.06 to $0.47 per second. 480P is the default and is genuinely fine on a phone; the sharper tiers cost what they cost.

- Re-renders. Every regeneration is a full charge. Preview one segment before committing an episode — that is what the free preview is for.

- Credit price per pack. Larger packs carry bonus credits, so the effective price per credit ranges across $0.0090–0.0096.

- Membership discount. : 5 include a monthly credit allowance and a standing discount on purchases. Total discount is capped at 50% by design.

- Script length. Affects analysis and shot lists only — a few percent of the total. Alphabetic scripts count 3,333 characters as a thousand, which is the cost-aligned rate rather than a promotion.

## What you can try without paying

New accounts on SceneMixer get 50 credits. The first project additionally includes 15 free reference images — five each for characters, locations and props — and the first episode's shot list free. There is one free video preview of up to 5 seconds. That is enough to run a real script through analysis, see your cast, read a shot list and watch a few seconds of rendered output before spending anything.

## How the money is actually handled

- Credits are frozen, then spent. Submitting a job reserves its cost; success converts the reservation into a charge, and a confirmed upstream failure returns it.

- SceneMixer does not charge for delivery failures. If a job produced nothing usable on our side, the reservation is held for review rather than spent.

- Every movement is in a ledger. The credits panel shows each freeze, charge and refund with its job.

- Price shown is price charged. Every button's label comes from the same quote endpoint the charge uses, so the number on the button is the number billed.

## What it will not do

- Quote a project before analysis. Shot-list and video costs depend on how the script actually divides. Analysis is quoted up front from the text you paste.

- Bill by the month for generation. Memberships give allowance and discount; SceneMixer always bills generation per unit of work.

- Refund a delivered render you dislike. The model ran and the upstream cost was incurred. Preview first.

- Charge for deleting things. Deleting projects and assets is free; note that deleting a target mid-generation still costs the generation, because the model already ran.

## Current rates

Everything above is pulled from the live rate table, so it is current as of this page load. The pricing page lists packs, tiers and per-tier video rates in full; micro drama economics works through the per-episode arithmetic for a vertical series.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Frequently asked questions

### What dominates the cost of a series?

Video, by roughly nine to one. In an 8×90-second worked example, rendering and assembly are about 9,000 and 900 credits against 36 for analysis and 90 for all eight shot lists. Seconds rendered and the video tier are the only big levers.

### Is editing free?

Yes. Outlines, shot scripts, character and prop descriptions, names — all free to change as often as you like. On SceneMixer, only generating and re-generating costs credits.

### Does a re-render cost full price?

Yes, because it is the same amount of upstream work. This is why the free preview exists: check one segment before committing a whole episode.

### Can I estimate a project before uploading the script?

Only roughly, because shot-list and video costs depend on how your script divides into episodes. Paste the script and SceneMixer quotes analysis exactly before you start it; from there the episode list tells you the rest.

### What do memberships change?

They add a monthly credit allowance and a standing discount on purchases, capped in total at 50%. Generation is still billed per unit of work — SceneMixer has no unlimited plan.


---

# AI Short Drama for Vertical Feeds: What Ships and What Does Not

URL: https://scenemixer.com/ai-drama-for-tiktok-and-reels/
Vertical frame, hooked openings and 60–120 second episodes are decided before rendering; captions go on after export, because no video model spells reliably.

Vertical feed delivery

# AI Drama for TikTok & Reels

Updated November 3, 2026

Quick answer: three of the four things a short-form feed demands are production decisions, not export settings: the 9:16 frame is set before any art is generated, the hook is built into the episode breakdown, and the 60–120 second length is paced in the outline. Captions are the exception and are deliberately left off the video, because no current model spells reliably and burned-in text would freeze one language into the file.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What the platforms actually demand of the file

Short-form feeds want a tall frame, a fast opening, speech that reads without sound on, and a length that fits a single sitting. Those are production decisions, not export settings, and three of the four have to be made before any video exists.

Feed requirements and where they are decided

| Requirement | Decided | How |

| 9:16 vertical | Step 1, before any art | Aspect ratio is a project setting; reference art and every shot are composed for it. Not a crop. |

| Immediate hook | Episode breakdown | Openings start on the problem; episodes end unresolved, with the kind of ending rotating. |

| 60–120 seconds | Episode outline | Short-drama pacing caps at 240s; vertical series usually settle in the 60–120 band. |

| Captions | After export | Deliberately not burned in — no video model spells reliably, so text goes on at the platform. |

| Dialogue in the viewer's language | Step 1 | Language is chosen once and the lines are authored in it, rather than dubbed afterwards. |

## Why the video model never draws captions

Two reasons, and neither is laziness. Current video models cannot produce reliable lettering — asked for text they return plausible-looking garble, which is why SceneMixer renders no words in frame at all. And even if they could, burned-in captions would freeze one language and one style into the file: changing them would mean re-rendering shots rather than editing a caption track. So the video comes out clean. If you want dialogue subtitles burned in, accounts that have bought credits can add them over the finished video or download an SRT file; otherwise caption it on the platform.

The same applies to hook cards and end cards for a series episode: overlay them on the platform, change them per post, re-render nothing.

## What comes out of an export

- Two renditions per segment. A full-quality file for download and assembly, and a lighter one used for in-app playback. The download is the full-quality one.

- A finished episode. SceneMixer assembles segments into one file at 1 credit per second of runtime.

- An AI badge, by default. A small corner mark is applied at delivery. Paid members can turn it off; the toggle sits next to the model selector and applies to subsequent renders, not past ones.

- Share links with previews. Any segment or finished episode can be shared as a page with proper social preview metadata, without publishing your project.

- No platform upload. There is no direct post-to-feed integration. You download and upload.

## Practical sequence for a feed series

- Set the frame and the dialogue language first. Both propagate into reference art and every line; changing them later means regenerating art.

- Read the episode list before rendering anything. Each outline states how that episode ends. Fixing endings as text is free; fixing them as video is not.

- Render one segment. SceneMixer's free preview covers up to 5 seconds. Watch it on a phone, not a laptop.

- Render one full episode before committing a season. At 10 credits a second, a 90-second episode is about 990 credits including assembly.

- Caption and publish per platform. Same clean master, different caption tracks and covers.

## What it will not do

- Reframe a landscape video to vertical. Vertical comes from generating vertical. There is no crop-and-track pass.

- Burn in captions, hook text or an end card. By design, as above.

- Post to a platform for you. No feed integrations. Export and upload.

- Advise you on a platform's AI-content rules. Those are the platform's, they change, and they are yours to read. What this product does is stamp a badge by default and let paid members remove it.

- Produce one continuous take of a whole episode. Segment caps are per tier; runtime is stacked.

## What it costs

Video from 10 credits per second of output (≈ $0.06/s at the lowest credit price), assembly 1 credit per second, shot lists 1 credit per hundred characters, analysis 4 credits per thousand. New accounts on SceneMixer get 50 credits, 15 free reference images and one free preview. The cost breakdown prices out a full eight-episode series; the vertical page covers what changes in a tall frame.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## Frequently asked questions

### Do I get captions burned into the video?

Only if you ask for them. No current video model produces reliable lettering, so the generated picture never contains text. Accounts that have bought credits can add dialogue subtitles over the finished video or download an SRT file; a caption track on the platform still works if you would rather change captions per post.

### Can I turn a landscape episode into a vertical one?

Not by cropping — there is no reframing pass. On SceneMixer, vertical output comes from setting the project to 9:16 before reference art is generated, so every shot is composed for the tall frame.

### How long should a feed episode be?

Sixty to a hundred and twenty seconds is the working band for vertical short drama, and the outline paces for it. The format's ceiling on SceneMixer is 240 seconds per episode.

### Is there an AI watermark on the output?

SceneMixer applies a small corner badge at delivery on free accounts; members' videos are badge-free by default. Paid members can switch it off for subsequent renders. Whatever a given platform requires in the way of AI disclosure is the platform's rule and worth reading yourself.

### Can it upload to TikTok or Reels directly?

No. SceneMixer has no posting integration — you download the finished episode and upload it yourself, which is also what lets you caption the same master differently per platform.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.


---

# AI Drama Generator: Pick a Genre, Get a Consistent Series

URL: https://scenemixer.com/ai-drama-generator/
Generate a drama series by genre: a script from one line or your own, a cast drawn once, scenes that stay put, dialogue in 15 languages, and per-second prices up front.

# AI Drama Generator

Quick answer: Choose a genre, write one line or paste your own script, and SceneMixer turns it into a drama: it reads the story into episodes, draws every character, place and prop once, writes a shot-by-shot storyboard, and generates each episode in short segments with the dialogue spoken. Video is charged by the finished second (10 credits a second on the default tier), and your first drama comes with free setup images and a free first-episode storyboard.

Brief for AI Create: A contract-marriage drama: a broke architect agrees to a one-year marriage with the heir she once humiliated at a design contest; ten episodes, office and penthouse, a hook at the end of each.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What makes a drama different from a clip

A clip generator answers one prompt with a few seconds of footage. A drama needs the same face in episode 9 as in episode 1, a living room that keeps its sofa where it was, a voice that does not drift, and a script that knows who is speaking. That is the work SceneMixer does before any video is made: the read lists the cast, the locations and the props, and step 2 draws each of them once as a reference that every later shot uses. Why a whole series stays consistent.

Then the storyboard turns each episode into numbered shots with seconds, positions and lines, and the video is generated segment by segment from that text. You read and edit the storyboard before you spend on video, which is where most mistakes are cheapest to fix. Editing the storyboard.

## Start from a genre

Every genre is made the same way; what changes is the cast, the sets and the look you pick in step 1. The genre guides give the settings that suit each one:

| Genre | Settings that suit it | Guide |
| Billionaire romance | Short Drama, 9:16, Digital HD, a modern city | Billionaire romance |
| Revenge and rebirth | Two states of one face; the death never shown | Rebirth revenge |
| Werewolf and vampire | Soft Glow or Shot on Film; the turn as a state | Werewolf romance |
| Mafia romance | Night streets, one mark per family | Mafia romance |
| Horror | A light source in every dark shot; no gore | Horror short |
| Fantasy and dragons | The creature as a character with its own sheet | Dragon fantasy |
| Period and palace | Court dress by rank; one great hall | Royal drama |
| Crime and courtroom | Every clue a prop; long speeches split across shots | Courtroom drama |
| Comedy and sitcom | One room, a small cast, timing in the shot seconds | Sitcom |
| Anime | A drawn look chosen once for the whole series | Anime series |

The full list, with more than seventy genres and formats, is on the genre overview.

## Short drama, TV series or film

Type is the first real decision. Short Drama writes episodes of up to 240 seconds with a hook at the end of each, for a vertical feed. TV Series allows up to 300 seconds and a slower build for ensemble stories. Film has no episode cap and lets a scene end on a look. SceneMixer fixes it once the script is read, because the episodes were cut for it. Short drama, series or film.

## What the first episode costs

Reading the script is 4 credits per 1,000 characters (English counts lighter than Chinese). Your account's first 15 setup images are free on the default image model, and episode 1's storyboard is free on your first drama. Video is the main cost: 10 credits a second on the default tier (≈ $0.06 a second at the lowest credit price), plus 1 credit a second to join the episode. A 90-second episode on the default tier comes to 990 credits of video and joining, before any retakes. Where credits go in one episode.

## What it will not do

- Real people and famous characters. SceneMixer refuses celebrity likenesses, known film and anime characters and brand mascots. IP check.
- Gore and explicit content. Violence is shown before and after, not as wounds. Content rules.
- One long take. Each generation is a segment of up to 30 seconds on the tiers built for long takes; an episode is a stack of them.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### What is an AI drama generator?

A tool that turns a story into a drama with recurring characters, places and dialogue, generated episode by episode, rather than one clip from one prompt.

### Do I need a script?

No. AI Create writes one from a single line; you can also paste or upload your own in TXT, DOCX or Markdown.

### Which genres can I make?

Any original story: romance, revenge, fantasy, horror, crime, comedy, anime and more. Real people, famous characters, gore and explicit content are not allowed.

### Will the characters look the same in every episode?

SceneMixer draws each character once in step 2 and uses that sheet for every shot, so faces, outfits and rooms hold across the series.

### Is it free?

New accounts on SceneMixer get 50 credits, 15 free setup images, a free first-episode storyboard and a free preview of about 5 seconds. Reading scripts and generating video are paid.

### Can it make vertical dramas?

Yes. Choose 9:16 Portrait in SceneMixer's step 1 for a phone feed, or 16:9 Landscape for YouTube and TV.

### What languages can the characters speak?

Any of 15 languages, set once in step 1.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Movie Maker: From One Idea to a Finished Film

URL: https://scenemixer.com/ai-movie-maker/
Make a movie with AI from one idea: AI Create writes the script, Film type paces it, the cast is drawn once, and you render, fix single shots and join the cut.

# AI Movie Maker

Quick answer: Type an idea, set the content type to Film, and SceneMixer writes the screenplay, reads it into scenes, draws the cast and the locations once, plans every shot and generates the film segment by segment with spoken dialogue and a score. You watch each sequence before paying for the next, fix single shots instead of whole scenes, and join the result into one file. Video costs 10 credits a second on the default tier.

Brief for AI Create: A quiet film about a retired ferry captain who teaches his granddaughter to sail the week before his boat is sold; three days, one harbour, no villain.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The whole road, in order

| Stage | What you do | What you get |
| Idea | One line in AI Create, a length from Mini to Long | A full screenplay you can read and edit |
| Step 1 | Type Film, 16:9, a visual style, dialogue language | The script read into acts and scenes |
| Step 2 | Generate the cast, places and props | Reference sheets every shot will use |
| Storyboard | Generate and read the shot list | Numbered shots with seconds and lines |
| Video | Render, watch, fix single shots | Segments with dialogue and sound |
| Cut | Join the segments | One film file, music on or off |

SceneMixer shows the price of each stage before you press the button. Your first five minutes walks through the screens.

## Why Film type matters

Short-form drama forces a hook every ninety seconds and dialogue that says everything out loud. Film switches that off: scenes can breathe, lines can carry subtext, and a sequence can end on a look. It also removes the per-episode length cap, so the story is divided into acts rather than episodes. Choose it before the read; it cannot be changed afterwards. Short drama, series or film.

## Directing without a camera

You direct through text. The storyboard says who is in frame, where they stand, what they say and how long each shot holds. Change the words and the next render follows. Camera language such as slow push in or over-the-shoulder works when written into the shot. How to write one shot and write positions the world can read.

When a shot comes out wrong, you do not start again. Trim it, redub a line, change one thing with AI visuals, or regenerate that segment alone. Edit this clip.

## How long a movie, and what it costs

Length is built from segments of a few seconds up to 30, joined into a cut. On SceneMixer, a ten-minute short is 600 seconds of video at 10 credits a second, plus 1 credit a second for the join and whatever you retake. A feature is nine times that. The honest advice: make a short or one sequence first, watch it, and decide from there. Can AI make a full-length movie yet?

## Limits worth knowing

- No continuous long take: every shot is a generation of limited length.
- Crowd battles and long fight choreography are steered around; background people are scenery. Fights and crowds.
- No real actors' faces or famous characters.
- No timeline for footage you shot yourself; the film is generated from the storyboard.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can I make a movie with AI from just an idea?

Yes. SceneMixer's AI Create writes the screenplay from one line; then the same pipeline reads, casts, storyboards and renders it.

### How long can an AI movie be?

On SceneMixer, there is no cap on a Film project's length; it is built from segments and joined. Cost, not a limit, is what usually decides.

### Do the characters talk?

Yes. SceneMixer speaks dialogue in the generation itself, in the language you chose, with a voice per character.

### Can I use my own script?

Yes. Paste it or upload TXT, DOCX or Markdown to SceneMixer; screenplay layout works.

### Can I change a scene after it is made?

Yes. Edit the storyboard text and regenerate that segment, or trim, redub or retouch the clip.

### Is there music?

SceneMixer adds a score automatically under the voices; the Music switch gives a dialogue-only version.

### What does it cost to try?

Your first SceneMixer drama gets free setup images, a free first storyboard and a short free preview; video is charged by the second.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Film Maker for Filmmakers: Previs, Pitches and Festival Shorts

URL: https://scenemixer.com/ai-film-maker/
An AI film maker for people who make films: previs and animatics from your script, pitch trailers, mood reels and festival shorts, with an editable shot list.

# AI Film Maker for filmmakers

Quick answer: SceneMixer gives filmmakers a fast way to see a script before shooting it: upload the screenplay, choose Film, and you get an editable shot list, consistent characters and sets, and rendered sequences with dialogue. Use it for previs and animatics, a pitch trailer, a mood reel, or a festival short made entirely with AI. Members can download without the AI badge, raise clips up to 2K and export subtitles.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Five jobs it does for a production

| Job | What you make | Guide |
| Previs and animatics | Key scenes rendered from your script to test blocking and pace | Previs and animatics |
| Pitch trailer | A 60–120 second trailer with the same cast | Trailer |
| TV pilot | A whole pilot episode to show a buyer | Pilot |
| Festival short | A finished short in an AI category | Festival short |
| Key art | Three poster options with the title painted in | Poster |

## Your screenplay, as written

Scene headings, character cues, parentheticals and V.O. all read. Export from your screenwriting app as plain text or Word (rename Fountain to .txt); PDFs are refused. Keep one spelling per character and one name per place, or the read will make two. Scene headings and character cues.

SceneMixer's read lists every character, location and prop. Step 2 draws them as plain reference sheets, so the look you approve is the one every shot uses. You can upload your own concept art as a reference (a member feature). Use your own image.

## The shot list is yours to edit

The storyboard is a shot list with seconds: framing, who is in frame, positions, the line and the sound. Rewrite any of it and only the shots you touched are re-read. Write lenses and moves in plain words; a vague shot gets a generic picture. Editing the storyboard and camera movement.

SceneMixer tracks positions across cuts, so a character on the left of a table stays there in the reverse shot. Why characters swap sides.

## Delivery

- Aspect: 16:9 or 9:16, fixed per project.
- Resolution: chosen by tier; Enhance raises a finished clip a step at a time up to 2K. Enhance.
- Badge: members get badge-free downloads by default. The AI badge switch.
- Subtitles: burned in or as SRT on accounts that have made a purchase. Subtitles.
- Looks: Shot on Film, Black & White Film, Soft Glow among 21 visual styles. Visual styles.

## Where it stops

It does not edit footage you shot, it cannot hold one take longer than a single generation, and it will not put a real actor's face on a character. Treat it as a tool to see and sell a film, and to make short films outright; a feature is possible but is a budget decision. Short film budgets.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can I use AI for previs?

Yes. Upload the scenes you want to test, render them, and use the clips as an animatic to check blocking and timing.

### Does it read Final Draft or Fountain?

Export as plain text or Word; a Fountain file is plain text, so rename it to .txt. SceneMixer does not accept PDFs.

### Can I control camera angles?

Yes. SceneMixer keeps each shot's framing and movement as text in the storyboard, which you can rewrite before rendering.

### Can I get a clean file without the AI badge?

Members can switch the badge off; downloads then have none. SceneMixer's shared links always carry it.

### What resolution can I deliver?

It depends on the tier you render on; Enhance raises a finished clip up to 2K.

### Can I use my own concept art for characters?

Yes, as a member: upload it and use it as is, or let SceneMixer turn it into a reference sheet.

### Can I submit the result to festivals?

Many festivals have AI categories, each with its own rules on length and disclosure. Read them first.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Short Film Maker: Plan, Budget and Render a Short

URL: https://scenemixer.com/ai-short-film-maker/
Make a short film with AI: how long a short should run, what each length costs per second, how to keep it to one location and two characters, and how to deliver it.

# AI Short Film Maker

Quick answer: A short film is the right first project for AI video: long enough to tell a story, short enough to budget. In SceneMixer, write or upload a script, choose Film and 16:9, and render it in segments with the cast drawn once. At 10 credits a second on the default tier, a five-minute short is 300 seconds of that rate before retakes, plus 1 credit a second to join it.

Brief for AI Create: A five-minute short: a lighthouse keeper on the last night before the light is automated hears a ship's bell where no ship should be; two characters, one tower, a quiet ending.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Pick the length first

Most of the money is in seconds of finished video, so the runtime is the budget. Video only, on the default tier, before retakes:

| Runtime | Seconds of video | Segments of 15 s | Typical shape |
| 1 minute | 60 | 4 | One scene, one turn |
| 3 minutes | 180 | 12 | Setup, complication, ending |
| 5 minutes | 300 | 20 | Two or three scenes |
| 10 minutes | 600 | 40 | A full short with a second act |
| 15 minutes | 900 | 60 | Close to the upper end for many festivals |

Multiply the seconds by the tier's rate: 10 credits a second on the default tier, more on sharper ones. On SceneMixer, every retake is charged again. Plan for at least one retake in five. Spend fewer credits.

## Write small

- One location, two or three characters. Fewer sheets, fewer things to keep consistent, more budget for the shots that matter.
- Faces and dialogue over action. Conversations, looks and reveals render well; crowd battles and long fights do not.
- A turn you can see. A letter, a light, a sound. A prop that changes is a state, drawn once. States.
- An original world. No real people, no famous characters.

## Make it, step by step

- Script: paste it, or write one in AI Create at the Brief or Short length.
- Step 1: Type Film, 16:9, a considered look such as Shot on Film, Black & White Film, the dialogue language.
- Step 2: draw the cast and the location; fix faces here, where it is cheap.
- Storyboard: read every shot; set the segment length to Long take in SceneMixer if you want longer continuous shots.
- Video: render the first scene, watch it, then the rest. Fix single shots instead of regenerating scenes.
- Join the segments, turn music on or off, download.

## Delivery and festivals

Festivals set their own rules for AI work, runtime, disclosure and format; read them before you write. Members can download without the AI badge, raise clips with Enhance up to 2K, and export an SRT. A poster comes from the project's cover art. Festival guide.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### How much does an AI short film cost?

Mostly seconds of video: a five-minute short is 300 seconds at 10 credits a second on SceneMixer's default tier, plus retakes and a small join fee.

### How long should an AI short film be?

Three to ten minutes is a good range: enough for a story, small enough to fix shot by shot.

### Can I make it without writing a script?

Yes. SceneMixer's AI Create writes one from a single line; read and edit it before you continue.

### Which look suits a short film?

Shot on Film or Black & White Film for live action, or one drawn style used throughout.

### Can I enter it into a festival?

Many festivals have AI categories; check each one's rules on AI, length and disclosure.

### Can I fix one bad shot?

Yes. Trim, redub, retouch or regenerate a single segment without touching the rest.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# Long AI Video Generator: Minutes, Episodes and Whole Series

URL: https://scenemixer.com/long-ai-video-generator/
Generate long AI videos: segments of up to 30 seconds stacked into episodes of up to 60 segments, many episodes per project, joined into one file with continuity kept.

# Long AI Video Generator

Quick answer: No AI model renders ten minutes in one take; long video is built. SceneMixer generates segments of a few seconds up to 30 on the long-take tiers, holds up to 60 segments in one episode, and lets a project run to many episodes, all sharing one cast and one set of locations. Each segment opens where the last one ended, and the episode is joined into one file.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## How the length is built

| Unit | How long | What decides it |
| Shot | A few seconds | The storyboard, shot by shot |
| Segment | Up to 15 or 30 seconds | The ceiling you pick per episode and the tier |
| Episode | Up to 60 segments | Type: Short Drama up to 240 s, TV Series 300 s, Film uncapped |
| Project | Many episodes | The length of the script you upload |

A segment is one generation. You choose Standard 15s or Long take 30s for each episode's storyboard, and the shots inside are written to fit. 15 or 30 second segments.

For the shorter end of the scale, two pages work the same arithmetic the other way: 30 to 60 seconds in one or two segments and three to five minutes as one episode.

## Keeping it continuous

- Same faces: every character has one reference sheet used in every segment. Same face in every shot.
- Same rooms: a location is drawn once and becomes the set. Same room in every episode.
- Same positions: who stands where carries across cuts. Everyone stays put.
- Clean joins: each segment opens where the previous one ended. Why the next segment opens like the last one ended.

## Plan the length before the cost

On SceneMixer, the rate does not change with length; the count of seconds does. On the default tier at 10 credits a second:

| Target | Seconds | Segments of 15 s | Segments of 30 s |
| 1 minute | 60 | 4 | 2 |
| 5 minutes | 300 | 20 | 10 |
| 10 minutes | 600 | 40 | 20 |
| 30 minutes | 1,800 | 120 | 60 |

Add 1 credit a second for the join and something for retakes. A thirty-minute run is a series of episodes rather than one. How to make an AI video longer than one minute.

## Generate a long run without babysitting it

Generate several segments at once, up to your plan's limit; they run on the server, so you can close the page. When all are done, join the episode. Episodes can be storyboarded and rendered in any order. Generate all segments and multi-episode series.

## Limits

One continuous take is capped at one segment. A single SceneMixer episode cannot exceed 60 segments. A script longer than one project can read is split into volumes, each its own project. Older model tiers stop at 15 seconds a segment.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI generate a 10-minute video?

Yes, SceneMixer builds it as a sequence of segments joined into one file; no model renders it in one take.

### How long can one AI video segment be?

A SceneMixer segment is up to 30 seconds on the long-take tiers, 15 on older ones.

### How many segments can an episode have?

Up to 60.

### Do long videos cost more per second?

No. The rate per second is the same; you pay for more seconds, retakes and the join.

### Will characters change halfway through?

Every SceneMixer segment uses the same reference sheets, so faces, outfits and rooms hold.

### Can I stop and continue later?

Yes. SceneMixer saves episodes and segments; generate the rest whenever you like.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# Story to Video: Turn Any Story Into a Video With AI

URL: https://scenemixer.com/story-to-video/
Turn a story into a video with AI: paste a short story, a fable or your own anecdote; it is read into scenes, cast, voiced and rendered, with prices shown first.

# Story to Video

Quick answer: Paste a story of at least a paragraph and SceneMixer turns it into a video: the read finds the characters and places, splits the story into scenes and episodes, and writes the dialogue; you approve the cast, read the storyboard and render it with voices and music. Prose works as well as a script. Reading costs 4 credits per 1,000 characters (English counts lighter), and video 10 credits a second on the default tier.

Paste a story like: My grandmother kept a jar of buttons on the windowsill. Every button was a coat she had mended for someone in the village, and on the day she died the jar was full…

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Any story that has people in it

| You have | What comes out | Tip |
| A short story | One or a few episodes | Keep names consistent |
| A family anecdote | A short film | Invented faces; no real people |
| A fable or fairy tale | A cartoon short | Public-domain tales, your own designs. Fairy tales |
| A chapter of a novel | A drama episode | Whole books: book to video |
| A Reddit-style story you wrote | A dramatized short | Reddit story videos |
| A bedtime story | A gentle kids' episode | Kids' story videos |

## Prose becomes scenes

You do not need to rewrite a story as a screenplay. In SceneMixer, the read turns narration into scenes: who is there, where, what they do and what they say. Lines in quotation marks become dialogue; narration can stay as a narrator's voice-over if you want it heard. Dialogue, voice-over and captions.

The minimum is about a paragraph, roughly 330 characters of English; anything shorter has nothing to split into scenes. Links are not fetched; paste the text. What you can feed in.

## What you choose, what is done for you

You choose: how it looks (21 visual styles from live action to watercolour), portrait or landscape, the dialogue language, and Short Drama, TV Series or Film. Done for you: the cast list, a voice per character, the location plates, the shot list and the music. You can change any of it before rendering. Choosing the visual style.

## Narrated or acted?

A narrated story video has one voice telling the story over pictures. An acted one has characters speaking their own lines. Both are possible; acted versions hold attention better on short-form platforms. If you want narration, say so in the story or in the storyboard with a voice-over line. Story video formats.

## Before you start

- Use a story you wrote or have the rights to. A recording or likeness of a real person needs their permission.
- Keep one spelling per name, or one character becomes two.
- Read the summary after the read: it is the cheapest place to catch a misunderstanding. Checking the summary.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI turn a written story into a video?

Yes. Paste the story into SceneMixer; it is read into scenes, cast, storyboarded and rendered with voices.

### Does my story need to be a script?

No. Prose works; dialogue in quotation marks becomes spoken lines.

### How short can the story be?

SceneMixer needs about a paragraph, roughly 330 characters of English, at minimum.

### Can I have a narrator?

Yes. Write the narration as voice-over lines, or keep the story's narration and ask for it in the storyboard.

### Can I use photos of my family?

Real people's likenesses need consent; uploading photos is a SceneMixer member feature, and the video provider refuses faces that look like real photos.

### What does it cost?

Reading is 4 credits per 1,000 characters (English counts lighter); video is charged per second.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# Book to Video: Turn a Book Into an AI Series or Film

URL: https://scenemixer.com/book-to-video/
Turn a book into video with AI: novels, memoirs, picture books and public-domain classics read into episodes, cast once, voiced in 15 languages. Long books in volumes.

# Book to Video

Quick answer: Upload the text of a book you wrote or have the rights to, and SceneMixer reads it into episodes or acts, draws every character and place once, and renders it with spoken dialogue. One project reads up to 300,000 characters of English text (longer texts are read in two passes); a longer book is split into volumes, each its own project with the same cast carried across. Reading costs 4 credits per 1,000 characters, English counted lighter.

Paste chapter one of your book, choose TV Series for a season or Film for a single adaptation, and read the summary before you go on.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | The text of a book you wrote or have the rights to; .txt, .docx and .md uploads; a long book is split into volumes with the same cast |

## Which books work

| Kind of book | Works as | Notes |
| Your own novel | A series, one chapter or two per episode | The usual case; novel to video covers web novels in depth |
| A memoir or family history | A short film or a few episodes | Invented faces for real people unless you have their consent |
| A children's picture book | A short cartoon | A drawn look such as Watercolor Storybook |
| A public-domain classic | An adaptation in any style | Check the text is public domain where you publish |
| Non-fiction | Dramatized scenes | Works when there are people and moments; pure argument does not |

The book must be yours or licensed to you. In SceneMixer, famous characters from other people's books, and well-known film versions of them, are refused by the IP check. IP check.

## How a long book is read

Up to 10,000 Chinese-equivalent characters a script is read in one pass; beyond that, in two passes: first the cast, places and props, then the structure, with each episode's outline expanded when you open it. That is how one project can take 300,000 characters of English. Past that ceiling the box refuses the text rather than cutting it off silently; split the book into volumes. Long scripts read in two passes and length caps.

Files: paste the text, or drop .txt, .docx or .md. In SceneMixer, PDFs and e-book files are refused; export or copy the text. Which files work.

## One cast across volumes

Each volume is its own project, but the cast does not have to be redrawn. Pick existing characters, scenes and voices from your library in the next project, and the hero of volume two has the face from volume one. Reuse characters across projects.

## Adaptation choices you make

- Type: TV Series for a season of 300-second episodes; Short Drama for a vertical, hook-driven cut; Film for one long adaptation.
- Narration: keep the narrator's voice as voice-over lines, or let scenes play with dialogue only.
- Region and language: the same book can be cast for another market and spoken in another of the 15 dialogue languages. Second language version.
- Trailer first: a 60–120 second book trailer is a cheap way to test the look. Book trailer.

## What a book costs to adapt

The reading is small; the video is not. A season of ten two-minute episodes is 1,200 seconds of video at 10 credits a second on SceneMixer's default tier, plus 1 credit a second for joining. Make episode 1, watch it, fix the cast, then decide on the rest. What a micro-drama costs.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI turn a book into a movie?

Yes. Upload the text to SceneMixer; it is read into acts or episodes, cast and rendered. Long books are split into volumes.

### How long a book can I upload?

SceneMixer takes up to 300,000 characters of English in one project; longer books in several projects.

### Can I upload a PDF or EPUB?

No. Paste the text or export it to TXT or DOCX.

### Can I adapt someone else's book?

Only with the rights. Public-domain texts are fine; famous characters from protected works are refused.

### Will the characters look the same in volume two?

Yes, if you pick them from your SceneMixer library instead of generating new ones.

### Can it keep the narrator?

Yes. Narration can be voice-over lines spoken by a SceneMixer narrator character.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Story Video Generator: Narrated or Acted Story Videos

URL: https://scenemixer.com/ai-story-video-generator/
Generate story videos with AI: narrated storytime videos or fully acted scenes, characters that keep their faces, a voice per character, vertical or landscape.

# AI Story Video Generator

Quick answer: A story video can be told or acted. SceneMixer does both from the same text: a narrator's voice-over over consistent pictures, or characters who speak their own lines, or a mix of the two. Paste or write the story, choose the look and the shape, and it is cast, storyboarded and rendered in segments. Video is 10 credits a second on the default tier; your first drama's setup images and first storyboard are free.

Brief for AI Create: A storytime video: a nurse on her first night shift hears a patient call her by her childhood nickname; told by the nurse in voice-over, three scenes, a twist in the last line.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Three ways to tell it

| Format | What you hear | Best for | How to set it up |
| Narrated | One voice telling the story | Storytime channels, history, bedtime | Write it as voice-over lines for a narrator character |
| Acted | Characters speaking their own lines | Drama, comedy, short-form feeds | Write dialogue with a speaker for each line |
| Mixed | A narrator framing acted scenes | Mystery, memoir, true-story style | Narrator opens and closes; scenes in between |

How lines are written decides which you get. Dialogue lines, voice-over and captions.

## Pictures that match the story

Story videos made from separate prompts drift: the hero's hair changes, the house gets a new door. In SceneMixer, each character, place and prop is drawn once as a reference, and every segment is generated against those references, so the pictures match the telling from the first line to the last. Same face in every shot.

## A voice per character

Every character comes out of the read with a voice described in words. Match it to a preset from the built-in library (free, and you can listen first) for a voice that holds across episodes; members can upload a clip they have the right to use. The narrator is a character like any other. Pick character voices.

## Shape and length for each platform

- TikTok, Shorts, Reels: 9:16 Portrait, Short Drama type, a hook in the first seconds. Posting to short-form.
- YouTube long-form: 16:9 Landscape, TV Series or Film, episodes joined into one file. Faceless YouTube channel.
- Subtitles: burned in or as SRT on accounts that have made a purchase.

## Story sources

Your own writing, a public-domain tale, a story you have permission to adapt. Stories about real, identifiable people need care: invented faces, changed names, nothing defamatory. Story to video and Reddit story videos.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### What is an AI story video generator?

A tool that turns a written story into a video with pictures, voices and music, narrated or acted.

### Can it narrate my story?

Yes. Write the story as a narrator's voice-over and pick a voice for the narrator.

### Can the characters speak instead?

Yes. Dialogue with a speaker per line is performed by those characters, in the language you choose.

### Will the main character look the same throughout?

Yes. SceneMixer uses one reference sheet per character for every segment.

### Can I make vertical story videos?

Yes. Choose 9:16 Portrait in SceneMixer's step 1.

### Can I add subtitles?

Yes, burned in or as an SRT file, on accounts that have made a purchase.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Cartoon Series Generator: Recurring Characters, Every Episode

URL: https://scenemixer.com/ai-cartoon-series-generator/
Make a cartoon series with AI: one drawn cast for every episode, 3D and 2D looks from Family 3D to Watercolor Storybook, voices per character, episodes in order.

# AI Cartoon Series Generator

Quick answer: A cartoon series lives on recurring characters. SceneMixer draws your cast once in the cartoon look you choose, then makes every episode from the same sheets, places and voices, so episode 12's hero is episode 1's hero. Pick a 3D or 2D style in step 1, upload or write the episodes, and render them one by one at 10 credits a second on the default tier.

Brief for AI Create: A cartoon series about a small robot who runs the lost-and-found at a busy train station; each episode returns one lost thing to its owner; ten episodes, gentle humour.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Choose the look once

The visual style is fixed for the project, so it is the most important choice for a series. The cartoon-friendly presets:

| Style | Feels like | Good for |
| Family 3D Animation | Big eyes, rounded, glowing skin | Family comedy, adventures |
| 3D Toy Cartoon | Matte toy-like figures, bright colours | Preschool, toy worlds |
| Claymation | Handmade clay | Gentle comedy, small sets |
| Cel-shaded 3D | Anime characters built in 3D | Action, older kids |
| Watercolor Storybook | A picture book in motion | Bedtime, fables |
| American Comic | Inked lines, bold colour | Superhero, comedy |

All 21 styles: visual styles one by one. The sample Rocket Pup shows a 3D cartoon made this way.

## A cast that comes back

SceneMixer gives every character one sheet, one voice and a set of states (a costume, a hat, wet fur) drawn once. Every episode uses them. Non-human characters work too: animals, robots, fruit, described plainly. Non-human characters.

Recurring places are drawn once as well: the treehouse, the classroom, the station hall. Same room in every episode.

## Writing a season

- One premise, many small stories. A series is easier to keep consistent when each episode is one problem solved.
- Upload the season at once and it is split into episodes; or add episodes later to the same project. Add episodes.
- Type: Short Drama for 1–4 minute episodes with a hook; TV Series for longer, calmer ones.

## For a channel

Kids' channels have their own rules on platforms (made-for-kids settings, ads, comments). The production side is here; the publishing side is in running a kids' cartoon channel. For single cartoon videos rather than a series, see the cartoon video generator.

## Limits

SceneMixer refuses famous cartoon characters or studio looks by name; design your own. Very fast slapstick with many bodies in contact is hard for current video models; one action per shot works best. Writing action.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make a whole cartoon series?

Yes. SceneMixer draws the cast and places once and generates every episode from them.

### Which cartoon styles are there?

3D looks such as Family 3D Animation and 3D Toy Cartoon, and 2D looks such as Watercolor Storybook, American Comic and Claymation, among 21 presets.

### Can I change the style later?

No; SceneMixer fixes the style after step 1. Start a new project with the same script for another look.

### Can animals talk?

Yes. Give each animal a voice; cartoon animals can have hands and expressions.

### How long can an episode be?

Up to 240 seconds in Short Drama, 300 in TV Series.

### Can I use a famous cartoon character?

No. Known characters are refused; design your own.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI TV Show Generator: Seasons, Ensembles and Episode Arcs

URL: https://scenemixer.com/ai-tv-show-generator/
Make a TV show with AI: TV Series type with 300-second episodes, an ensemble cast drawn once, several story threads, a season split from one script, and a pilot first.

# AI TV Show Generator

Quick answer: Set the type to TV Series and SceneMixer reads your season into episodes of up to 300 seconds, with room for a scene to build and endings that need not all be cliffhangers. The whole ensemble, the recurring sets and the props are drawn once and shared by every episode. Make the pilot first, watch it, then the rest.

Brief for AI Create: A TV series about a night-shift team at a small-town radio station; four regulars, a caller each episode, one long mystery under the season; eight episodes.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## TV Series versus short drama

| | Short Drama | TV Series |
| Episode length | Up to 240 seconds | Up to 300 seconds |
| Opening | Conflict in the first seconds | Fifteen to twenty seconds of establishment |
| Ending | A hook every time | A hook, a feeling or an unfinished decision |
| Cast | Two or three leads | An ensemble with several threads |
| Platform | Vertical feeds | YouTube, TV-shaped screens |

Short drama, series or film explains what the type changes in the script.

## An ensemble that stays itself

Ensemble shows have more faces to keep straight. Each regular is one sheet with a specific face, a fixed outfit and states for costume changes; give them different builds and colours so they read at a glance. Leads, supporting and voice-only characters and describing a face.

SceneMixer draws recurring sets (the studio, the diner, the precinct) once, and each becomes the set for every episode.

## Planning a season

- Write or upload the whole season; it is split into episodes and keeps preparing later ones in the background.
- Storyboard and render the pilot; fix faces and voices now, while one episode depends on them.
- Run the rest, several at once; each episode is joined into its own file. Multi-episode series.
- Make a trailer and a cover for the season. Pilot and pitch trailer.

## Budget a season

Eight five-minute episodes are 2,400 seconds of video. At 10 credits a second on the default tier, plus 1 credit a second to join, that is the main line of the budget; reading and storyboards are small next to it. Higher tiers multiply the per-second rate. Which tier.

## Limits

Crowd scenes with many speaking extras are hard; background people are scenery. Very long single takes are not possible. SceneMixer refuses real actors and famous shows.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make a TV show?

Yes. SceneMixer's TV Series type splits a season into episodes of up to 300 seconds with one shared cast and set of locations.

### How many characters can a show have?

SceneMixer allows up to 100 characters per project, though a smaller regular cast is easier to keep distinct.

### Can I make a pilot first?

Yes. Render episode 1 fully, then parse and render the rest.

### Do episodes have to end on cliffhangers?

Not in TV Series; an ending can be a feeling or a decision left open.

### Can I add episodes later?

Yes, to the same project, with the same cast.

### Is it landscape?

Your choice in step 1: 16:9 for TV-shaped screens, 9:16 for phones.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# Storyboard to Video: An Editable Shot List That Renders

URL: https://scenemixer.com/storyboard-to-video/
Storyboard to video with AI: the script becomes a shot-by-shot storyboard with seconds, framing and lines; edit any shot in plain text, then render it.

# Storyboard to Video

Quick answer: In SceneMixer the storyboard is text: numbered shots, each with seconds, framing, who is in frame, what they do and say. It is written from your script, you edit it in plain words, and the video is generated from exactly what it says. Generating the storyboard for your first drama's first episode is free; editing it costs nothing; only rendering video is charged. Drawn storyboard panels cannot be uploaded; a written shot list can be pasted as the script.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a storyboard here looks like

A segment's storyboard has a short header (where, time, weather) and then the shots:

| Shot | What it says |
| [Shot 1] 3s | Wide shot. Rain on the window of an all-night diner; @Mara alone in the booth by the door. |
| [Shot 2] 4s | Medium shot on @Mara. @Mara (quietly): You said you'd never come back here. (voice: Mara) |
| [Shot 3] 3s | Over-the-shoulder from @Mara to @Theo in the doorway, shaking off the rain. No dialogue. |

Names are linked chips from the cast, so the right face is used; seconds add up to the segment length. The segment header.

## Edit, then render

- Change words, not pictures. Rewrite framing, action or a line, save, and only the shots you touched are re-read. What you can change.
- Ask AI to revise. Describe a change for one segment or the whole episode. AI edit.
- Add or remove a segment. Insert or delete a segment.
- Restore an earlier text. Text history.

## If you already have a storyboard

You cannot upload drawn panels or an animatic. You can paste a written shot list or a script with shots described: the read turns it into scenes and the storyboard is rebuilt in the format above, keeping your order and your lines where they are clear. Then edit any shot that came out differently from your plan. Concept art for characters can be uploaded as reference (a member feature).

## Why text instead of panels

Video models read words. A storyboard written in the model's language, with each person linked to a reference sheet and each position stated, is what keeps a face, a seat and a prop the same across cuts. A drawing would still have to be described. Write positions the world can read and shot-to-shot continuity.

## Cost

On SceneMixer, the first episode's storyboard on your first drama is free; later storyboards are charged by length and shown on the button. Editing is free. Video is 10 credits a second on the default tier and is charged only when you render a segment.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can I turn a storyboard into a video with AI?

Yes. Each segment of the written storyboard is rendered as video with the cast it names.

### Can I upload my own storyboard drawings?

No. Paste a written shot list or script instead; SceneMixer rebuilds the storyboard as text you can edit.

### Is editing the storyboard free?

Yes. Only rendering video is charged.

### Can I set shot lengths?

Yes. Each shot in SceneMixer has seconds; they add up to the segment length, up to 15 or 30 seconds.

### Can I control camera angles?

Yes, by writing the framing and movement in the shot.

### Can AI rewrite the storyboard for me?

Yes. Describe the change and SceneMixer's AI edit applies it to a segment or the whole episode.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Video Generator With Dialogue: Lines, Voices and 15 Languages

URL: https://scenemixer.com/ai-video-generator-with-dialogue/
An AI video generator with dialogue: characters speak your lines in the same generation, a voice per character, 15 dialogue languages, redub and subtitles.

# AI Video Generator With Dialogue

Quick answer: In SceneMixer, dialogue is spoken in the video itself: each line in the storyboard carries its speaker and a voice, and the model performs it with the picture, mouths moving, in the language you set (any of 15 languages). No separate dubbing pass and no extra charge per line; video is 10 credits a second on the default tier. A line can be changed or redubbed afterwards.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## How a line becomes speech

Write the line in the script with the speaker's name. The read assigns it to that character; the storyboard places it in a shot with a delivery note and the character's voice tag; the generation performs it. Write dialogue so it is picked up and dialogue and sound in one generation.

| You write | On screen |
| MARA (quietly): Read it before you say no. | Mara says it, softly, in her voice |
| THEO (on the phone): I'm outside. | Theo's voice from a phone; he need not be in frame |
| NARRATOR (voice-over): That was the last winter. | A narrator over the picture |

## Voices that stay the same

- Described in words (default, free): age, weight, pace; performed fresh each shot.
- A preset from the library (free to match and to change): used as an audio reference on every line, so the voice holds.
- Your own clip (members): one you have the right to use.
Pick character voices and keep a voice the same.

## Languages

In SceneMixer, the dialogue language is set in step 1 and every line is spoken in it, with shot timing computed from that language's speaking rate. The same story can be made again in another language for another market. Choose the dialogue language and native dialogue versus dubbing.

## Fixing a line after the video is made

- Change the words: redub with lip-sync keeps the picture and changes the line. Change a line.
- Change the voice: Change a voice in a finished clip.
- A line you never wrote: what to do.
- Subtitles: matched to your script's words; burned in or SRT, on accounts that have made a purchase.

## Limits

Two people talking over each other and long monologues in one shot are weak spots; split long speeches across shots. SceneMixer cannot clone real people's voices.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI videos have dialogue?

Yes. SceneMixer generates the lines with lip movement, spoken by the character who says them.

### Is there an extra charge for dialogue?

No. Video is charged per second on SceneMixer whether or not anyone speaks.

### Which languages can characters speak?

Any of 15 languages.

### Can each character have a different voice?

Yes. Each character has its own description in SceneMixer, and can be matched to a preset or, for members, an uploaded clip.

### Can I change a line after rendering?

Yes. Redub the clip in SceneMixer with the new line; the picture stays.

### Can I clone a celebrity's voice?

No. SceneMixer does not support real people's voices, which need their permission.

Updated November 3, 2026


---

# Free AI Drama Generator: What You Can Make Without Paying

URL: https://scenemixer.com/free-ai-drama-generator/
Start an AI drama for free: welcome credits, 15 free setup images, a free first-episode storyboard, three free posters and a free video preview. What is paid, plainly.

# Free AI Drama Generator

Quick answer: A new SceneMixer account starts with 50 credits and, on your first drama, 15 free setup images, a free storyboard for episode 1, a free set of three posters and one free video preview of about 5 seconds with your own cast. That is enough to see your characters, your shot list and a real clip before you pay anything. Generating full episodes of video is paid, at 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Everything that is free

| Free item | How much | Conditions |
| Welcome credits | 50 credits | On sign-up; spend them on anything |
| Setup images | 15 per account | Characters, places, props on the default image model; the script must have been read |
| Storyboard | Episode 1 of your first drama | One first drama per account |
| Promo posters | One set of three options | Your first drama; drawn from your leads |
| Video preview | About 5 seconds, once | One shot of your own episode at 480P |
| Editing text | Unlimited | Storyboard edits, summaries, settings cost nothing |

Details and edge cases: what you get for free and free sheets are per account.

## A free first session, step by step

- Sign up; the welcome credits land at once.
- Paste a short script or write one with AI Create at a short length; the welcome credits cover reading a short script. Reading is 4 credits per 1,000 characters, English counted lighter.
- Generate the cast and places with the free setup images.
- Generate episode 1's storyboard, free, and read it.
- Use the free preview: one shot of your episode, with your cast speaking.
- Look at the three free posters in Cover art.

## What is paid

Video, mainly. On SceneMixer, each second of finished video is charged at the tier's rate (10 credits a second on the default tier, ≈ $0.06 at the lowest credit price), plus 1 credit a second to join an episode. Reading longer scripts, AI writing, storyboards after the first and images beyond the free ones are charged too, each shown on its button before you press it. Why video costs more than everything else.

## Ways to get more free credits

- Share your first video: : 100 credits, once, within 14 days of your first clip. Share your first video.
- Invite friends: : 20% of what they pay, as credits, for 12 months. Referral program.

## Free videos carry the badge

Videos from free SceneMixer accounts carry a small AI badge in the corner; members can switch it off. Anything shared through a share link carries it either way. AI video without a watermark. What you generate with free credits is yours to use, commercially included, under the terms. What you get to keep.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Is there a free AI drama generator?

You can start one free on SceneMixer: welcome credits, free setup images, a free first storyboard, free posters and a short free preview. Full episodes of video are paid.

### How many free credits do I get?

SceneMixer gives 50 on sign-up, plus 100 for sharing your first video within 14 days.

### Can I make a whole episode for free?

Not the video. The cast, the storyboard and a few seconds of preview are free; each second of episode video is charged.

### Do I need a credit card to start?

No. Sign up and the free items are there; you pay only when you buy credits or a plan.

### Is the free preview my own story?

Yes. It is one shot of your episode, with your characters, about 5 seconds long, rendered by SceneMixer.

### Can I use free videos commercially?

Under the SceneMixer terms, output made with free credits is yours to use, commercially included, with the AI badge on unless you are a member.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Video Generator Without a Watermark: How the Badge Works

URL: https://scenemixer.com/ai-video-generator-no-watermark/
Make AI videos without a watermark: members' videos are badge-free by default, the switch applies to new clips, downloads keep your setting, shared copies keep the badge.

# AI video without a watermark

Quick answer: SceneMixer adds a small AI badge to the corner of generated video unless you turn it off, and turning it off is a membership benefit: members' new videos are badge-free by default. Downloads keep whatever the switch said when the clip was made. Copies sent through share links always carry the badge. Plans start at $7.99 a month.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Who gets a clean file

| Account | New videos | Download | Shared link |
| Free | Badge on | Badge on | Badge on |
| Member, default | No badge | No badge | Badge added to a shared copy |
| Member, badge switched on | Badge on | Badge on | Badge on |

The badge is a small mark in the bottom-right corner, not a full-frame watermark. The AI badge switch.

## Where the switch is

In the episode editor, next to the model picker: a pill that reads AI Badge or No Badge. On a phone it is at the bottom of the model sheet; the account settings have it too. It is read when a segment is generated, so it changes new clips only; to clean an old clip, regenerate it under the new setting.

## Make a whole episode clean

In SceneMixer, the episode file is joined from the clips as they are, so mixed clips give a mixed episode. Set the switch before you generate, render every segment under it, then join. Download the original-quality file, not the play version. Original versus play version.

## Why shared copies keep it

A shared link travels without you: reposted, clipped, embedded. The badge tells whoever meets it that it was made with AI, and many places' labelling rules expect that. So the share path always uses a badged copy, while your own downloads follow your setting. Why shared videos carry the badge.

## Your labelling duty does not go away

Removing the visual badge does not remove the rules of the platforms you post to. TikTok, YouTube and others have their own AI-content labels; use them. SceneMixer's switch panel says it plainly: once removed, labelling is your responsibility. Do AI videos need a label?

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can I make AI videos without a watermark?

Yes, as a SceneMixer member: new videos are badge-free by default and downloads have no badge.

### Do free accounts get a watermark?

Free SceneMixer accounts' videos carry a small AI badge in the bottom-right corner.

### Can I remove the badge from a video I already made?

Regenerate that clip with the badge off; the old version stays in SceneMixer's Version History.

### Why does my shared link show the badge?

SceneMixer's shared copies always carry it; use Download for a clean file.

### How much is a membership?

SceneMixer's plans start at $7.99 a month and include monthly credits.

### Is the badge a big watermark?

No. SceneMixer's badge is a small mark in one corner.

Updated November 3, 2026


---

# AI Anime Video Generator: Anime Shorts With Consistent Characters

URL: https://scenemixer.com/ai-anime-video-generator/
Generate anime videos with AI: anime and webtoon looks, characters drawn once and kept in every shot, fights in one move per shot, dialogue in Japanese and more.

# AI Anime Video Generator

Quick answer: Pick an anime look in step 1 (Anime, Watercolor Anime, Painterly Anime, Cel-shaded 3D or Korean Webcomic), write or paste the story, and SceneMixer draws each character once in that style and keeps them in every shot. Scenes are storyboarded shot by shot, with dialogue spoken in Japanese or any of the other dialogue languages. Video is 10 credits a second on the default tier.

Brief for AI Create: An anime short: a delivery girl on a flying bicycle has one hour to get a letter across a city of floating islands before the bridges close; one rival, one storm, a last-second landing.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The anime looks

| Style | Looks like | Suits |
| Anime | Clean lines, flat colour, expressive eyes | Most series; action and comedy |
| Watercolor Anime | Soft washes, light, sky | Romance, slice of life |
| Painterly Anime | Thick painted light and detail | Cinematic shorts, city at night |
| Cel-shaded 3D | Anime characters in 3D | Fights, mecha, camera moves |
| Korean Webcomic | Webtoon colour and faces | Romance fantasy, office drama |

The sample Rust City Knockout is a painterly anime boxing story made this way. Getting a consistent anime look.

## Characters that stay on model

Anime characters are recognised by hair, eyes and one costume. Each is drawn once in SceneMixer as a reference sheet in the chosen style, and every shot uses it, so the hair colour and the uniform do not drift. A transformation, a battle-worn outfit or a school uniform versus casual clothes is a state of the same character. States.

## Fights and action

One clear move per shot: a dash, a block, a hit, a reaction. Power-ups are light and effect around the character. Wounds stay off screen. Anime fight scenes has a shot-by-shot example.

## Short video or series?

For one anime short, choose Film and keep it to a few scenes. For a series with episodes, see the anime series generator; for a channel, an anime series for YouTube; for manga you wrote, manga to anime.

## Limits

SceneMixer refuses existing anime characters, studios' house styles by name and famous series; design your own world. Crowded battles and long choreographed fights are weak spots of current video models.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make anime videos?

Yes. Choose an anime style in step 1; SceneMixer draws characters, places and every shot in it.

### Will my anime character look the same in every shot?

Yes. SceneMixer uses one reference sheet per character for every segment.

### Can the characters speak Japanese?

Yes. Japanese is one of SceneMixer's dialogue languages; set it in step 1.

### Can I make fight scenes?

Yes, written one move per shot, with no gore.

### Can I use characters from a famous anime?

No. Known characters are refused; design your own.

### Vertical or landscape?

Either: 9:16 for Shorts and TikTok, 16:9 for YouTube.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Cartoon Video Generator: 3D, 2D and Claymation Shorts

URL: https://scenemixer.com/ai-cartoon-video-generator/
Make cartoon videos with AI: 3D toy, family 3D, claymation, comic and storybook looks; talking animals and objects; one cartoon short or a run of them.

# AI Cartoon Video Generator

Quick answer: Choose a cartoon style, describe your characters (people, animals, a talking teapot) and paste or write the story; SceneMixer draws the cast once in that style and renders the cartoon shot by shot with voices and music. Single shorts and whole series are made the same way. Video is 10 credits a second on the default tier; your first drama's setup images and first storyboard are free.

Brief for AI Create: A two-minute cartoon: a grumpy garden gnome and an overexcited sprinkler must work together when a hedgehog gets stuck in the pond; slapstick, no dialogue except grunts.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Cartoon looks

| Style | Character feel | Try it for |
| 3D Toy Cartoon | Matte, rounded, toy-like | Toy worlds, preschool |
| Family 3D Animation | Big eyes, warm light | Family comedy |
| Claymation | Handmade clay, small sets | Gentle comedy, stop-motion feel |
| American Comic | Inked lines, bold colour | Superhero, satire |
| Watercolor Storybook | A picture book moving | Fables, bedtime |
| Chinese Painted Style | Brush lines, ink tones | Folk tales, legends |

Claymation short, 3D cartoon movie.

## Who can be a character

Anything you can describe: a person, a dog, a robot, a fruit, a sock. Cartoon animals can have hands and faces; realistic animals stay on four legs. Give each a name and a short, exact description. Non-human characters and talking animals.

## Comedy timing

Cartoons live on timing, and timing is in the shot seconds. A two-second reaction after a three-second set-up lands; a long shot with three gags in it does not. Write one gag per shot and let SceneMixer's storyboard hold the beat. Shot seconds and timing.

## Sound

Voices, ambient sound and effects are generated with the picture; a score is added under the episode, with a switch for a version without it. Background music.

## One short or a series

A single cartoon video is one project with one episode. For recurring characters across many episodes, see the cartoon series generator. For children's stories, kids' story videos.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make cartoon videos?

Yes. Pick a cartoon style; SceneMixer draws the cast once in it and renders every shot in the same look.

### Which cartoon styles are there?

3D Toy Cartoon, Family 3D Animation, Claymation, American Comic, Watercolor Storybook and more among 21 visual styles.

### Can objects and animals talk?

Yes. Give them a voice; they speak their lines in the video.

### Can I make a cartoon without dialogue?

Yes. Write shots with no dialogue; SceneMixer keeps ambient sound and music.

### Can I copy a famous cartoon's style?

No famous characters or studio styles by name; SceneMixer's built-in looks are yours to use.

### How long can a cartoon be?

A few seconds to many minutes, built from segments and joined.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Kids Story Video Generator: Gentle Stories, Safe Looks

URL: https://scenemixer.com/ai-kids-story-video-generator/
Make kids' story videos with AI: bedtime stories, fables and lessons in storybook or 3D cartoon looks, a calm narrator, short episodes and strict content rules.

# AI Kids Story Video Generator

Quick answer: Write or paste a children's story, choose a gentle look such as Watercolor Storybook or Family 3D Animation, and SceneMixer turns it into a short video with a narrator or speaking characters, soft music and the same characters every episode. Content rules already exclude violence and anything unsuitable. Video is 10 credits a second on the default tier.

Brief for AI Create: A bedtime story: a small cloud is afraid of thunder until an old lighthouse teaches it to count between the flash and the rumble; three minutes, a calm narrator, ends with the cloud asleep.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Choose a gentle setup

| Setting | Choose | Why |
| Visual style | Watercolor Storybook, Family 3D Animation, 3D Toy Cartoon | Soft, warm, readable for small children |
| Type | Short Drama for 1–4 minute episodes | Short enough for attention spans |
| Aspect | 16:9 for TV and tablets, 9:16 for phones | Where it will be watched |
| Voices | A calm narrator preset; distinct, friendly character voices | Matched from the library, free |
| Music | On, under the voices; off for a quiet version | The Music switch |

## Writing for children

- One problem, one lesson, one ending. Bedtime stories end calm; lessons end with one sentence to remember.
- Repeat the hero's look and name. Children recognise characters by colour and shape; keep them simple.
- Say it, show it. The narrator says what happens while the picture shows it.
- Nothing scary on screen. A storm is noise and a dark window, not danger.
Bedtime story series and educational cartoons.

## Story sources

Your own stories, and public-domain fables and fairy tales with your own character designs. SceneMixer refuses famous children's characters. A story starring your own child as a character is possible with an invented, cartoon look; photos of real children are not a good idea and realistic photo faces are refused by the video provider's face check. Fairy tale videos.

## Publishing for kids

Platforms treat content made for children differently (comments, ads, recommendations). Set the made-for-kids options where you post. Production is here; channel strategy is in running a kids' cartoon channel.

## Cost of a bedtime story

A three-minute story is 180 seconds of video at 10 credits a second on the default tier, plus 1 credit a second to join. The cast is drawn once and reused, so episode 2 costs only its video. Your account's first 15 setup images are free.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make children's story videos?

Yes. Write or paste the story, pick a gentle look, and SceneMixer renders it with a narrator or speaking characters.

### Is the content safe for kids?

SceneMixer blocks violence, sexual content and other unsuitable material for everyone; write the story gently and review every episode before you publish.

### Which style is best for little children?

Watercolor Storybook, Family 3D Animation or 3D Toy Cartoon are soft and easy to read.

### Can I make my child the hero?

As an invented cartoon character, yes. Real photos of children are not recommended, and SceneMixer refuses photo-like faces.

### Can I add a narrator?

Yes. Write the story as voice-over lines and match the narrator to a calm preset voice.

### How long should a kids' story be?

Two to five minutes is a comfortable length for one sitting.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Horror Movie Generator: Dread Without Gore

URL: https://scenemixer.com/ai-horror-movie-generator/
Make horror with AI: dark looks that stay readable, sound before sight, possessions and monsters as character states, and scares that work within the no-gore rule.

# AI Horror Movie Generator

Quick answer: Horror on AI video is built from the ordinary going wrong, sound before sight, and a cut at the worst moment, none of which needs gore, which is blocked. In SceneMixer, choose Film for a single short or Short Drama for a scare series, pick a look that can hold darkness, write a light source into every dark shot, and make the monster or the possession a state of the character.

Brief for AI Create: A horror short: a night-shift guard in an empty care home keeps hearing the call bell from room 14, sealed for a year; at 3 a.m. he opens the door; slow dread, one scare, no blood.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Looks that hold darkness

| Style | Effect |
| Black & White Film | Classic, shadows read as shapes |
| Shot on Film | Grain and warm practical light; found-footage feel |
| Digital HD | Clean modern horror; needs strong practical lights |
| Painterly Anime | Atmospheric anime horror |

Sets with built-in mood: Gothic Dark, Hospital Ward, a rainy alley. The sample Crimson Vow plays in a rain-soaked alley and church. Scene and costume styles.

## Scares that work within the rules

| Scare | How to write it |
| The sound first | Put the creak, breathing or bell in the shot's ambient line before anything is seen |
| The wrong detail | A door open that was shut; a chair turned to the wall |
| The reveal | A close-up on the reaction, then the thing, for two seconds |
| The possession | A second state of the same character: eyes, posture, voice |
| The aftermath | Show the room after, not the act |

Violence is shown before and after, not as wounds; SceneMixer refuses bodies and blood. Content rules.

## Darkness the model can render

A shot described as pitch black renders as murk. Name the light: a phone torch, a flickering tube, lightning through a window, a candle. The darker the scene, the more important the one light. Say what is there, not what is not.

## Monsters and ghosts

A creature is a character with its own sheet: describe its shape and texture plainly, without naming a famous monster. A ghost is a character with a ghostly state; the house is a location drawn once. Ghost and haunted house and folk horror.

## Length and cost

Horror shorts work at three to eight minutes; scare series at one to two minutes an episode with a hook. Every second is 10 credits a second on SceneMixer's default tier. Make an AI horror short has a full shot-by-shot example.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make horror movies?

Yes, horror built on dread, sound and reveals; SceneMixer blocks gore and graphic violence.

### Why are my dark scenes muddy?

Name a light source in every dark shot; pitch black renders as murk.

### Can I have a monster?

Yes, as a character with its own sheet on SceneMixer, described plainly and not copied from a famous film.

### Can a character be possessed?

Yes. Make the possessed look a state of the same character.

### Is blood allowed?

No. Show the moment before and the room after.

### Which style is scariest?

Black & White Film and Shot on Film hold shadows well; Digital HD needs strong practical lights.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Trailer Generator: Trailers for Films, Books, Series and Games

URL: https://scenemixer.com/ai-trailer-generator/
Make a trailer with AI: a 60–120 second script in three acts, the same cast as your story, a narrator, rising cuts, a score, and a poster to go with it.

# AI Trailer Generator

Quick answer: A trailer is a short script of moments, not the whole story. Write 60 to 120 seconds in three beats (the world, the disruption, the rise) and a final button, set the type to Film, and SceneMixer draws the cast once, storyboards each moment and renders the trailer with a narrator and a score. If the story already exists as a project, reuse its cast so the trailer and the film match.

Brief for AI Create: A 90-second trailer for a mystery series: a village where every clock stopped at 4:12 on the same night twenty years ago; a returning daughter, a narrator, the title spoken at the end.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A 60–120 second trailer script in three beats and a button, or the cast of an existing project |

## The shape of a trailer

| Beat | Seconds | What it does |
| The world | 15–25 | Calm. Who, where, what is normal. |
| The disruption | 20–30 | The thing that changes everything, in one clear image. |
| The rise | 25–45 | Shorter shots, bigger moments, the stakes out loud. |
| The button | 5–10 | Silence, one last line, then the title. |

A full shot-by-shot example is in making an AI movie trailer.

## Trailers for different things

| Trailer for | What to show | Guide |
| A film or short | Three or four key scenes with the real cast | Movie trailer |
| A TV pilot pitch | The world, the hero, the hook, the conflict | Pitch trailer |
| A book | Mood and premise without spoiling the ending | Book trailer |
| A game or tabletop campaign | The world and its stakes, original characters | Write it as a short script |
| Your own series | The season's best moments | Reuse the series' cast |

## Same faces as the real thing

In SceneMixer, a trailer made in a separate project can pick its characters, places and voices from your library, so the hero of the trailer is the hero of the film. Reuse characters across projects.

## Titles and text

Video models are unreliable at drawing legible words, so let the narrator speak the title, or add it in an editor afterwards. For the poster, Cover art paints the title into a key visual. AI movie poster.

## Cost

A 90-second trailer is 90 seconds of video: on the default tier that is 990 credits including the join, before retakes. Trailers have many short shots, so plan a retake or two for the key moment.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make a movie trailer?

Yes. Write a short trailer script; SceneMixer casts, storyboards and renders it with a narrator and score.

### How long should a trailer be?

60 to 120 seconds: three beats and a final button.

### Can the trailer use my film's characters?

Yes. Pick them from your library in the trailer project.

### Can it write the title on screen?

Not reliably. Have the narrator say it, add it in an editor, or use the poster for the title.

### Is there music?

Yes. SceneMixer adds a score under the voices; it can be switched off.

### Can I make a trailer for someone else's film?

Only original material; SceneMixer refuses famous films and characters.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Movie Poster Generator: Key Art From Your Own Cast

URL: https://scenemixer.com/ai-movie-poster-generator/
Generate a movie poster with AI: three key-art options painted from your own characters, the title drawn in, a tagline and episode numbers on top, portrait and landscape.

# AI Movie Poster Generator

Quick answer: Open step 2 of a project, pick Cover art, type the title, choose the lettering, the genre mood and up to two characters, and press Generate 3 options. Three posters are painted from your leads' reference art with the title in the picture; you pick one, and the tagline and episode numbers are laid on top. Each option is priced as one image on your image tier; your first drama's first set is free.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | Your project's title, a genre mood and up to two of your own cast; three poster options per generation |

## What you get

- Three key-visual options with your title painted in, built from your characters' sheets, so the faces are your series' faces.
- A tagline and episode numbers set by the site on top, editable any time without repainting.
- Portrait 9:16 and landscape 16:9: the project's own shape first; the other is redrawn from the one you pick at no extra cost.
- Episode covers with EP 1, EP 2 and so on.
- The share card: the cover becomes the picture people see when you share the series.

## The choices that shape it

| Choice | What it changes |
| Title | The words painted into the poster |
| Lettering | The style of the title type |
| Genre mood | Light, colour and composition: romance, thriller, fantasy and others |
| Cover characters | Up to two leads, from their reference sheets |
| Tagline | A line set on top by the site, not painted |
| Image tier | Which image model paints it; set in Settings before generating |

Make a promo cover poster covers every option.

## Why from your own cast

A poster generated from a prompt alone shows someone who is not in your film. SceneMixer's cover art starts from the reference sheets the video uses, so the poster, the thumbnail and the episodes show the same people. At least one lead needs a character image first.

## Price

Each option costs one image on the image tier you have selected (7 credits each on the default tier). Your account's first project gets its first set of three free; on accounts created since late October 2026 that set is made for you once your leads' sheets are ready. Later sets (a new title, other characters, another mood) are charged.

## Limits

Ad projects have no cover art. Posters of real actors, famous films or franchises are refused. Fine print such as credits blocks is not painted reliably; add it in a design tool if you need it.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make a movie poster?

Yes. Cover art paints three poster options from your own characters with the title drawn in.

### Is the poster free?

Your first drama's first set of three is free; later sets are priced as images.

### Can I get both vertical and horizontal posters?

Yes. Pick one and SceneMixer redraws the other shape from it at no extra cost.

### Can I change the tagline later?

Yes. It is laid on top by SceneMixer and can be changed without repainting.

### Can I make a poster without a project?

No. SceneMixer paints the poster from the project's character sheets, so it needs at least one lead image.

### Can I use an actor's face?

No. Real people and famous characters are refused.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# Script to Movie: Turn a Screenplay Into an AI Film

URL: https://scenemixer.com/script-to-movie/
Turn a screenplay into a movie with AI: keep scene headings and character cues, read a feature-length script in one project, cast once, render act by act in Film mode.

# Script to Movie

Quick answer: Export your screenplay as plain text or Word, upload it, choose Film, and SceneMixer reads scene headings, character cues and parentheticals into a cast, a set of locations and acts. One project reads up to 300,000 characters of English, enough for a feature screenplay. Each act is storyboarded and rendered with spoken dialogue; you watch one before paying for the next.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Prepare the screenplay

| Element | Keep it? | Tip |
| Scene heading (INT./EXT., place, time) | Yes | One name per place throughout |
| Character cue | Yes | One spelling per character |
| Dialogue | Yes | For exact lines, NAME (note): line on one line |
| Parenthetical | Yes | Becomes the delivery note |
| V.O. and O.S. | Yes | Written as (voice-over), (on the phone) |
| Transitions | Optional | The storyboard plans its own cuts |

Files: .txt, .docx, .md or paste. PDFs are refused; Fountain is plain text, rename it to .txt. Scene headings and character cues.

## What the read gives you

A summary, a cast list with descriptions and voices, every location, the props that matter, and the script divided into acts (Film) or episodes (Series). SceneMixer reads a long screenplay in two passes: the cast and places first, then the structure, with each act's outline expanded when you open it. Read the summary before going on: a misread name is cheapest to fix here. Checking the summary.

## From page to shots

Each act becomes a storyboard of numbered shots with seconds, framing and the lines from your page. Dialogue is kept; long action paragraphs are condensed to what can be seen. Edit any shot in plain text before rendering. Storyboard to video.

## Script to movie, or script to video?

For a screenplay meant as a film, Film type keeps your pacing: no forced hooks, room for silence. For short vertical episodes from a script, see script to video; for what Film mode changes and costs, the AI movie generator.

## Budget by the page

A screenplay page runs about a minute on screen. So a ten-page short is roughly 600 seconds of video at 10 credits a second on SceneMixer's default tier, plus 1 credit a second for the join, and a 100-page feature ten times that. Start with the act you care most about. Can AI make a full-length movie yet?

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI turn a script into a movie?

Yes. Upload the screenplay; it is read into a cast and acts, storyboarded and rendered with dialogue.

### Which screenplay formats work?

Plain text, Word and Markdown. Export from your screenwriting app; rename Fountain to .txt; PDFs are not accepted.

### How long a screenplay can I upload?

SceneMixer takes up to 300,000 characters of English in one project, enough for a feature.

### Are my exact lines kept?

Dialogue is kept; for word-for-word lines, put the name and line on one line.

### Can I change shots before rendering?

Yes. Every shot in SceneMixer is editable text.

### How much does a feature cost?

Mostly seconds of video; a page is about a minute, so budget by the page and render one act first.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# Text to Movie: From a Paragraph to a Film With AI

URL: https://scenemixer.com/text-to-movie/
Text to movie with AI: from one line or a paragraph to a written script, a cast and a rendered film. How text length maps to episodes and minutes, and what it costs.

# Text to Movie

Quick answer: Text-to-video tools turn a prompt into one clip. Text to movie needs more: a script, characters who come back, places that stay put. In SceneMixer you type a line or a paragraph, AI Create writes a script at the length you pick (Mini to Long), the read turns it into a cast and episodes, and each episode is storyboarded and rendered. Video is 10 credits a second on the default tier.

Brief for AI Create: An old watchmaker repairs one clock each night for a stranger who never shows his face; on the last night the clock is his own.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## From text length to film length

AI Create offers five script lengths, counted in Chinese-equivalent characters (English uses more letters for the same story, and the price already allows for it). The read then splits the script into episodes of a minute or two each:

| Length option | Script size | What it usually becomes |
| Mini | 1,000 | One scene: a test |
| Brief | 2,500 | One short episode or two |
| Short | 5,000 | A handful of episodes: the usual first choice |
| Medium | 10,000 | A short season |
| Long | 20,000 | A full series |

More text means more episodes rather than longer ones. Options too long to read afterwards are hidden from the menu. Choosing script length and cost.

## Text-to-video versus text-to-movie

| | One prompt, one clip | Text to movie here |
| Characters | New faces every prompt | Drawn once, used in every shot |
| Places | Re-imagined each time | Drawn once as a set |
| Dialogue | Rare or none | Every line spoken by its character |
| Length | Seconds | Segments joined into episodes |
| Control | Rewrite the prompt | Edit the script and each shot |

## Write a better line

The brief matters more than the length: a person, a want, an obstacle and a turn. Inspire Me suggests one if you have nothing. Then read the script it writes and change what you do not like before the read. Writing the one-line brief.

## Settings before the read

Type (Short Drama, TV Series or Film), aspect, visual style, region and dialogue language are set in SceneMixer's step 1 and fixed after the read. Film type gives slower, quieter pacing. Style, region and dialogue.

## Costs, in order

Writing the script and reading it are priced by length and shown on their buttons. Your first drama's setup images (15) and first storyboard are free. Video is the main cost, per second at the tier's rate, plus 1 credit a second to join an episode.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make a movie from text?

Yes. SceneMixer turns a line or paragraph into a script, a cast and episodes, rendered shot by shot.

### How much text do I need?

One line is enough for AI Create; for your own text, at least a paragraph.

### How long will the movie be?

It depends on the script length you pick; more text gives more episodes of a minute or two each.

### Can I edit the script before it is made?

Yes. Read and edit the written script, then each shot in the storyboard.

### Is text-to-movie the same as text-to-video?

No. Text-to-video makes one clip; SceneMixer's text to movie keeps a cast and places across many shots and episodes.

### What does it cost?

Writing and reading are small and shown on their buttons; video is charged per second.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Reddit Story Video Generator: Acted, Not Just Narrated

URL: https://scenemixer.com/ai-reddit-story-video-generator/
Make Reddit-style story videos with AI: your own story or one you have permission to use, acted by consistent characters or narrated, vertical for TikTok and Shorts.

# AI Reddit Story Video Generator

Quick answer: Reddit-style story videos usually put a voice reading a post over unrelated footage. SceneMixer acts the story instead: paste a story you wrote (or have the author's permission to use), and it is read into characters and scenes, with the poster as a narrator if you want one, in 9:16 for TikTok, Shorts and Reels. Names and details that identify real people should be changed.

Paste a story like: AITA for refusing to give my sister the wedding dress our grandmother left me? She says it should go to whoever marries first… (your own, rewritten as scenes).

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A story you wrote or have permission to use, pasted as text; acted by characters in 9:16 |

## Narrated or acted

| Format | What it looks like | How to set it up |
| Narrated | The poster's voice telling it, over consistent scenes | Narrator voice-over lines; scenes show what is said |
| Acted | The people in the story speak | Rewrite the post as dialogue with a speaker per line |
| Hybrid | The poster opens and closes; the conflict is acted | Voice-over at the start and end |

Acted versions keep viewers longer because something happens on screen. Turn a Reddit story into an AI drama has a full worked example.

## Whose story it is

- Your own posts are yours to adapt.
- Someone else's post belongs to its author; ask first.
- Real people in it get invented faces and changed names; nothing that identifies or defames them. Content rules.
- SceneMixer keeps usernames and screenshots out of the video.

## A format that fits the feed

- 9:16 Portrait, Short Drama type: the question in the first seconds, the verdict at the end.
- One to three minutes per episode; a long saga becomes a series with a hook between parts.
- One or two locations: the kitchen, the car, the wedding venue.
- Subtitles burned in, for sound-off viewing (accounts that have made a purchase).

## Posting

Label AI content where the platform asks for it. Reddit communities have their own rules about AI and self-promotion. Where to post AI videos on Reddit and posting to TikTok, Shorts and Reels.

## Cost

A two-minute story is 120 seconds of video at 10 credits a second on SceneMixer's default tier, plus 1 credit a second to join. The cast is reused across parts, so a series costs only its video after the first.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make Reddit story videos?

Yes, narrated or acted, from a story you wrote or have permission to use.

### Can I use any Reddit post?

Only your own or with the author's permission; change details that identify real people.

### Do I need gameplay footage in the background?

No. SceneMixer acts the story with its own characters in its own places.

### Can the narrator sound like the poster?

Pick a narrator voice from SceneMixer's preset library, or upload your own clip as a member.

### What length works for TikTok?

One to three minutes per part, vertical, with the question up front.

### Are subtitles available?

Yes, burned in or SRT, on accounts that have made a purchase.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Talking Baby Video Generator: Babies With Grown-Up Voices

URL: https://scenemixer.com/ai-talking-baby-video-generator/
Make talking baby videos with AI: a baby character drawn once, a grown-up preset voice, comedy timing in short vertical videos, and the rules on real kids.

# AI Talking Baby Video Generator

Quick answer: A talking baby video is a comedy of contrast: a baby's face and a grown-up voice with opinions. In SceneMixer, describe the baby as a character, give it a voice from the preset library (a gravelly businessman, a weary grandmother), write short scenes with a punchline, and render them in 9:16. The baby is drawn once and stays the same baby in every episode. Photos of real children are not the way in; invented babies are.

Brief for AI Create: A talking baby comedy: a ten-month-old reviews his parents' cooking like a tired restaurant critic; three dishes, three verdicts, one standing ovation for mashed banana.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Design the baby

Write the baby like any character: age in months, hair, eyes, what it wears, one detail that makes it this baby (a sock always missing, a tiny bow tie). A look such as Family 3D Animation, 3D Toy Cartoon makes the joke obvious; Digital HD makes it uncanny, which can also be the joke. Describe a face.

## The voice is the joke

The read gives each character a voice in words; for a baby, write the grown-up voice you want ("a deep, tired voice of a man in his sixties"). Then match it to a preset from SceneMixer's library, free to try and change, so the same voice comes back every episode. Pick character voices.

| Baby | Voice | Running joke |
| The critic | Weary restaurant critic | Reviews every meal |
| The boss | Brisk executive | Runs the nursery like a company |
| The philosopher | Slow, gentle elder | Finds meaning in a sock |

## Write for timing

Set-up shot, reaction shot, line. Two to four seconds each. The baby's line lands better after a beat of silence, so give the reaction shot its own seconds. Shot seconds and timing and making talking baby videos.

## Real children

Uploading photos is a member feature, and photo-like faces are refused by the video provider's real-face check, so a video of your own child is not something to plan on. Invented babies are the format. Any sexualisation of minors is zero tolerance. Real-face check.

## Length and cost

Twenty to sixty seconds per video works on TikTok and Shorts. Thirty seconds is 30 seconds of video at 10 credits a second on SceneMixer's default tier. The baby is drawn once, so a series only adds video.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make a talking baby video?

Yes. Describe a baby character, give it a grown-up voice, write the scene and render it.

### Can I use a photo of my baby?

Not reliably: uploading is a member feature and photo-like faces are refused by the face check. Invented babies work.

### How do I give the baby an adult voice?

Describe the voice in words and match it to a preset from the library.

### Will it be the same baby in every video?

Yes. SceneMixer reuses the character sheet in every shot and episode.

### Which style is funniest?

3D cartoon looks make the joke clear; realistic looks make it stranger.

### How long should it be?

Twenty to sixty seconds for short-form platforms.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Animation Generator: 17 Animated Looks, Script to Episode

URL: https://scenemixer.com/ai-animation-generator/
Turn a script into animated episodes: 17 animated looks across 3D and 2D, from Family 3D and Anime to Claymation and Ink Wash, with the same characters in every shot.

# AI Animation Generator

Quick answer: SceneMixer is an AI animation generator for stories: pick one of 17 animated looks, paste or write the script, and it draws the cast once in that look and animates every shot with voices, sound and music, episode after episode. It makes new animated scenes from a script; it does not animate a drawing or photo you already have frame by frame. Video is 10 credits a second on the default tier; the first 15 setup images on an account are free.

Brief for AI Create: A three-minute animated short: a lighthouse keeper's cat who is afraid of the sea has to row out to the rock when the lamp goes dark; no dialogue, only the storm and one meow.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## 3D animation: six looks

| Look | What it looks like | Try it for |
| Family 3D Animation | Big-eyed, rounded characters with glowing skin | Family comedy, kids' series |
| 3D Toy Cartoon | Rounded, matte toy-like figures in bright colours | Preschool, toy worlds |
| Cel-shaded 3D | Anime characters built in 3D, flat colour bands | Action anime, games |
| Realistic 3D | Near-human 3D characters and skin, like film CG | Sci-fi, fantasy epics |
| Chinese 3D Animation | Detailed characters, flowing hair, glossy jade and metal | Xianxia, legends |
| Game Engine Realism | Next-gen game graphics: sharp materials, no grain | Game trailers, cyberpunk |

## 2D and handmade: eleven looks

| Look | What it looks like | Try it for |
| Anime | Clean lines and flat colours, like TV anime | Series, shonen, romance |
| Painterly Anime | Lavish, saturated scenery; simple-lined characters | Films, emotional stories |
| Watercolor Anime | Watercolour-on-paper look, fine lines, natural colours | Gentle fantasy, slice of life |
| Korean Webcomic | Slender, polished characters, soft airbrushed shading | Webtoon romance and revenge |
| Painted Manhua | No line art, painted volume, rich contrast | Cultivation, action |
| American Comic | Bold ink lines and heavy black shadows | Superheroes, noir |
| Watercolor Storybook | Transparent watercolour and pencil, like a children's book | Bedtime stories, fables |
| Chinese Painted Style | Fine brush lines, muted ink tones, silk texture | Folk tales, palaces |
| Ink Wash Animation | Shapes painted in ink washes on rice paper | Poetry, wuxia |
| Classical Oil Painting | Visible brushwork over warm brown, like a museum painting | Myths, period stories |
| Claymation | Hand-pressed clay figures and miniature sets | Gentle comedy, small worlds |

Visual Style: 3D Toy Cartoon

Visual Style: Cel-shaded 3D

Visual Style: Anime

Visual Style: Painterly Anime

Visual Style: Korean Webcomic

Visual Style: Watercolor Storybook

Visual Style: Claymation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

In SceneMixer, the look is chosen in Step 1 and fixed once you continue; the location and costume pictures switch to drawn versions to match. Four further looks are live action. Every visual style, one by one.

## From script to animation

- Step 1: the look, the format (Short Drama, TV Series or Film), 9:16 or 16:9, region and dialogue language. Then the script is read.
- Step 2: every character, location and prop is drawn once in the look; these sheets are what every shot copies.
- Storyboard: each episode is written as numbered shots with seconds, framing, lines and sound; you can edit any of it.
- Animation: each segment is rendered with the voices and sound in the same pass, then the episode is joined with a background score.
The whole workflow · Animated looks on each tier.

## The same character in every shot

Animation falls apart when the hero's face changes between shots. In SceneMixer, each character has one sheet in the chosen look, and every shot they appear in is generated from it, across every episode. Drawn and 3D faces also sit far from real people, so they are rarely refused by the face check. Same face in every shot · Animals, objects and creatures as characters.

## What it is not

It does not take an existing drawing or photo and set that picture moving frame by frame: it draws new scenes from your script. If you already have character art made as a sheet (full figure, plain background), it can be uploaded on a paid plan and used as is. Uploading your own image. Ads are the one place that starts from photos: upload a product photo and an ad is written and shot around it. An ad from one photo.

## Where to go next

Anime video generator · cartoon video generator · cartoon series · kids' stories · animated short film · 3D cartoon movie · claymation short · 2D animation · 3D animation · stop motion and claymation · comic to video · Chinese animation.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### What is an AI animation generator?

A tool that turns a script or idea into animated video. SceneMixer draws the cast once in an animated look and animates every shot, with voices and music.

### Which animation styles are there?

Seventeen: six 3D looks (Family 3D Animation, 3D Toy Cartoon, Cel-shaded 3D and more) and eleven 2D and handmade looks (Anime, Watercolor Storybook, Claymation and more).

### Can it animate my drawing or photo?

Not frame by frame. SceneMixer makes new animated scenes from a script; character art made as a sheet can be uploaded on a paid plan.

### Do characters stay the same between shots?

Yes. Each character has one sheet in the chosen look, which SceneMixer uses in every shot and every episode.

### Can I change the style later?

No. SceneMixer fixes the look after Step 1; start a new project to try another.

### Does the animation have sound?

Yes. Voices, ambient sound and effects come with each segment; SceneMixer puts a background score under the episode.

### How long can an animation be?

A few seconds to a whole series: segments are joined into episodes.

### Is it free?

New accounts on SceneMixer get welcome credits, free setup images and a free first storyboard; rendering video costs credits.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Video Remake: Re-shoot a Clip With a New Cast, Look and Language

URL: https://scenemixer.com/ai-video-remake/
Remake a video with AI: a clip is read shot by shot into a cast, sets and a storyboard, then shot again with new characters, a new look and a new language. Not a filter.

# AI Video Remake

Quick answer: Upload an MP4 or MOV and SceneMixer reads it shot by shot: where the cuts fall, who is on screen, what they do and what they say. That becomes a cast, a set of locations, a prop list and a finished storyboard. Then you pick a visual style, a wardrobe and a region, and every shot is generated again with your own characters. The source clip is never shown to the video model, so the result is a new picture, not a filter over the old one. The read is priced by the clip's length in minutes; everything after it is priced like any drama.

Try it: a one-minute scene you shot on a phone, read into a storyboard, then rendered as Anime with the two leads moved to Tokyo and speaking Japanese. The beats are yours; the people, the streets and the language are new.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | An MP4 or MOV you upload, read shot by shot and re-shot with your own cast, look and language; the source clip never reaches the video model |

## What a remake is, and what it is not

Searches for “remake video AI”, “recast” and “reimagine video” mean different things to different people, so here is the plain version. A remake in SceneMixer keeps the structure of the clip: the order of shots, how long each holds, who is in frame, the camera move, the spoken lines. It replaces the picture: faces come from your character sheets, rooms from your location plates, clothes from the wardrobe you choose. It is closer to re-shooting a scene with a new cast than to running a style filter over footage.

Two things it therefore cannot do. It will not turn you into an anime character: your face is not kept, a drawn character stands where you stood. And it will not preserve a real actor's likeness from the source: the read produces a written description of each person, and the sheet drawn from it is a new face. How a remake works, step by step · Where your uploaded clip goes.

## Who this is for

| If you are… | What you make here | Start with |
| A creator who wants the same beat in a new skin | The clip's timing and staging, re-shot as anime, 3D cartoon or film | Pick the look in Step 1 after the read |
| A team localising a scene | The same scene with the cast moved to another region and speaking another language | Target region |
| A writer studying a scene's blocking | A full storyboard of a clip you admire, to learn from or build on | Previs and animatics |
| Someone with old footage and a new idea | A remake of your own clip with a different cast, then new episodes that continue it | Add episodes |

Use material you have the rights to. Reading a clip is not a licence to its story, and the result is a new picture that you then own. What you get to keep.

## Video to anime, video to cartoon: how it works here

If what you want is a live-action clip turned into anime or a cartoon, SceneMixer's remake does it with one difference from a converter: the people are redrawn as new characters, not traced. Choose Anime, Painterly Anime, Korean Webcomic, Family 3D Animation or Claymation in Step 1 and the whole clip is shot in that look, with the cast, sets and props drawn once so they match from shot to shot. Anime video generator · Cartoon video generator.

Visual Style: Painterly Anime

Visual Style: Korean Webcomic

Visual Style: Family 3D Animation

Visual Style: Claymation

Visual Style: Digital HD
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## The read, in four passes

- Find the cuts. The clip is split into shots.
- Identify the cast, scenes and props. Every person who appears becomes a character with a written appearance and a voice description; every place becomes a location; every prop that matters becomes an item.
- Read each shot. Framing, camera move, who is in frame, what is said. Dialogue is transcribed and attributed to the speaker, in the language it was spoken.
- Check and lay out the storyboard. Numbered shots with seconds inside segments, ready for the editor.
A fast-cut clip has more shots to read than a slow one, so it takes longer, but it does not cost more: the read is priced by length alone, rounded up to the next minute, and the price is on the button before you press it. Limits, read time and price · How dialogue and music are read.

## Then it is a normal project

After the read you are in the same SceneMixer editor as every drama: Step 1 for the look, the format and the dialogue language; Step 2 for the cast and locations (the first 15 setup images on an account are free); the storyboard, which you can edit line by line; then video segment by segment at 10 credits a second on the default tier, and a join at 1 credit a second. Members can swap one character for another after the fact, or replace a voice. Recast a remake · Swap a character in a clip · Change a voice.

## Limits worth knowing

- SceneMixer takes one clip per remake, MP4 or MOV, within the length and size the composer shows (the length cap is 300 seconds on the default setting).
- Dialogue is read in the language spoken; to change the language, set it in Step 1 and the lines are performed in the new one.
- Real people's faces, famous characters and brand mascots are refused by the IP check, whether they came from the clip or from a photo. IP check.
- Music is written for the remake; the source's soundtrack is read for mood, not copied.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Reading the source | Upload an MP4 or MOV: it is read shot by shot into a cast, locations, props and a storyboard; the clip itself never reaches the video model | Watch the clip, write the shot list and character descriptions yourself |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI remake a video?

Yes. SceneMixer reads a clip shot by shot into a cast, locations and a storyboard, then generates every shot again with new characters in the look you choose.

### Is this a video-to-anime filter?

No. In SceneMixer, the people are redrawn as new characters and the whole clip is re-shot in an anime look; your own face is not kept.

### Can I keep the actors' faces from the original?

No. SceneMixer's read writes a description of each person; the character sheet drawn from it is a new face. Real people's likenesses are refused.

### What does a remake cost?

SceneMixer's read is priced by the clip's length in minutes, rounded up, shown on the button. Images, storyboards and video afterwards are priced like any drama.

### Can the remake be in another language?

Yes. Choose the dialogue language in SceneMixer's Step 1; the lines are performed in it. Any of 15 languages.

### Can I change who plays whom afterwards?

SceneMixer members can recast a character or replace a voice in a finished clip without re-reading the source.

### What clips are allowed?

SceneMixer accepts clips you shot or have the rights to, within the length and size shown on the upload box. Reading a clip is not a licence to its story.

### Does it keep the music?

No. SceneMixer reads the soundtrack for mood and writes new music for the remake.

Updated November 3, 2026


---

# 1-Minute AI Video Generator: 30 and 60 Second Videos With Dialogue

URL: https://scenemixer.com/1-minute-ai-video-generator/
Make a 30- or 60-second AI video: one or two segments with dialogue, sound and music, vertical or landscape, for ads, TikTok, Reels and Shorts. What it costs.

# 1-Minute AI Video Generator

Quick answer: A minute of AI video in SceneMixer is one or two generated segments: up to 15 seconds each on the standard setting, up to 30 on the long-take tiers, each with dialogue, ambience and music generated in the same pass, then joined into one file. A 60-second episode on the default tier costs sixty seconds at 10 credits a second plus a small join fee; a 15-second ad is quoted as one price before it starts. Vertical 9:16 for feeds, 16:9 for YouTube and sites.

Brief for AI Create: A 60-second short: a night-shift nurse finds a child's drawing of her own face in a stranger's locker; two shots in a corridor, one in the car park, no narration, one line at the end.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a minute is made of

| You want | Segments | Shots inside | Where it fits |
| 15 seconds | 1 | 3–5 | An ad, a hook, a Reel |
| 30 seconds | 1 long take or 2 standard | 6–10 | An ad, a TikTok, a teaser |
| 60 seconds | 2 long takes or 4 standard | 12–20 | Shorts (up to 3 minutes), Reels, a one-minute episode |

A shot is a few seconds; a segment is one generation and the unit you pay for; the episode is the file you download. For a minute you choose the segment length per episode, and SceneMixer's storyboard writes the shots to fit. 15 or 30 second segments · Shot seconds and timing.

## Who this is for

| If you are… | What you make here | Start with |
| Selling something | A 15, 30 or 60 second ad from one to four photos, with an end card | AI ad video generator |
| Running a TikTok or Reels account | Vertical one-minute dramas with a hook in the first three seconds | AI drama for TikTok and Reels |
| Posting Shorts | 60-second episodes that stand alone, or a serial with a cliffhanger each | Posting to Shorts |
| Testing an idea before a longer piece | A one-minute version of a scene, watched before you spend on the rest | The free preview |

## What a minute costs, and what is free

Video is charged by the finished second: 10 credits a second on SceneMixer's default tier, about $0.06 a second at the lowest credit price, so a 60-second episode is sixty times that plus 1 credit a second to join it. Reading a script is 4 credits per 1,000 characters; the first 15 setup images on an account are free, as is the first episode's storyboard on your first drama and a preview of about 5 seconds. New accounts start with 50 credits. An ad is quoted as one all-in figure before it starts. Where credits go in one episode · The ad price is a cap.

## Why a minute is still a story

A minute is long enough for a setup, a turn and a last look, and short enough that one weak shot shows. Three habits help. Write one action per shot (a run, a glance, a line), not a paragraph. Put the hook in the first shot, because that is all a feed shows before the scroll. End on a question if there is a next episode, and on a still if there is not. Hooks and cliffhangers · How to write one shot.

SceneMixer performs dialogue in the video itself, in the language you set in Step 1, so a 60-second piece can carry four to eight spoken lines without a dubbing pass. AI video generator with dialogue.

## Vertical or landscape

Choose 9:16 Portrait in Step 1 for TikTok, Reels and Shorts, and every sheet, plate and shot is composed tall from the start; 16:9 Landscape for YouTube, a website or a screen. The format is fixed once the script is read, so decide where the minute will live before you start. Vertical AI video generator · Vertical or landscape.

Other lengths: 3 to 5 minutes: one episode · 10 minutes and beyond: a series.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### How do I make a 1-minute AI video?

Write or paste a short script in SceneMixer, choose the format in Step 1, generate the cast, read the storyboard, then render two long-take segments or four standard ones and join them.

### Can it make a 30-second video?

Yes: one long-take segment, or two standard ones, joined. Ads on SceneMixer come in 15, 30 and 60 seconds.

### How long does a minute take to generate?

SceneMixer renders each segment on its own, and you can queue them together; a minute is usually ready in minutes, longer on busy tiers.

### Is there dialogue?

Yes. SceneMixer speaks the lines in the generation, in the language set in Step 1, with a voice per character.

### What does 60 seconds cost?

Sixty seconds at 10 credits a second on SceneMixer's default tier plus the join, before any retakes; at the lowest credit price that is sixty times $0.06 for the video alone.

### Can I make it vertical for TikTok?

Yes. Choose 9:16 Portrait in Step 1; SceneMixer composes the whole project tall.

### Is the first one free?

The first setup images, the first episode's storyboard and a short preview are free on SceneMixer; the minute of video itself is paid.

### Can I add subtitles?

Yes, on accounts that have made a purchase. SceneMixer can burn the captions into the picture or give you an SRT file.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# 5-Minute AI Video Generator: One Episode, Up to 60 Segments

URL: https://scenemixer.com/5-minute-ai-video-generator/
Make a 3- to 5-minute AI video as one episode: up to 60 segments joined into one file, the same cast throughout, for YouTube, Facebook and story channels.

# 5-Minute AI Video Generator

Quick answer: Five minutes of AI video is an episode: a stack of segments, each a generation of up to 15 or 30 seconds, joined into one file with the same faces, rooms and voices from the first second to the last. Short Drama episodes run to 240 seconds, TV Series to 300, Film without a cap; one episode holds up to 60 segments. Three to five minutes is the natural length for a YouTube story video, a Facebook episode or a lesson, and it is where consistency starts to matter more than any single shot.

Brief for AI Create: A five-minute TV Series pilot: two estranged sisters inherit a failing seaside hotel and have one season to turn it round; three scenes, one hotel, a cliffhanger.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## How five minutes is built

| Target | Long-take segments (30 s) | Standard segments (15 s) | Type to choose |
| 3 minutes | 6 | 12 | Short Drama (cap 240 s) |
| 4 minutes | 8 | 16 | Short Drama |
| 5 minutes | 10 | 20 | TV Series (cap 300 s) or Film |

Each segment opens where the last one ended: the system tracks who stands where, what they hold and what has changed, so a five-minute scene does not reset at every cut. What is tracked between segments · Why the next segment opens like the last one ended.

## Who this is for

| If you are… | What you make here | Start with |
| Running a YouTube story channel | A three- to five-minute acted or narrated story per upload | AI story video generator |
| Writing a series | A pilot episode of about five minutes, then the next ones from the same cast | AI TV show generator |
| Teaching | A five-minute lesson as a small drama or an explainer with a recurring guide | Videos for your classroom |
| Pitching | A five-minute proof of a longer idea, watched before anyone commits | A TV pilot or pitch |

## What five minutes costs

Video is 10 credits a second on SceneMixer's default tier, so 300 seconds is three hundred times that, plus 1 credit a second for the join, before retakes. At the lowest credit price that is about $0.06 a second. The honest advice for a first five-minute piece: draft it at 480P, watch it, fix the two or three shots that missed, then raise the keepers. Draft at 480P, finish at 1080P · Spend fewer credits.

## Generate it without babysitting

Press Generate all and the segments queue in order; you can close the tab and come back to a finished episode. The storyboard is the place to spend your attention: read all twenty segments before rendering any, because a fix there costs nothing and a fix after costs a segment. Generate all segments at once · Checklist before you generate.

## Time, and what to do with it

A five-minute episode takes longer than the sum of its segments because SceneMixer renders them in a queue shared with everyone; expect something between a coffee and a meal, longer on the higher tiers. Look around the set while a shot renders, or read the next episode's outline. Why a generation takes time.

Other lengths: 30 to 60 seconds: one or two segments · 10 minutes and beyond: a series.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make a 5-minute video?

Yes, as one episode of up to 60 segments that SceneMixer joins into one file. Choose TV Series or Film type for the 300-second cap or no cap.

### How many segments is five minutes?

On SceneMixer, five minutes is ten long-take segments of 30 seconds, or twenty standard segments of 15.

### Will the characters stay the same for five minutes?

Yes. SceneMixer draws each character once and renders every segment against that sheet; positions and props are tracked between segments.

### How much does a 5-minute AI video cost?

Three hundred seconds at 10 credits a second on SceneMixer's default tier plus the join, before retakes; at the lowest credit price about $0.06 a second for the video alone.

### How long does it take?

SceneMixer renders segments in a shared queue; a five-minute episode usually takes a good part of an hour, longer on the top tiers.

### Can I make it longer than five minutes?

Yes. Film type on SceneMixer has no episode cap, and a project can hold many episodes; see the long video page.

### Can it be 3 minutes instead?

Yes. Six long-take segments or twelve standard ones; Short Drama type on SceneMixer caps an episode at 240 seconds.

### Is there a free part?

The first setup images, the first episode's storyboard and a short preview are free on SceneMixer; video is paid by the second.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI 2D Animation Generator: Eleven Drawn Looks, Script to Series

URL: https://scenemixer.com/ai-2d-animation-generator/
Make 2D animated videos from a script in eleven 2D looks, from Anime and Korean Webcomic to Watercolor Storybook and Ink Wash, with one drawn cast in every shot.

# AI 2D Animation Generator

Quick answer: SceneMixer makes 2D animation from a story: choose one of eleven 2D looks in Step 1, paste or write the script, and it draws every character, place and prop once in that look, then animates each shot with the dialogue, ambience and music generated together. It is for making new 2D scenes from text; it does not animate a drawing you already have. Video is 10 credits a second on the default tier and the first 15 setup images on an account are free.

Brief for AI Create: A 2D animated short: a girl who runs the last analogue cinema in a city of screens gets one night to show a film nobody remembers; one street, one projector room, no villain.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The eleven 2D looks

| Look | What it looks like | Suits |
| Anime | Clean lines and flat colours, like TV anime | Series, romance, action |
| Painterly Anime | Lavish, saturated scenery; simple-lined characters | Films, emotional stories |
| Watercolor Anime | Watercolour on paper, fine lines, natural colours | Gentle fantasy, slice of life |
| Korean Webcomic | Slender, polished characters, soft airbrushed shading | Webtoon romance and revenge |
| Painted Manhua | No line art, painted volume, rich contrast | Cultivation, action |
| American Comic | Bold ink lines and heavy black shadows | Superheroes, noir |
| Watercolor Storybook | Transparent watercolour and pencil, like a children's book | Bedtime stories, fables |
| Chinese Painted Style | Fine brush lines, muted ink tones, silk texture | Folk tales, palaces |
| Ink Wash Animation | Shapes painted in ink washes on rice paper | Poetry, wuxia |
| Classical Oil Painting | Visible brushwork over warm brown | Myths, period stories |
| Claymation | Hand-pressed clay figures and miniature sets | Small worlds, gentle comedy |

Visual Style: Painterly Anime

Visual Style: Watercolor Anime

Visual Style: Korean Webcomic

Visual Style: Painted Manhua

Visual Style: American Comic

Visual Style: Watercolor Storybook

Visual Style: Ink Wash Animation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

In SceneMixer, the look is chosen once and fixed when you continue past Step 1; the location and costume pictures switch to drawn versions to match. Every visual style, one by one.

## Who this is for

| If you are… | What you make here | Start with |
| Making an anime or webtoon series | Episodes in Anime or Korean Webcomic with the same drawn cast | Anime video generator |
| Making a children's channel | Watercolor Storybook or Family 3D stories with a recurring guide | Kids' story videos |
| Adapting a comic | Panels that become moving shots with spoken lines | Comic to video |
| Making a short film | A drawn short with a score and no dialogue, or with it | Animated short film |

## How 2D animation is made here

- Step 1: the 2D look, the type (Short Drama, TV Series or Film), 9:16 or 16:9, region and dialogue language. Then SceneMixer reads the script.
- Step 2: every character, location and prop is drawn once in the look. These sheets are what every shot copies.
- Storyboard: each episode as numbered shots with seconds, framing, lines and sound; edit any of it as text.
- Animation: each segment is rendered with the voices and sound in the same pass, then joined with a score.
The whole workflow · Drawn looks on each tier.

## What 2D does well, and what to write for it

Drawn characters hold their faces better than live-action ones, and drawn faces are rarely refused by the real-person check, so 2D is the easiest place to keep a cast consistent across twenty episodes. Write fights as one move per shot, give each shot one light source, and let the scenery carry the mood: a Painterly Anime sky can do what a page of description cannot. Anime fight scenes · Same face in every shot.

## What it is not

SceneMixer does not animate a picture you upload, frame by frame; it draws new scenes from a script. It does not imitate a named studio or a named show: the looks are generic styles, and famous characters are refused. Members can upload a reference photo for a character, which is redrawn in the look rather than traced. Use your own photo · The IP check.

Other animation pages: 3D animation · Stop motion and claymation · Chinese animation · All 17 animated looks.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### What is an AI 2D animation generator?

A tool that turns a script into 2D animated video: in SceneMixer, eleven drawn looks, a cast drawn once, and every shot animated with dialogue and sound.

### Can I animate my own drawing?

No. SceneMixer draws new scenes from a script. Members can upload a photo as a reference for a character, which is redrawn in the look.

### Which 2D style is best for anime?

Anime for a TV look, Painterly Anime for films, Watercolor Anime for gentle stories, Korean Webcomic for webtoon romance.

### Do the characters stay the same?

Yes. SceneMixer gives each character one sheet in the chosen look and every shot that features them is generated from it.

### Is it free?

New accounts on SceneMixer get 50 credits, 15 free setup images, a free first-episode storyboard and a short preview. Video is paid by the second.

### Can the characters talk?

Yes. On SceneMixer, lines are performed in the generation, in any of 15 languages, with a voice per character.

### Can I mix 2D and live action?

Not in one project; SceneMixer fixes the look for the whole series. Start a second project for a different look.

### Which tier for 2D?

Drawn looks render well on SceneMixer's default tier; raise the shots you keep to 1080P or 2K afterwards.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI 3D Animation Generator: Six 3D Looks, One Cast, Every Shot

URL: https://scenemixer.com/ai-3d-animation-generator/
Make 3D animated videos from a script in six 3D looks, from Family 3D and 3D Toy Cartoon to Realistic 3D and Game Engine Realism, one cast kept in every shot.

# AI 3D Animation Generator

Quick answer: SceneMixer makes 3D animation from a story: choose one of six 3D looks in Step 1, give it a script, and it designs every character, place and prop once in that look, then renders each shot with the dialogue, ambience and music generated together. There is no modelling, rigging or keyframing: you direct in text and the look does the rest. Video is 10 credits a second on the default tier; the first 15 setup images on an account are free.

Brief for AI Create: A 3D animated family short: a retired delivery robot adopts a stray kitten the week the city bans pets on the tram line; one neighbourhood, five episodes of two minutes, one tram.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The six 3D looks

| Look | What it looks like | Suits |
| Family 3D Animation | Big-eyed, rounded characters with glowing skin | Family comedy, kids' series |
| 3D Toy Cartoon | Rounded, matte toy-like figures in bright colours | Preschool, toy worlds |
| Cel-shaded 3D | Anime characters built in 3D, flat colour bands | Action anime, games |
| Realistic 3D | Near-human 3D characters and skin, like film CG | Sci-fi, fantasy epics |
| Chinese 3D Animation | Detailed characters, flowing hair, glossy jade and metal | Xianxia, legends |
| Game Engine Realism | Next-gen game graphics: sharp materials, no grain | Game trailers, cyberpunk |

Visual Style: 3D Toy Cartoon

Visual Style: Cel-shaded 3D

Visual Style: Realistic 3D

Visual Style: Chinese 3D Animation

Visual Style: Game Engine Realism
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

SceneMixer fixes the look once you continue past Step 1, and the location and costume pictures are rendered in 3D to match. Every visual style.

## Who this is for

| If you are… | What you make here | Start with |
| Making a kids' or family channel | Family 3D or 3D Toy Cartoon episodes with a recurring cast | Cartoon series generator |
| Making a game trailer or cutscene | Game Engine Realism scenes of your own world and characters | Game trailer or cutscene |
| Telling a big story | Realistic 3D or Chinese 3D for sci-fi, fantasy and legends | Space opera |
| Making a 3D movie | A film in acts, in one 3D look, with a score | 3D cartoon movie |

## No rigging, no keyframes

Traditional 3D means modelling a character, rigging it, animating every movement and lighting every scene. In SceneMixer, the model does that work from your text: a character sheet in the 3D look is the model, the shot's words are the animation, and the segment's time and weather are the lighting. What you keep is the direction: who is in frame, what they do, what they say, how long the shot holds. How to write one shot · Write positions the world can read.

## What 3D does well, and where to be careful

3D looks hold materials and lighting well: glossy armour, wet streets, fur, glass. They are forgiving about faces, because a stylised 3D face sits far from any real person. Be careful with crowds and fast choreography, which every video model struggles with: write one move per shot and keep background people as scenery. Fights, crowds and group scenes · Background people.

## What it is not

SceneMixer does not export a rig, a mesh or a scene file; the output is video. It does not animate a model you already have. It does not copy a named studio's look or a named character: the looks are generic, and famous characters and brand mascots are refused. Content rules.

Other animation pages: 2D animation · Stop motion and claymation · Chinese animation · All 17 animated looks.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### What is an AI 3D animation generator?

A tool that turns a script into 3D animated video without modelling or rigging: in SceneMixer, six 3D looks and a cast designed once for every shot.

### Can I export the 3D model?

No. SceneMixer's output is video; there is no mesh, rig or scene file.

### Which look is closest to a family animated film?

Family 3D Animation. 3D Toy Cartoon is simpler and brighter, for preschool.

### Can it make game-style animation?

Yes. On SceneMixer, Game Engine Realism renders like next-generation game graphics; Cel-shaded 3D looks like a 3D anime game.

### Do the characters stay consistent?

Yes. Each character has one sheet in the 3D look and SceneMixer generates every shot from it.

### Is there dialogue?

Yes, performed in the generation with a voice per character, in any of 15 languages.

### What does it cost?

Video is 10 credits a second on SceneMixer's default tier; the first 15 setup images and the first episode's storyboard are free.

### Can I upload a 3D character I designed?

Members can upload a reference image for a character; SceneMixer redraws it in the chosen look rather than importing it as a model.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Stop Motion Video Generator: Claymation Looks From a Script

URL: https://scenemixer.com/ai-stop-motion-video-generator/
Make stop-motion and claymation style videos with AI: clay figures and miniature sets drawn once and kept in every shot, from a script, with voices and music.

# AI Stop Motion Video Generator

Quick answer: SceneMixer's Stop-motion group holds the Claymation look: hand-pressed clay figures, thumbprint textures, miniature sets, the slightly uneven motion of a film shot a frame at a time. Choose it in Step 1, write the story, and the cast and sets are drawn once in clay, then every shot is generated in that look with the dialogue and music in the same pass. No armatures, no lights, no eleven months in a garage. Video is 10 credits a second on the default tier.

Brief for AI Create: A claymation short: a lighthouse keeper's teapot comes to life every night to argue about the weather; one kitchen, one lamp room, three two-minute episodes, no narrator.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What the look does

Claymation reads as handmade. Figures have fingerprints and seams; sets are miniatures with visible paint; light is a single warm lamp. The model keeps those tells in every shot, which is the whole charm, and the look forgives what video models get wrong: a slightly stiff walk looks intentional in clay. Expect gentle comedy, small worlds and talking objects to land best; expect fast action and large crowds to fight the look.

Visual Style: Family 3D Animation

Visual Style: Watercolor Storybook

Visual Style: 3D Toy Cartoon
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

How to make an AI claymation short walks through a complete one.

## Who this is for

| If you are… | What you make here | Start with |
| Making a children's or family series | Clay episodes with the same figures every time | Cartoon series generator |
| Making a short film | A clay short with a score, for a festival or a feed | Animated short film |
| Advertising a small brand | A clay version of your product's world, with an end card | AI ad video generator |
| Telling a fable | Talking animals and objects in miniature sets | Fairy tale videos |

## Objects and animals as characters

Clay is the natural home for a talking teapot. Give the object a name in the script and it is read as a character: it gets a sheet, a voice and a place in every shot it appears in, and the lines you write for it are spoken. Non-human characters · Talking animal videos.

## Writing for stop motion

- One movement per shot. A turn, a hop, a reach. Clay motion is read as deliberate; let each shot do one thing.
- Small sets, few people. Two or three figures per shot keep the handmade feel; crowds become a blur of clay.
- Light from a lamp. Write the light source into the shot: a single desk lamp from the left. One source is how miniature sets are lit.
- Textures in the words. Fingerprints on the cheek, a cardboard sea, a cotton-wool cloud on a wire all come through.
How to write one shot · Describe the cause, not the result.

## What it is not

It is a look, not a technique: nothing is photographed a frame at a time, and no clay is involved. SceneMixer will not turn footage you shot into stop motion, and it will not reproduce a named studio's characters. Members can upload a reference image for a character, which is redrawn in clay rather than copied. Use your own photo.

Other animation pages: 2D animation · 3D animation · Cartoon video generator · All 17 animated looks.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make stop motion animation?

It can make video in a stop-motion look: SceneMixer's Claymation style renders clay figures and miniature sets from a script, with the handmade tells kept in every shot.

### Is anything actually photographed frame by frame?

No. In SceneMixer, stop-motion is a visual look generated from text, not a filming technique.

### Can I turn my own footage into stop motion?

Not as a filter. A remake reads a clip into a storyboard and re-shoots it with new clay characters; see the remake page.

### Can objects talk?

Yes. Name an object in the script and SceneMixer turns it into a character with a sheet, a voice and lines.

### Which stories suit claymation?

Gentle comedy, small worlds, fables, talking objects and animals. Fast action and crowds fight the look.

### Do the figures stay the same?

Yes. SceneMixer draws each character once in clay and generates every shot from that sheet.

### What does it cost?

Video is 10 credits a second on SceneMixer's default tier; the first 15 setup images and the first episode's storyboard are free.

### Is there music?

Yes, SceneMixer writes a score for the episode; a dialogue-only version is one switch away.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Comic to Video: Turn a Comic, Manga or Webtoon Story Into Animation

URL: https://scenemixer.com/ai-comic-to-video/
Turn a comic, manga or webtoon story into animation with AI: the script becomes shots, the cast is drawn once in a comic look, every panel a moving shot with lines.

# AI Comic to Video

Quick answer: SceneMixer turns a comic story into video: paste the script or the panel-by-panel text of your comic, choose a comic look (American Comic, Korean Webcomic, Painted Manhua or Anime), and it draws your cast once in that look, writes each panel as a shot with seconds and spoken lines, and generates the episode with voices and music. It works from your words, not from scanned pages: the panels are redrawn as new shots, so the result moves like animation rather than a page turning.

Brief for AI Create: From my webtoon, chapter 1: a delivery courier in a flooded megacity finds a sealed letter addressed to herself, dated ten years from now; six panels become one 60-second episode, Korean Webcomic look.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | The script or panel-by-panel text of your comic; panels are redrawn as new shots, scanned pages are not used |

## Two ways in

| You have | What to paste | What happens |
| A script or a chapter outline | The text, with character names and lines | Read into episodes, cast, locations and a storyboard |
| Finished panels with dialogue | The panel text in order: who is in it, what happens, what is said | Each panel becomes a shot; the lines are spoken by the right character |
| Only an idea | One sentence in AI Create | A script is written, then the same pipeline runs |

Scanned pages are not read. Members can upload a reference image for a character so the drawn sheet follows your design; it is redrawn in the look, not traced. Use your own reference image · Turn a manga or webtoon into an anime series.

## Who this is for

| If you are… | What you make here | Start with |
| A webtoon or comic author | An animated episode per chapter, in a look close to your pages, to post beside them | Anime series for YouTube |
| A writer without an artist | Your comic script as a finished animated short | Anime video generator |
| A publisher or studio | A trailer or a pilot episode of a comic property to test it | AI trailer generator |
| A teacher or librarian | A comic strip with a lesson, animated with voices | Educational cartoon |

## The comic looks

| Look | What it looks like | Closest to |
| American Comic | Bold ink lines and heavy black shadows | Superhero and noir comics |
| Korean Webcomic | Slender, polished characters, soft airbrushed shading | Webtoons |
| Painted Manhua | No line art, painted volume, rich contrast | Chinese manhua |
| Anime | Clean lines and flat colours | Manga adapted for TV |
| Cel-shaded 3D | Anime characters built in 3D | Game-style manga |

Visual Style: Korean Webcomic

Visual Style: Painted Manhua

Visual Style: Anime

Visual Style: Cel-shaded 3D
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## From panel to shot

A panel is a frozen moment; a shot is a few seconds of movement with sound. SceneMixer turns each panel into a shot with a duration, a framing, who is in frame and the line that is spoken, and adds the connective tissue a page leaves out: the walk between two panels, the reaction after a line. Speech bubbles become spoken dialogue with a voice per character; sound effects written in the panel (CRASH) become ambience. Text on screen is not drawn into the picture; captions are added after export. Dialogue in the storyboard · Text on screen.

## What it is not

It does not animate your existing artwork or scan your pages. It does not reproduce published characters: famous manga and comic characters are refused by the IP check, so this is for your own stories. The IP check.

Other animation pages: 2D animation · Anime video · Chinese animation · All 17 animated looks.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI turn a comic into a video?

SceneMixer turns a comic's script or panel text into animated video: each panel becomes a shot with movement and spoken lines, in a comic look.

### Do I upload my pages?

No. Paste the text of the panels or the script into SceneMixer. Members can upload a character reference so the drawn cast follows your design.

### Can I adapt a published manga?

No. SceneMixer's IP check refuses published characters; use your own story and characters.

### Which look is closest to a webtoon?

Korean Webcomic. American Comic for ink and shadow, Painted Manhua for Chinese manhua, Anime for a TV adaptation feel.

### Are the speech bubbles spoken?

Yes. Each line is performed by its character's voice in the language you set in SceneMixer; subtitles can be burned in or exported as SRT.

### How long is one chapter as video?

It depends on the panels: a six-panel chapter is usually a 30- to 60-second episode; a long chapter can run to several minutes.

### Does it keep my character designs?

Members can upload a reference per character; SceneMixer redraws it in the look and then uses it for every shot.

### What does it cost?

Reading the text is 4 credits per 1,000 characters, English counted lighter; video is 10 credits a second on SceneMixer's default tier; the first 15 setup images are free.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Chinese Animation Generator: Donghua Looks, Xianxia Worlds

URL: https://scenemixer.com/ai-chinese-animation-generator/
Make donghua-style Chinese animation with AI: Chinese 3D, Painted Manhua, Chinese Painted Style and Ink Wash looks, xianxia and wuxia settings, 15 dialogue languages.

# AI Chinese Animation Generator

Quick answer: SceneMixer has four looks built for Chinese animation: Chinese 3D Animation (detailed characters, flowing hair, glossy jade and metal), Painted Manhua (no line art, painted volume), Chinese Painted Style (fine brush lines, muted ink tones, silk texture) and Ink Wash Animation (ink on rice paper). Pair one with scene and costume presets such as Ink Jianghu, Immortal Sect Robes or Tang Opulence, write the story, and the cast is drawn once in that look and kept through every episode, speaking Chinese or any of the other 14 languages.

Brief for AI Create: A xianxia donghua: an outer-sect disciple who cannot cultivate discovers her broken sword remembers every duel it ever lost; one mountain sect, one sword, ten episodes with a cliffhanger each.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The four looks

| Look | What it looks like | Suits |
| Chinese 3D Animation | Detailed 3D characters, flowing hair and robes, glossy jade and metal | Xianxia, legends, epics |
| Painted Manhua | No line art, painted volume, rich contrast | Cultivation, action, revenge |
| Chinese Painted Style | Fine brush lines, muted ink tones, silk texture | Folk tales, palace stories |
| Ink Wash Animation | Shapes painted in ink washes on rice paper | Poetry, wuxia, fables |

Visual Style: Painted Manhua

Visual Style: Chinese Painted Style

Visual Style: Ink Wash Animation

Scene Style: Ink Jianghu

Costume Style: Immortal Sect Robes

Costume Style: Tang Opulence

Costume Style: Jianghu Travelwear
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

Scene and costume presets set the world: Ink Jianghu and Silk Road City for places, Immortal Sect Robes, Jianghu Travelwear, Tang Opulence and Qing Palace Robes for dress. Every scene and costume preset.

## Who this is for

| If you are… | What you make here | Start with |
| Adapting a cultivation web novel | A donghua series, one episode per arc, the same disciples throughout | Web novel to drama series |
| Making a wuxia short | Duels written one move per shot, in Ink Wash or Painted Manhua | Xianxia or wuxia drama |
| Telling a palace or folk story | Chinese Painted Style with Tang or Qing dress | History drama |
| Reaching viewers abroad | The same donghua with dialogue in English, Indonesian or Spanish | A second-language version |

## Cultivation, sects and powers on screen

Donghua lives on things no camera can film: qi as light, a sword that flies, a realm breakthrough. Write them as visible causes: golden light gathers in her palm, the blade lifts and hangs point-down, the air cracks and the ground frosts. Give each power a colour and keep it; give each sect a dress and a mark on the sheet. SceneMixer writes ranks and realms into the character sheet so they do not drift. Describe the cause, not the result · Costume changes and states.

## Dialogue in Chinese, or not

Set the dialogue language in Step 1: Chinese, or any of the other 14 languages. The lines are performed in the generation itself, so a donghua for an Indonesian or Spanish audience is the same project with a different setting, not a dubbing job. Names stay as written in the script. Choose the dialogue language · Fifteen languages, three layers.

## What it is not

It does not reproduce a named studio's style or a named series' characters; the looks are generic, and published characters are refused. It does not animate artwork you upload, frame by frame. Fights follow the one-move-per-shot rule, and injuries are shown before and after, not as wounds. Content rules.

Other animation pages: 2D animation · 3D animation · Anime video · All 17 animated looks.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## Can you make money with AI dramas on TikTok and YouTube?

Yes, within each platform's rules, and the rules are about the account and the video, not about how the video was made. TikTok's Creator Rewards Program (checked November 2, 2026) is open to personal accounts aged 18+ in the United States, United Kingdom, Germany, Japan, South Korea, France, Mexico and Brazil with at least 10,000 followers and 100,000 video views in the last 30 days; a rewarded video must be original, longer than one minute and reach 1,000 views on the For You feed. YouTube's Partner Program needs 1,000 subscribers plus either 4,000 watch hours in the last 12 months or 10 million Shorts views in the last 90 days.

Both platforms require realistic AI-generated content to be labelled, and YouTube states that disclosing AI use alone does not limit a video's audience or its eligibility to earn. SceneMixer episodes are longer than a minute by design, carry an AI badge by default (members can switch it off and add the platform's own label instead), and are yours to publish. The full answer, with sources.

## FAQ

### Can AI make donghua?

SceneMixer makes Chinese-style animation from a script in four looks built for it, with a cast drawn once and kept through every episode.

### Which look is closest to 3D donghua?

Chinese 3D Animation: detailed characters, flowing hair, glossy jade and metal.

### Can the characters speak Chinese?

Yes. Mandarin Chinese, set in Step 1, performed in the generation. SceneMixer offers fourteen other languages for the same project.

### Can I adapt a published web novel?

Only one you own or have the rights to. SceneMixer's IP check refuses published characters.

### How do I show cultivation powers?

Write them as visible causes with a fixed colour per power; SceneMixer renders them as light, motion and weather.

### Do the robes and sect marks stay the same?

Yes. SceneMixer keeps them on the character sheet, and every shot inherits them; a change of robes becomes a state.

### Is it free to try?

New accounts on SceneMixer get 50 credits, 15 free setup images, a free first-episode storyboard and a short preview.

### Can I make wuxia in ink wash?

Yes. Ink Wash Animation with Ink Jianghu scenes and Jianghu Travelwear is built for it.

### Is SceneMixer like ReelShort or DramaBox?

No. ReelShort and DramaBox are apps for watching short dramas. SceneMixer is for making them: it turns your script or novel into episodes with a consistent cast and spoken dialogue, and you publish the episodes wherever you like, including those platforms' creator programmes where they exist.

Updated November 3, 2026


---

# AI Film Making: Script, Cast, Storyboard and Cut in One Place

URL: https://scenemixer.com/ai-film-making/
AI film making from script to finished cut: character sheets, locations, a shot-by-shot storyboard, spoken dialogue and a joined film, with each price shown first.

# AI Film Making

Quick answer: AI film making in SceneMixer follows the same order as a real production, with each department's job done by one step. You hand over a script (or a one-line idea) and it is read into a cast, locations, props and an outline; every character and place is drawn once; a storyboard lists numbered shots with seconds, framing and lines; each segment of up to 15 or 30 seconds is rendered with its dialogue and sound; and the segments are joined into a cut. Set the type to Film and the length has no hard cap: a film is as long as the script you give it, paid by the finished second.

## The crew you do not have to hire

| On a film set | Who does it | Here, in SceneMixer |
| Script and breakdown | Writer, script supervisor | Upload a script, or write the idea in AI Create. The read lists characters, locations, props and an episode or sequence outline you can edit. |
| Casting | Casting director | Every character gets a full-body sheet, drawn once in Step 2. The first 15 setup images on an account are free. |
| Production design | Art department | Each location and prop is drawn once and becomes the set for every shot that uses it. |
| Shot list and storyboard | Director, DP | Numbered shots with seconds, framing, camera move and lines. You can edit any line; editing is free. |
| Shooting | Camera, sound | Each segment renders with dialogue and ambience in one pass, at 10 credits a second on the default tier. |
| Editing | Editor | Segments are joined into one cut at 1 credit a second; each segment keeps a version history so a retake never loses the good take. |

The one job it does not take is yours: deciding what the film is about. A tour of the editor · What you can change in the storyboard.

## Set the type to Film before the script is read

Type is chosen once, in Step 1, and it rewrites the rules the read and the shot list are written under. Short Drama assumes a phone and a hook to close every unit; TV Series allows a slower opening and a closing beat; Film has no hard length cap and no forced hook, so a scene can end on a look and a line can land sideways. SceneMixer divides the film into sequences rather than episodes, and each sequence is billed and rendered on its own, which means you can finish sequence one and watch it before you pay for sequence two. AI movie generator: how the Film type works · Short drama, series or film.

## A retake is a shot, not a day

Film making is mostly deciding what to redo. Here a single shot can be rewritten and regenerated without touching the shots on either side, and a clip fix tells you its price and whether you are refunded when it fails before you confirm. Because the cast and the sets are drawn once, a retake keeps the same faces and the same room. What a clip fix costs · Why faces and rooms stay the same.

## Use it before the real shoot, too

Many people use AI film making as previs: a rough, fully voiced version of a scene to test pacing, blocking and dialogue before committing a crew or a budget, or as the proof-of-concept that goes with a pitch. Previs and animatics · A pilot or pitch trailer · A short film for a festival.

## What it does not replace

- SceneMixer does not perform: a character's acting is written into the shot, then generated, not captured from an actor.
- It will not use a real person's face or a famous character; the IP check refuses them, whether they arrive in text or in a photo.
- A feature is a very large number of paid segments (up to 60 per episode or sequence), so plan the length and the budget together: Choosing script length and cost.
- Fine control of one exact camera path is limited to what the shot description can express; the storyboard is where you steer it.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### What is AI film making?

Making a film with generative AI for the pictures and sound: you supply the script and the decisions, and each step (cast, sets, storyboard, shots, cut) is generated and editable.

### Can I make a whole film with it?

Yes, as a set of sequences. On SceneMixer, the Film type has no hard length cap; each sequence is made and paid for separately, so you can build a long film one sequence at a time.

### Do I need filmmaking experience?

No, but film vocabulary helps in the storyboard: framing, camera move, who is in the shot. The editor shows each shot as plain text you can rewrite.

### How much does it cost?

Video is charged per finished second: 10 credits a second on SceneMixer's default tier (about $0.06 a second at the lowest credit price), plus 1 credit a second to join. Reading a script is 4 credits per 1,000 characters.

### Is anything free?

New accounts on SceneMixer start with 50 credits, the first 15 setup images are free, the first episode's storyboard on your first drama is free, and there is a free preview of about 5 seconds.

### Can I use my own script?

Yes. Paste it or upload .txt, .docx or .md. SceneMixer reads a script of up to 300,000 characters of English in one project.

### Does the film have dialogue?

Yes. Each character is given a voice by SceneMixer and speaks in the video itself, in any of 15 languages.

### Can I use it for previs or a pitch?

Yes. A rough voiced version of a scene is a common use, and a pitch trailer is a short project of its own.

Updated November 3, 2026


---

# AI Movie Making: A First Movie From Idea to Download, Step by Step

URL: https://scenemixer.com/ai-movie-making/
AI movie making for beginners: what to write first, how long a movie really is in segments and credits, what is free, and the five steps from idea to a downloaded file.

# AI Movie Making

Quick answer: Making a movie with AI comes down to five moves: write it (or let AI Create write it), pick the look and the length, check the cast and the places, read the storyboard before you render anything, then generate and join. A movie is a stack of segments of up to 15 or 30 seconds, so a 5-minute movie is 10 to 20 segments and a 30-minute one is 60 to 120. Video costs 10 credits a second on the default tier; reading, setup images, the first storyboard and a short preview are free or small, so most of the money is decided by how many seconds you choose.

## Movie, film, video: the words do not matter, the length does

People search for AI movie making, film making and video making and mean the same thing: a story with characters who stay the same from the first shot to the last. What changes the work is the length. On the default tier these are the sizes, using the same arithmetic as the long video page:

| Movie length | Segments of 15 s | Segments of 30 s | A fair name for it |
| 1 minute | 4 | 2 | A teaser or an ad |
| 5 minutes | 20 | 10 | A short film, one episode |
| 10 minutes | 40 | 20 | A short film with a second act |
| 30 minutes | 120 | 60 | A series of episodes or sequences, not one file |

## The five moves

- Write it small. One setting, two or three characters, one turn. If you have nothing yet, type one line in AI Create and read what it drafts. Write in your own language.
- Pick the look and the length. Step 1: type (Short Drama, TV Series or Film), 9:16 or 16:9, a look, a region and the dialogue language. Choosing the settings.
- Check the cast and the places. Step 2 draws every character, location and prop once. Fix a face here, not in shot forty. The first 15 images on a SceneMixer account are free.
- Read the storyboard. It is numbered shots in plain text. Rewrite anything that is wrong; editing is free. The first episode's storyboard on your first drama is free.
- Generate, then join. Segments render on their own and can be queued together; when they are done, join them at 1 credit a second. Compositing an episode.

## What a first movie costs, and where it goes

Almost all of it is video: 10 credits a second on the default tier, about $0.06 a second at the lowest credit price. A 3-minute movie is a hundred and eighty of those seconds plus the join. Reading the script is 4 credits per 1,000 characters, English counted lighter, and storyboard editing is free. New accounts start with 50 credits and can preview about 5 seconds before spending. Where credits go in one episode · Spending fewer credits.

## Mistakes a first movie makes

- Starting too big. A 20-minute plan for a first try. Make a 2-minute version and learn what the storyboard does.
- Too many people. Every extra character is another sheet to approve. Two or three carry a short movie.
- Skipping the storyboard. The video is the expensive part; the storyboard is the free one. Read it like a screenplay and cut what you would not shoot.
- Fixing in the wrong place. A wrong face is a Step 2 problem; a wrong line is a storyboard problem; a wrong move in one shot is a single regeneration.
A step-by-step short film · Making a multi-episode series.

## Where to go next

Related: AI movie generator · AI movie maker · AI film making · A 5-minute movie · Script to movie.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### How do I make a movie with AI?

Write or generate a short script, set the type, format and look in SceneMixer's Step 1, check the cast and places in Step 2, read the storyboard, then generate the segments and join them.

### How long can an AI movie be?

Each segment is up to 15 or 30 seconds and an episode or sequence holds up to 60 segments. SceneMixer's Film type has no hard cap, so longer movies are several sequences.

### How much does an AI movie cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join. A 3-minute movie is 180 seconds of that; reading and setup are small by comparison.

### Can I make one for free?

On SceneMixer, you can read the script, draw the first 15 setup images, get the first storyboard and a 5-second preview free. The video itself is paid.

### Do the characters talk?

Yes. Each character has a voice and speaks in the video, in any of 15 languages.

### Can I make it vertical?

Yes. Choose 9:16 Portrait in SceneMixer's Step 1 for a phone-first movie, or 16:9 Landscape.

### What if I do not have a script?

Type a one-line idea in AI Create and it drafts the script and episodes; you edit before anything is rendered.

Updated November 3, 2026


---

# AI Cartoon Maker: Cartoons With the Same Characters in Every Scene

URL: https://scenemixer.com/ai-cartoon-maker/
Make a cartoon with AI: pick a 3D toy, family 3D, claymation, watercolor or 2D look, draw the cast once, and get spoken multi-scene cartoons with the same characters.

# AI Cartoon Maker

Quick answer: An AI cartoon maker is only useful if the hero looks the same in scene two as in scene one. SceneMixer draws each character once in the look you choose (3D toy, family 3D, claymation, watercolor storybook, 2D cel and more), then every shot is generated from that drawing, with the voices spoken in the video. A cartoon here is a story of scenes and episodes, not a single looping clip: a short cartoon is a minute or two, a kids' series is a run of 2 to 5 minute episodes. Video costs 10 credits a second on the default tier; the first 15 setup images are free.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Pick a look, then everything follows it

The look is chosen in Step 1 and fixed once you continue; the characters, places and props are all drawn in it. Five that cartoon makers use most:

| Look | What it looks like | Try it for |
| 3D Toy Cartoon | Rounded, matte, toy-like figures in bright colours | Preschool, toy worlds |
| Family 3D Animation | Big-eyed, rounded characters with glowing skin | Family comedy, kids' series |
| Claymation | Hand-pressed clay figures and miniature sets | Gentle comedy, small worlds |
| Watercolor Storybook | Transparent watercolour and pencil, like a children's book | Bedtime stories, fables |
| Anime | Clean lines and flat colours, like TV anime | Series, action, romance |

Visual Style: Claymation

Visual Style: Watercolor Storybook

Visual Style: Anime

Visual Style: American Comic
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

All seventeen animated looks, one by one: AI animation generator · Animated looks on each tier.

## A cartoon in four steps

- Write the story or type one line. A cartoon wants one problem and a few repeating characters. A kids' bedtime series.
- Choose the look and the format. 16:9 for YouTube and TV, 9:16 for Shorts and Reels.
- Approve the cast. Each character sheet is drawn once; every later shot copies it.
- Read the storyboard, render, join. SceneMixer speaks dialogue in the video, one voice per character.

## From one cartoon to a channel

A single cartoon is a project; a channel is the same cast reused. Characters, scenes and voices can be carried into the next project, so episode two does not redraw the hero. Run a kids' cartoon YouTube channel · Reuse a cast across projects · Cartoon series generator.

## What it will not draw

- Existing cartoon characters and studio styles by name: the IP check refuses famous characters and brand mascots. Pick a look by its description, as above, and make your own cast.
- Celebrities, brand mascots and well-known characters are refused. Members can upload a photo of themselves and have it redrawn as a cartoon character: using your own photo.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### What is the best AI cartoon maker?

The one that keeps your characters the same across scenes. SceneMixer draws each character once and uses that drawing for every shot, with spoken dialogue.

### Can I make a cartoon for kids?

Yes. Looks such as 3D Toy Cartoon, Family 3D Animation and Watercolor Storybook suit children's stories, and SceneMixer can reuse the same cast across a series.

### Can it make a cartoon series?

Yes. On SceneMixer, episodes of 2 to 5 minutes are joined from up to 60 segments each, with the same characters carried across.

### Is there a free AI cartoon maker?

You can start free on SceneMixer: 50 credits on signup, the first 15 setup images, a free first storyboard and a 5-second preview. The video itself is paid.

### How much does a cartoon cost?

Video is 10 credits a second on SceneMixer's default tier (about $0.06 a second at the lowest credit price) plus 1 credit a second to join.

### Do the characters talk?

Yes, in any of 15 languages, one voice per character.

### Can I make a vertical cartoon?

Yes. Choose 9:16 Portrait in Step 1.

### Can I use famous cartoon characters?

No. Famous characters and brand mascots are refused; invent your own.

Updated November 3, 2026


---

# AI Anime Maker: Anime Scenes and Episodes With a Steady Cast

URL: https://scenemixer.com/ai-anime-maker/
Make anime with AI: TV-anime, painterly, watercolor, webtoon and cel-shaded 3D looks, a cast drawn once, spoken dialogue, and genres from shonen to slice of life.

# AI Anime Maker

Quick answer: SceneMixer makes anime as scenes and episodes, not a single moving picture. Choose one of the anime looks in Step 1 (Anime, Painterly Anime, Watercolor Anime, Korean Webcomic, Cel-shaded 3D), draw each character once, and every shot is generated from that drawing with the dialogue spoken in the video. The same pipeline reads a script, a web novel or a manga's text, so an anime episode of up to 300 seconds is a handful of segments joined into one file, at 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The anime looks

| Look | What it looks like | Try it for |
| Anime | Clean lines and flat colours, like TV anime | Series, shonen, romance |
| Painterly Anime | Lavish, saturated scenery; simple-lined characters | Films, emotional stories |
| Watercolor Anime | Watercolour-on-paper look, fine lines, natural colours | Gentle fantasy, slice of life |
| Korean Webcomic | Slender, polished characters, soft airbrushed shading | Webtoon romance and revenge |
| Cel-shaded 3D | Anime characters built in 3D, flat colour bands | Action anime, games |

Visual Style: Painterly Anime

Visual Style: Korean Webcomic

Visual Style: Cel-shaded 3D
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

How the looks hold up on each video tier · AI anime video generator.

## Genres people make

- Shonen and action: shonen anime and fight scenes that read clearly.
- Romance: shojo romance anime.
- Fantasy: isekai anime.
- Everyday life: slice-of-life anime.
- Feature length: an anime movie.

## From a manga, a webtoon or a web novel

If the story already exists as text, paste it. SceneMixer reads the characters, places and episode structure out of it and redraws everything as new shots; your own character sheets can be used as references. Scanned pages are not read, only text. Turn a manga or webtoon into an anime series · Turn a web novel into a series.

## Making it a channel

Anime channels need an episode a week and the same faces every time. Cast, scenes and voices carry from one project to the next, and each episode can be 9:16 for Shorts or 16:9 for a channel. An anime series for YouTube · AI anime series generator.

## Limits

- SceneMixer's IP check refuses existing anime characters and studio styles by name. Describe the look, as in the table, and invent the cast.
- Members can upload a photo and have the face, hair and clothing redrawn in the anime look as a character sheet. Using your own photo.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make anime?

Yes. SceneMixer generates anime scenes and episodes in several looks, with a cast drawn once and dialogue spoken in the video.

### Is there a free AI anime maker?

You can start free on SceneMixer: 50 credits, the first 15 setup images, a free first storyboard and a 5-second preview. Video is paid.

### How long can an anime episode be?

On SceneMixer, Short Drama runs up to 240 seconds, TV Series up to 300, Film without a hard cap; each episode holds up to 60 segments.

### Can I make an anime from my manga?

Yes, from its text: SceneMixer reads the characters, places and structure and redraws them. Scanned pages are not used.

### Does it keep the same character design?

Yes. Each character is drawn once in SceneMixer and that sheet is used for every shot.

### Can characters speak Japanese?

Yes. On SceneMixer, dialogue can be spoken in any of 15 languages, including Japanese, chosen in Step 1.

### Can I make vertical anime for Shorts?

Yes. Choose 9:16 Portrait in Step 1.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier (about $0.06 a second at the lowest credit price) plus 1 credit a second to join.

Updated November 3, 2026


---

# Book to Movie AI: Turn Your Book Into a Film or a Series

URL: https://scenemixer.com/book-to-movie-ai/
Turn a book you wrote into a movie with AI: chapters are read into a cast, places and sequences, drawn once, and rendered with spoken dialogue. Limits and costs.

# Book to Movie AI

Quick answer: Yes, with a book you wrote or have the rights to. Upload the text (.txt, .docx or .md) and SceneMixer reads it into characters, places, props and a sequence-by-sequence outline, draws each once, and renders the film with the dialogue spoken in the video. One project reads up to 300,000 characters of English, roughly 50,000 words; a longer book is split into volumes that share one cast. Set the type to Film for a single adaptation with no length cap, or TV Series for a season. Reading costs 4 credits per 1,000 characters (English counted lighter); the video is paid by the finished second.

## Film or series: decide before the book is read

| You want | Type | What it does to the book |
| One film from a short book or one story | Film | Sequences instead of episodes, no hard length cap, no forced hook; silence and subtext allowed |
| A season from a novel | TV Series | Episodes of up to 300 seconds, a closing beat on each |
| A fast vertical cut | Short Drama | Episodes of up to 240 seconds, conflict early, a hook to close each |

The type is part of the read, so choose it first. Short drama, series or film · Book to video: the full guide.

## What happens to a book

- It is read. SceneMixer lists the characters, places and props, and cuts the story into sequences or episodes with an outline for each. A long book is read in two passes. How long scripts are read.
- You check it. Merge two names that are one person, remove a place you do not need. Nothing is rendered yet.
- The cast is drawn once. Each character gets a full-body sheet; each place becomes a set. The first 15 setup images are free.
- Sequence one is made first. Storyboard, video and join for the first sequence; watch it before you pay for the next.

## Which books work

| Kind of book | Works as | Notes |
| Your own novel | A film, or a series of one or two chapters per episode | The usual case |
| A memoir or family history | A short film or a few episodes | Real people become invented faces unless you have their consent |
| A children's picture book | A short cartoon | A drawn look such as Watercolor Storybook |
| A public-domain classic | An adaptation in any style | Check the text is public domain where you publish |
| Non-fiction | Dramatised scenes | Works when there are people and moments; pure argument does not |

The book must be yours or licensed to you. Other people's characters, and well-known film versions of them, are refused by the IP check.

## Length and cost, honestly

Reading a book is cheap. Making it is not: on SceneMixer, the film is rendered at 10 credits a second on the default tier (about $0.06 a second at the lowest credit price) plus 1 credit a second to join, so a 90-minute film is a very large number of paid segments. The practical route is to adapt in stages: a trailer first, then the opening sequence, then the rest as the budget allows. A book trailer first · Choosing script length and cost.

## Files and languages

Paste the text or drop a .txt, .docx or .md file. PDFs and e-book files are refused; export or copy the text. The film can be cast for another market and spoken in any of 15 dialogue languages, so one book can have a second-language version. Which files work · Casting for another region.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI turn a book into a movie?

Yes, for a book you wrote or have the rights to. SceneMixer reads the text into characters, places and sequences and renders a film with spoken dialogue, one sequence at a time.

### How long a book can I upload?

SceneMixer accepts up to 300,000 characters of English per project, about 50,000 words. Longer books are split into volumes that share one cast.

### Which file types work?

.txt, .docx and .md, or pasted text. PDFs and e-book files are refused.

### Can I adapt someone else's book?

Only if you have the rights. SceneMixer's IP check refuses famous characters and other people's IP.

### How much does it cost?

Reading is 4 credits per 1,000 characters on SceneMixer, English counted lighter. Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Should I make a film or a series?

A short book or one story suits Film; a novel usually suits a TV Series season, one or two chapters per episode.

### Is there a free way to try it?

Paste chapter one into SceneMixer, read the summary and the first storyboard for free, and use the 5-second preview before you spend.

### Does it keep the narrator?

You can keep the narrator as voice-over lines in SceneMixer, or let scenes play with dialogue only.

Updated November 3, 2026


---

# Script to Animation AI: Turn a Script Into an Animated Video

URL: https://scenemixer.com/script-to-animation/
Turn a script into an animation with AI: scene headings become sets, character cues become voiced characters and action lines become shots in the look you pick.

# Script to Animation AI

Quick answer: Paste a script and choose an animated look: 2D anime, 3D toy cartoon, claymation, watercolor storybook and more. SceneMixer reads the scene headings as locations, the character cues as voiced characters and the action lines as shots, draws each character and place once in that look, and renders the dialogue in the video. A script written in plain screenplay form, or just paragraphs of prose, works; one project reads up to 300,000 characters of English. Video costs 10 credits a second on the default tier.

## What each part of your script becomes

| In the script | In the animation |
| Scene heading (INT. KITCHEN - NIGHT) | A location, drawn once in your look and used as the set whenever the scene returns |
| Character name above a line | A character with a full-body sheet and one voice, the same in every scene |
| Dialogue | Spoken in the video, in the language you set in Step 1 |
| Action lines | Numbered shots with seconds, framing and camera move in the storyboard |
| A prop that matters (a key, a letter) | An item drawn once, so it looks the same each time it appears |

Using scene headings and character cues · Writing dialogue so it is picked up.

## Choose how it looks

In SceneMixer, the look is one choice in Step 1, and the same script reads very differently in each:

| Look | Best for a script about… |
| Anime | Series, school, romance, action |
| 3D Toy Cartoon / Family 3D Animation | Kids, family comedy, small worlds |
| Claymation | Gentle comedy, tiny stakes, a town |
| Watercolor Storybook | Fables, bedtime stories, memory |
| Painterly Anime | A film with big scenery and quiet scenes |

Visual Style: Family 3D Animation

Visual Style: Claymation

Visual Style: Watercolor Storybook

Visual Style: Painterly Anime
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
All animated looks.

## Write a script that animates well

- One action per line. A run, a glance, a door. The storyboard turns each into a shot.
- Keep the cast small. Every character is another sheet and another voice; four or five carry a short.
- Let places repeat. A script with three recurring places is cheaper and steadier than one with twelve.
- Say it in the dialogue. Animation has no inner monologue; if a thought matters, give it to a line or a voice-over.
How to write one shot · Vertical script format.

## From the script to a file

Upload the script (.txt, .docx or .md) and read the cast, places and outline it produces. In SceneMixer, Step 2 draws the sheets; the first 15 setup images are free. Then the storyboard, which you can edit line by line for free, then video segment by segment, then the join at 1 credit a second. Upload a script and read it · Which files work.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### How do I turn a script into an animation?

Upload or paste it, choose an animated look in Step 1, check the cast and places, read the storyboard, then render the segments and join them.

### Does it understand screenplay format?

Yes: SceneMixer reads scene headings, character names above dialogue and action lines as locations, characters, speech and shots. Plain prose also works.

### What styles can I choose?

Anime, Painterly Anime, Watercolor Anime, Korean Webcomic, 3D Toy Cartoon, Family 3D, Claymation, Watercolor Storybook, Ink Wash and more.

### Do the characters speak?

Yes. Each character is voiced by SceneMixer and speaks in the video, in any of 15 languages.

### How long a script can I use?

Up to 300,000 characters of English per project, about 50,000 words.

### How much does it cost?

Reading is 4 credits per 1,000 characters. Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Is it free to try?

New accounts start with 50 credits; the first 15 setup images, the first storyboard and a 5-second preview are free.

### Can I animate a one-page script?

Yes. A one-page script is a minute or two of animation, and a good size for a first try.

Updated November 3, 2026


---

# AI Animated Story Generator: Turn a Story Into an Animated Video

URL: https://scenemixer.com/ai-animated-story-generator/
Turn a short story into an animated video with AI: one cast drawn once in a cartoon, anime or storybook look, scenes with spoken lines or voice-over, 2 to 5 minutes long.

# AI Animated Story Generator

Quick answer: Give SceneMixer a story, a few paragraphs or a few thousand words, and it comes back as an animated video: the characters and places are read out of the text, drawn once in the look you choose (storybook watercolor, 3D toy, claymation, anime), and each scene is generated with the lines spoken in the video. A story that fits 2 to 5 minutes is one episode of up to 60 segments; you can keep the narrator as voice-over lines or let scenes play with dialogue alone. Video costs 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What kind of story animates well

| Story shape | Why it works | A look to try |
| A fable with one animal and one lesson | Few characters, one place, a clear turn | Watercolor Storybook |
| A bedtime story with a gentle ending | Soft pacing, a voice-over carries it | 3D Toy Cartoon, Claymation |
| An adventure with a chase | Plenty of action lines to become shots | Family 3D Animation, Anime |
| A family comedy about a messy day | A small cast and one house | Family 3D Animation |
| A fantasy quest in three scenes | Each scene is a new place, drawn once | Painterly Anime |

Bedtime story series · Educational cartoons for kids.

## From the text to the screen

- Paste the story. Prose is fine; it does not have to be in script form. Or type one line in AI Create and read what it writes.
- Choose the look and the format. 16:9 for a channel or a screen, 9:16 for Shorts.
- Check the characters. Each gets a sheet drawn once; the places become sets. The first 15 setup images are free on SceneMixer.
- Read the storyboard. Narration, if you want it, appears as voice-over lines you can edit; so does every spoken line. Editing is free.
- Render and join. Segments are generated on their own, then joined at 1 credit a second.

## Voice-over or dialogue

A story written in the third person (“the fox stopped at the river”) can keep its narrator as voice-over lines, or you can rewrite the key moments as speech and let the characters say them. Most children's stories use a bit of both. Spoken lines are performed in the video, one voice per character, in any of 15 languages. Writing dialogue so it is picked up.

## Length and cost

A two-minute story is 120 seconds of video; at 10 credits a second on SceneMixer's default tier that is the main cost, about $0.06 a second at the lowest credit price, plus 1 credit a second to join. Reading the text is 4 credits per 1,000 characters. New accounts start with 50 credits and a 5-second preview is free. Where credits go in one episode.

Related: AI story video generator · Kids story videos · AI cartoon maker · Script to animation.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI turn a story into an animated video?

Yes. SceneMixer reads the text into characters and places, each drawn once in an animated look, and every scene is generated with spoken lines or voice-over.

### Does the story have to be a script?

No. Prose works on SceneMixer; scripts with scene headings and character names work too.

### Can it narrate the story?

Yes. In SceneMixer, narration appears as editable voice-over lines in the storyboard.

### How long will the video be?

A short story is usually 2 to 5 minutes, one episode of up to 60 segments. SceneMixer splits longer stories into episodes.

### Which looks are available?

Storybook watercolor, 3D toy cartoon, family 3D, claymation, anime, painterly anime and more.

### Can I make it for children?

Yes. Pick a gentle look and keep the story simple; SceneMixer keeps characters the same across scenes.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; reading is 4 credits per 1,000 characters, English counted lighter.

### Is there a free try?

New accounts start with 50 credits; the first 15 setup images, the first storyboard and a 5-second preview are free.

Updated November 3, 2026


---

# AI Animated Movie Generator: Feature-Style Animation From a Script

URL: https://scenemixer.com/ai-animated-movie-generator/
Make an animated movie with AI: set the type to Film, pick a 2D, 3D, anime or clay look, draw the cast once and render it one sequence at a time with spoken dialogue.

# AI Animated Movie Generator

Quick answer: An animated movie in SceneMixer is a Film project in an animated look: no hard length cap, divided into sequences, each storyboarded, rendered and joined on its own. The cast and the sets are drawn once, so a hero keeps the same design from the first sequence to the last, and the dialogue is spoken in the video. A five-minute short is 10 to 20 segments; a feature is hundreds, paid by the finished second at 10 credits a second on the default tier, so most people finish a trailer and one sequence before committing to the rest.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What animated movie length means in segments

| Length | Segments of 15 s | Segments of 30 s | Usually |
| 2 minutes | 8 | 4 | A trailer or a very short film |
| 10 minutes | 40 | 20 | An animated short |
| 30 minutes | 120 | 60 | A TV special, in sequences |
| 90 minutes | 360 | 180 | A feature, in many sequences |

Each segment is a paid generation, so a feature is a plan, not a button. The Film type lets a project run as long as the script; each sequence is billed and rendered separately. Choosing script length and cost.

## Pick a look for the whole film

SceneMixer fixes the look in Step 1 and every character, place and prop is drawn in it, so choose with the genre in mind:

| Look | Suits a movie about… |
| Painterly Anime | Big scenery, quiet scenes, emotional endings |
| Anime | Action, school, romance |
| Family 3D Animation / 3D Toy Cartoon | Family comedy, kids, a world of toys |
| Claymation | Gentle comedy in a small place |
| Realistic 3D | Sci-fi and fantasy epics |

Visual Style: Anime

Visual Style: Family 3D Animation

Visual Style: Claymation

Visual Style: Realistic 3D
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

An anime movie · A 3D cartoon movie · An animated short film.

## Work in sequences

- Trailer first. A 60 to 120 second trailer tests the look and the cast cheaply. Make a movie trailer.
- Sequence one. Storyboard, render and join the opening; watch it with the sound on.
- Fix the cast, not the shots. If a face is wrong, change the character sheet once in SceneMixer and the next shots follow.
- Continue. Add the next sequence in the same project, or start a new one with the same cast. Add or continue.

## Voices and music

Every character has a voice and speaks in the video, in any of 15 languages, so an animated movie can be made in Japanese, Spanish or Arabic as easily as in English. Each episode is joined with a background score. AI video generator with dialogue.

## What to expect

- It makes scenes and sequences; it is not a single prompt that returns a 90-minute film.
- Famous characters and studio styles by name are refused; describe the look instead.

Related: AI animation generator · AI movie generator · Long AI video generator · AI film making.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make an animated movie?

Yes, as a SceneMixer Film project in an animated look: a cast drawn once, a storyboard per sequence, and video rendered and joined one sequence at a time.

### How long can the movie be?

SceneMixer's Film type has no hard length cap; each sequence holds up to 60 segments of up to 15 or 30 seconds.

### How much does an animated movie cost?

Video is 10 credits a second on SceneMixer's default tier (about $0.06 a second at the lowest credit price) plus 1 credit a second to join, so cost scales with the finished minutes.

### Which styles can I use?

2D anime, painterly anime, 3D toy cartoon, family 3D, claymation, realistic 3D and more; the look is chosen in Step 1.

### Will the characters look the same throughout?

Yes. SceneMixer draws each character once, and that sheet is used for every shot.

### Can the characters speak?

Yes, in any of 15 languages, one voice per character.

### Where should I start?

With a trailer and the first sequence, then decide how much of the rest to make.

Updated November 3, 2026


---

# AI Long Story Video Generator: Stories of 5 to 30 Minutes

URL: https://scenemixer.com/ai-long-story-video-generator/
Make a long story video with AI: 5 to 30 minutes built from up to 60 segments per episode, one cast drawn once, dialogue or voice-over, for YouTube and story channels.

# AI Long Story Video Generator

Quick answer: A long story video is not one long generation. SceneMixer builds it from segments of up to 15 or 30 seconds, up to 60 per episode, each generated from the same character and place drawings and joined into one file. That makes 5 minutes one episode, and 30 minutes a run of episodes or sequences. The cost is the finished seconds at 10 credits a second on the default tier; the storyboard that plans it is free to read and edit.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## How a long story is planned

| Story length | Segments (15 s / 30 s) | How to build it |
| 5 minutes | 20 / 10 | One TV Series episode |
| 10 minutes | 40 / 20 | Two TV Series episodes (up to 300 seconds each), or three Short Drama episodes |
| 20 minutes | 80 / 40 | Four to five episodes of one story |
| 30 minutes | 120 / 60 | A series of episodes or Film sequences |

Short Drama episodes run up to 240 seconds and TV Series up to 300, so a story longer than that is cut into numbered parts that share one cast. How length is built · A 5-minute video.

## Make the length a story, not a stretch

- Give every part a turn. A 20-minute story as five 4-minute parts needs five small reversals, not one long middle.
- End parts on a question. It makes the next one click. Hooks and cliffhangers.
- Keep the cast small and the places few. That is what makes a long story steady and affordable.
- Let a voice carry the gaps. A voice-over line covers a time skip better than a shot.

## Generate a long run without watching it

SceneMixer renders segments on the server, so you can queue many at once, up to your plan's limit, and close the page. When they are done, join the episode. Episodes can be storyboarded and rendered in any order, and a single weak shot is regenerated on its own. Segment versions · Why a long story stays consistent.

## Where long stories are posted

A 10 to 20 minute story fits a YouTube channel; the same story cut into 1 to 3 minute parts fits Shorts and Reels in 9:16. Decide before the script is read: the format is fixed in Step 1. A faceless YouTube channel · Vertical video.

## Cost, plainly

Ten minutes is 600 seconds of video: that many seconds at 10 credits a second on SceneMixer's default tier, plus 1 credit a second to join, plus a little for retakes. A series of ten 90-second episodes from a 30,000-character novel comes to about 10,026 credits, about $62 at the lowest credit price. Spending fewer credits.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### How long can an AI story video be?

On SceneMixer, one episode holds up to 60 segments of up to 15 or 30 seconds. Longer stories are several episodes or sequences with the same cast.

### Can I make a 10 minute story video?

Yes: 20 to 40 segments, as one TV Series episode of up to 300 seconds plus another, or as a Film sequence.

### Does the character look the same all the way through?

Yes. SceneMixer draws each character once and uses it for every segment of every episode.

### How much does a long story video cost?

Finished seconds at 10 credits a second on the default tier plus 1 credit a second to join.

### Can I make it free?

Reading, the first 15 setup images, the first storyboard and a 5-second preview are free on SceneMixer; the video is paid.

### Can I post it on YouTube?

Yes. Choose 16:9 Landscape in SceneMixer's Step 1. Members can turn the corner mark off in downloaded files; a shared link always carries it.

### Can I add voice-over?

Yes. SceneMixer writes narration as voice-over lines in the storyboard.

Updated November 3, 2026


---

# AI TV Series Creator: Plan a Season and Make Every Episode

URL: https://scenemixer.com/ai-tv-series-creator/
Create a TV series with AI: a season outline, a cast kept across episodes, a storyboard per episode and spoken video up to 300 seconds each, with one price per episode.

# AI TV Series Creator

Quick answer: To create a TV series in SceneMixer you set the type to TV Series, give it a script, a novel or one line, and it plans the season: characters, places, props and an episode-by-episode outline, each episode with its conflict and its closing beat. Episodes run up to 300 seconds, each built from up to 60 segments, and the cast is drawn once and carried across the season. You make episode one first, watch it, and then decide how many more to make.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A season, plainly

| Season shape | Episodes | Episode length | Good for |
| Mini-series | 4 to 8 | Up to 300 s | One story told in parts |
| Short season | 10 | 90 to 300 s | A first season; the worked example below |
| Long season | 20+ | 60 to 300 s | A web series with weekly releases |

TV Series differs from Short Drama in its pacing: a slower opening, and a closing beat that can be an unfinished decision rather than a twist. Short drama, series or film.

## The cast carries the season

A series lives or dies on the faces staying the same. Every character is drawn once and used in every episode; places become recurring sets; voices stay the same. When you add episode eleven, you pick the existing cast from your library. Why a series stays consistent · Reuse a cast across projects · Consistent character video.

## Make the series in order

- Outline the season. Read the outline; merge, cut and reorder episodes before anything is drawn. Episode outline generator.
- Approve the cast once. The first 15 setup images are free on SceneMixer.
- Make episode one. Storyboard (free for your first drama), render, join, watch.
- Adjust, then continue. Fix what episode one taught you, then add the next. Add episodes or continue.

## What a season costs

The worked example is ten episodes of 90 seconds from a 30,000-character novel, at the cheapest video tier: about 10,026 credits, about $62 at the lowest credit price, reading and joins included. Longer episodes cost in proportion; every price is shown before you press the button. What a micro-drama costs · Choosing script length and cost.

## Where to take it next

Related: AI TV show generator · AI episode generator · AI short drama generator · Novel to video.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI create a TV series?

Yes. SceneMixer plans a season from a script, a novel or an idea, draws the cast once, and makes each episode as numbered shots rendered with spoken dialogue.

### How long is an episode?

On SceneMixer, TV Series episodes run up to 300 seconds, built from up to 60 segments.

### Do the characters stay the same across episodes?

Yes. SceneMixer draws each character once and uses it in every episode.

### How many episodes can I make?

As many as the script supports; you add them one at a time, and SceneMixer charges only for the ones you render.

### How much does a season cost?

About 10,026 credits (about $62 at the lowest credit price) for ten 90-second episodes at the cheapest video tier.

### Can I make a series from a novel?

Yes. SceneMixer reads a novel of up to 300,000 characters of English into an episode outline in one project.

### Is the first episode free?

On SceneMixer, the first episode's storyboard on your first drama is free, with a 5-second video preview; the video itself is paid.

Updated November 3, 2026


---

# AI Animated Series Maker: Cartoon and Anime Series, Every Week

URL: https://scenemixer.com/ai-animated-series-maker/
Make an animated series with AI: choose a cartoon or anime look, draw the cast once, and produce episodes on a weekly rhythm with the same characters, voices and sets.

# AI Animated Series Maker

Quick answer: An animated series needs one thing a clip generator cannot give it: the same characters and places in episode twelve as in episode one. SceneMixer draws the cast and the sets once in the look you choose (anime, 3D toy, family 3D, claymation, watercolor) and reuses them for every episode, with spoken dialogue in 15 languages. An episode is up to 240 to 300 seconds, so a weekly series is a repeatable job: write, read the storyboard, render, join, post.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Kinds of animated series people make

| Series | Look | Episode length | Format |
| Kids' cartoon channel | 3D Toy Cartoon, Family 3D, Claymation | 2 to 4 minutes | 16:9 |
| Anime serial | Anime, Painterly Anime | 3 to 5 minutes | 16:9, or 9:16 for Shorts |
| Family comedy | Family 3D Animation | 2 to 3 minutes | 16:9 |
| Fantasy saga | Painterly Anime, Realistic 3D | 4 to 5 minutes | 16:9 |
| Webtoon-style romance | Korean Webcomic | 1 to 3 minutes | 9:16 |

A kids' cartoon channel · An anime series for YouTube.

## A weekly rhythm

- Monday: write. Next episode's script, a page or two, or one line for AI Create to draft.
- Tuesday: read. Read the storyboard and fix shots; editing is free on SceneMixer.
- Wednesday: render. Queue the segments together and let them run on the server.
- Thursday: join and check. Join the episode; regenerate any weak shot on its own.
- Friday: post. 16:9 for the channel, or a 9:16 cut for Shorts and Reels.

Nothing in this rhythm redraws the cast, because the characters, places and voices are reused. Reuse across projects.

## Keeping the look steady for months

The look is fixed in Step 1 and every sheet is drawn in it, so the series does not drift. Characters keep their outfits unless you set a costume change, and a place is the same place every time it returns. The same room in every episode · Costume changes.

## Cost per episode

An episode is its finished seconds at 10 credits a second on SceneMixer's default tier, plus 1 credit a second to join: a 90-second episode is about ninety of those seconds. Ten 90-second episodes from a 30,000-character novel come to about 10,026 credits (about $62 at the lowest credit price). Retakes cost one shot, not one episode. What a clip fix costs.

Related: Cartoon series generator · Anime series generator · AI cartoon maker · AI anime maker.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make an animated series?

Yes. SceneMixer draws the cast and sets once in an animated look and reuses them, so episodes keep the same characters, voices and places.

### How long is an episode?

Short Drama episodes on SceneMixer run up to 240 seconds and TV Series up to 300, each built from up to 60 segments.

### Can I make a weekly series?

Yes. The weekly loop is write, read the storyboard, render, join and post, with the cast reused.

### Which looks are available?

Anime, painterly anime, Korean webcomic, 3D toy cartoon, family 3D, claymation, watercolor storybook and more.

### How much does an episode cost?

Its finished seconds at 10 credits a second on the default tier plus 1 credit a second to join.

### Can the series be vertical?

Yes. Choose 9:16 Portrait in SceneMixer's Step 1 for Shorts and Reels.

### Do I need to redraw characters each episode?

No. Pick them from your library in SceneMixer when you add the next episode.

Updated November 3, 2026


---

# AI Short Movie Maker: Make a Short Movie of 2 to 10 Minutes

URL: https://scenemixer.com/ai-short-movie-maker/
Make a short movie with AI: 2 to 10 minutes, a cast drawn once, a storyboard you can edit, spoken dialogue and one joined file. Cost and structure explained.

# AI Short Movie Maker

Quick answer: A short movie in SceneMixer is a Film project of 2 to 10 minutes: you write or generate a short script, the cast and places are drawn once, a storyboard lays out numbered shots, and each segment of up to 15 or 30 seconds is rendered with its dialogue and sound, then joined into one file. A 5-minute short is 10 to 20 segments, about 300 seconds of video at 10 credits a second on the default tier. Everything before the video, including the first storyboard and a 5-second preview, is free or small.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Shape a short movie in three parts

| Part | Share of a 5-minute short | Segments (15 s) | What happens |
| Setup | About 1 minute | 4 | Who, where, what they want |
| Complication | About 3 minutes | 12 | It goes wrong; a choice is forced |
| Turn and last look | About 1 minute | 4 | The choice, its price, one closing image |

That is 20 segments, or 10 on a long-take tier. Keep to one place and two or three characters for a first short. A step-by-step short film · AI short film maker.

## Short movie, short film, short drama

The three words are used for nearly the same thing, but SceneMixer's Type setting in Step 1 treats them differently. Film has no forced hook and lets a scene end on a look. Short Drama assumes a phone and closes every unit on a hook. TV Series sits in between. Pick Film for a movie you want people to sit through, Short Drama for one that has to survive a scroll. Short drama, series or film.

## A first short, step by step

- Type one line in AI Create, or paste a page of script.
- Choose Film, 16:9 or 9:16, a look, and the dialogue language.
- Approve the characters and places; the first 15 setup images are free on SceneMixer.
- Read the storyboard and cut any shot you would not shoot.
- Render the segments, join them at 1 credit a second, and watch the whole thing once with the sound on.
The free preview · Composite and share.

## Cost of a short

Video is the cost: 120 seconds for a 2-minute short, 300 for a 5-minute short, 600 for a 10-minute one, each at 10 credits a second on the default tier (about $0.06 a second at the lowest credit price), plus 1 credit a second to join. A single weak shot is regenerated alone. New accounts start with 50 credits. Spending fewer credits.

## Festivals and sharing

Shorts are made to be shown. Check each festival's rules on AI-assisted work and runtime before you submit, and export at the resolution it asks for. Making a short for a festival · AI movie making · AI film making.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### How do I make a short movie with AI?

Write or generate a short script, set the type to Film in SceneMixer, check the cast and places, read the storyboard, then render the segments and join them.

### How long can a short movie be?

Typically 2 to 10 minutes: 8 to 40 segments of 15 seconds, half that on a 30-second long-take tier.

### How much does a short movie cost?

Its finished seconds at 10 credits a second on the default tier plus 1 credit a second to join.

### Is it free?

On SceneMixer, reading, the first 15 setup images, the first storyboard and a 5-second preview are free; the video is paid.

### Can it have dialogue?

Yes, spoken in the video in any of 15 languages, one voice per character.

### Can I make it vertical?

Yes. Choose 9:16 Portrait in Step 1.

### Can I enter it in a festival?

Many festivals accept AI-assisted work, but rules differ; check the festival's own entry rules.

Updated November 3, 2026


---

# AI Movie From a Prompt: Turn One Line Into a Movie You Can Edit

URL: https://scenemixer.com/ai-movie-from-prompt/
Make an AI movie from a prompt: type one line in AI Create, read the script, cast and storyboard it drafts, edit anything, then render. The prompt starts the film.

# AI Movie From a Prompt

Quick answer: Type one line into AI Create and SceneMixer drafts the rest: a script, a cast, the places and a numbered storyboard, all as plain text you can read and rewrite before anything is rendered. That is the difference between a movie from a prompt and a clip from a prompt: the prompt starts a project that has characters, sets and episodes, and the video is generated shot by shot from them. Reading the draft and editing it are free; video costs 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What one line becomes

| You type | You get back | You can change |
| A one-line idea | A script split into episodes or sequences | Every line, scene and ending |
| … which names the people | A cast with a written look and a voice each | Faces, ages, outfits, names |
| … and the places | A list of locations, each drawn once | Add, remove or merge places |
| … and the length | A storyboard of numbered shots with seconds | Order, framing, dialogue |

SceneMixer renders nothing until you press the video button, and shows its price first. A tour of the editor · What you can change.

## Prompts that make better movies

- Name a person and a want. “A night-shift nurse who needs to find a missing patient before dawn” beats “a hospital thriller”.
- Give it one place. A single setting drawn once is steadier and cheaper than five.
- Say how it ends, or how it should feel. The draft otherwise decides for you.
- Say the length. A 2-minute short and a 20-minute film get different structures.
Try: A two-minute short: a widowed lighthouse keeper receives a radio call in her late husband's voice; one room, one storm, one line at the end.

## From the draft to a movie

- Read the script and the cast. Cut anything you do not want.
- Choose Film, TV Series or Short Drama, 16:9 or 9:16, a look and the dialogue language in SceneMixer's Step 1.
- Approve the character and place drawings; the first 15 setup images are free.
- Read the storyboard; the first episode's storyboard on your first drama is free.
- Render the segments and join them at 1 credit a second.

## What a prompt cannot do

A prompt will not fix a weak idea, and it will not hold a two-hour plot: past a few thousand words the draft is better built from your own script. It also will not use a real person's face or a famous character; those are refused. If you already have a script or a book, start there: script to movie · book to movie · text to movie.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make a movie from a prompt?

Yes. One line in SceneMixer's AI Create drafts a script, cast, places and storyboard; you edit them, then render the movie shot by shot.

### How long should the prompt be?

One line is enough to start. A sentence that names a person, a want and a place gives a better draft.

### Is it free?

On SceneMixer, drafting, reading and editing the storyboard are free for your first drama, with the first 15 setup images and a 5-second preview; the video itself is paid.

### How much does the movie cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; the price is shown before you start.

### Can I change what the AI wrote?

Yes. Every line of the script and every shot of the storyboard is editable in SceneMixer.

### Do the characters talk?

Yes, in any of 15 languages, one voice per character.

### Can I make it vertical?

Yes. Choose 9:16 Portrait in Step 1.

Updated November 3, 2026


---

# Story to Movie AI: Turn Your Story Into a Short Film or a Series

URL: https://scenemixer.com/story-to-movie-ai/
Turn a story you wrote into a movie with AI: characters and places are read from the text, drawn once, and shot as scenes with spoken dialogue. No script needed.

# Story to Movie AI

Quick answer: Paste a story, a few paragraphs or a few thousand words, and SceneMixer reads it into characters, places and scenes, draws each once, and renders the movie with the dialogue spoken in the video. You do not have to format it as a script: prose works. Choose Film for a single piece or TV Series to split a longer story into episodes. Reading is 4 credits per 1,000 characters (English counted lighter); the video is paid by the finished second.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a story needs to become a movie

| In your story | In the movie |
| Who it is about | A character with a full-body drawing and one voice |
| Where it happens | A place drawn once and used as the set each time it returns |
| What they say | Dialogue spoken in the video, in the language you choose |
| What they do | Numbered shots with seconds and framing in the storyboard |
| What they think | Voice-over lines, or a look held a beat longer |

Inner thoughts are the one thing a camera cannot show. Either keep them as a narrator's voice-over, or turn the important ones into spoken lines. Writing dialogue so it is picked up.

## Story length and what it turns into

| Your story | Make it as |
| A short story that fits one sitting | Film: one project, sequences if it needs them |
| A longer story with chapters | TV Series: one chapter or two per episode, up to 300 seconds each |
| A fast, hook-driven cut for phones | Short Drama: up to 240 seconds an episode, a hook to close each |

One SceneMixer project reads up to 300,000 characters of English. A whole book · A web novel.

## Three things to do before you press go

- Read the cast list. Merge a nickname that is the same person as a full name.
- Cut what the screen does not need. Descriptions of the weather are fine; a paragraph of backstory is a voice-over or nothing.
- Test with the first scene. Make the opening alone, watch it, then decide how much more to make.
Upload and read a text · The free preview.

## Rights

The story must be yours or licensed to you. Other people's characters, and well-known film versions of stories, are refused by the IP check; invent your own cast or use a public-domain text. What you get to keep.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI turn my story into a movie?

Yes. Paste the story into SceneMixer; characters and places are read from it, drawn once, and shot as scenes with spoken dialogue.

### Does my story need to be a screenplay?

No. Prose works. Scripts with scene headings and character names also work.

### How long can the story be?

Up to 300,000 characters of English per project, about 50,000 words.

### How much does it cost?

Reading is 4 credits per 1,000 characters. Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Can I keep the narrator?

Yes, as voice-over lines in the storyboard.

### Can I try it free?

Yes: SceneMixer gives 50 credits on signup, the first 15 setup images, the first storyboard and a 5-second preview.

### Can it be in another language?

Yes. In SceneMixer, dialogue can be spoken in any of 15 languages.

Updated November 3, 2026


---

# Script to Film AI: Turn a Screenplay Into a Finished Film

URL: https://scenemixer.com/script-to-film-ai/
Turn a screenplay into a film with AI: scene headings become sets, character cues become voiced cast, action lines become shots, and the Film type drops the forced hook.

# Script to Film AI

Quick answer: Upload a screenplay (.txt, .docx or .md) and set the type to Film. SceneMixer reads the scene headings as locations, the character cues as a cast with one voice each, the dialogue as speech and the action lines as shots, then draws each person and place once and renders the film sequence by sequence. Film pacing allows silence and half-spoken lines, and there is no forced hook or length cap; up to 300,000 characters of English, about 50,000 words, are read in one project.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a screenwriter should know

| Your screenplay | What SceneMixer does with it |
| INT./EXT. scene headings | Each distinct place becomes a drawn location |
| Character names above dialogue | Each name becomes one character, one voice |
| Action lines | Cut into numbered shots with seconds, framing and camera move |
| Dialogue | Spoken in the video in the language you set in Step 1 |

Using scene headings and character cues · Script length caps and how characters count.

## Film type: what changes for a screenplay

Short drama assumes a phone and a hook to close every unit. Film assumes someone who has chosen to sit down. Under Film, the read splits your screenplay into sequences rather than episodes, dialogue is allowed to land sideways, and a scene can end on a look. Set it before the read, because it is the read that carries the rule. How the Film type works · Short drama, series or film.

## A screenplay's first week here

- Day one: upload, read the cast and places, fix the list. Nothing is rendered.
- Day two: approve the drawings; the first 15 setup images are free.
- Day three: read the storyboard for sequence one; rewrite any shot that does not match your intent.
- Day four: render sequence one and watch it. This is your pilot.

## Cost and honesty

Reading is 4 credits per 1,000 characters. Video is 10 credits a second on the default tier, plus 1 credit a second to join; a feature is a very large number of paid segments, so many writers make a trailer and one sequence first. Make a trailer first · A pilot or pitch trailer.

Related: Script to movie · Script to video · AI film making · A movie from a prompt.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI turn a screenplay into a film?

Yes. In SceneMixer, scene headings, character cues, dialogue and action lines are read into locations, a voiced cast and shots, then rendered sequence by sequence.

### Which file types work?

.txt, .docx and .md, or pasted text. PDFs are refused.

### How long can the screenplay be?

Up to 300,000 characters of English per project, about 50,000 words.

### Does it keep my dialogue?

Yes. Each line is spoken as written in SceneMixer, by the character it belongs to.

### What is different about the Film type?

SceneMixer's Film type has no hard length cap, no forced hook, sequences instead of episodes, and room for silence and subtext.

### How much does it cost?

Reading is 4 credits per 1,000 characters. Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Is it good for a pitch or a pilot?

Yes. Many writers make a trailer or the first sequence as proof of concept.

Updated November 3, 2026


---

# AI Film Script Generator: Draft a Film Script You Can Shoot

URL: https://scenemixer.com/ai-film-script-generator/
Generate a film script with AI: one line becomes a script with scenes, a cast and dialogue, read back as a storyboard you can edit, and then shot in the same place.

# AI Film Script Generator

Quick answer: Describe the film in one line in AI Create and SceneMixer drafts a script: scenes, characters, dialogue and an ending, split into sequences you can edit. Unlike a standalone script generator, the script does not stop on the page: the same project turns it into a cast, places and a numbered storyboard, and then into video. Drafting and reading are free to try; if you already have a script, upload it instead.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## From one line to a script

| You give | The draft contains | You decide |
| Genre and a one-line premise | Scenes, characters and dialogue | What to keep, cut or reorder |
| Who it is about | Named characters with a written look and voice | Names, ages, relationships |
| How long it runs | Sequences or episodes sized to it | Add or merge sequences |
| Where it takes place | A short list of locations | Fewer places is steadier |

Choose Film to draft with film pacing: no forced hook, room for silence. AI script generator · Episode outline generator.

## A script you can read like a script

The draft is plain text. Read it start to finish before you build anything: does each scene turn, does each line belong to the person who says it, does the ending cost the hero something? Rewrite whatever fails; editing is free. When it works, SceneMixer turns it into the storyboard without it being retyped. What you can change · Writing dialogue in the storyboard.

## If you already have a script

Upload it. The read lists the characters, places and props it finds and outlines the sequences. Upload and read · Script to film · Which files work.

## What it is not

- It is not a replacement for a writer's judgment: a draft is a starting point to cut and rewrite.
- It will not write with real people's likenesses or well-known characters.
- It drafts for the screen; a stage play or a novel chapter is better pasted in as your own text. Stage play to short film.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI write a film script?

Yes. One line in SceneMixer's AI Create drafts scenes, characters, dialogue and an ending that you can edit.

### Is the script formatted like a screenplay?

The script SceneMixer writes is plain text with scenes and named speakers; paste it into your own tool if you need strict screenplay formatting.

### Can I turn the script into a film here?

Yes. In SceneMixer, the same project reads it into a cast and places, writes the storyboard and renders the video.

### Is it free?

Drafting and editing are free to try on your first SceneMixer drama, with 50 credits on signup; video is paid.

### Can I use my own script instead?

Yes. Upload .txt, .docx or .md to SceneMixer.

### What languages can it write in?

You can write and read in your own language, and SceneMixer can speak dialogue in 15 languages.

### How long a script can it handle?

Up to 300,000 characters of English per project, about 50,000 words.

Updated November 3, 2026


---

# AI Actor Generator: Create a Cast That Looks the Same in Every Scene

URL: https://scenemixer.com/ai-actor-generator/
Create AI actors for a film or series: each character gets a full-body design and a voice, reused in every shot and episode. Invented faces only, no real people.

# AI Actor Generator

Quick answer: An AI actor generator is only worth it if the actor stays the same. In SceneMixer each character is designed once, as a full-body sheet from a written description (or, for members, a reference photo), given one voice, and reused in every shot and every episode, so the lead in scene forty is the lead in scene one. The cast is invented: real people's faces and famous characters are refused. The first 15 setup images on an account are free.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What an AI actor is here

| Part | How it is made | Where it is used |
| Face and body | A full-body sheet drawn from a written description in your chosen look | Every shot the actor is in |
| Outfit | Costume style set once; changes are explicit | Across scenes and episodes |
| Voice | A voice assigned to the character | Every spoken line |
| Role | Name, age, relationships and behaviour from the script | The storyboard's blocking and lines |

AI character sheet generator · AI character voices · Picking voices.

## Casting, step by step

- Describe the role. Age, build, hair, one distinguishing detail, how they carry themselves.
- Choose the look. Realistic film, anime, 3D cartoon or claymation; the same description reads differently in each.
- Approve the sheet. Redraw it if the face is wrong; this is where to fix a face, not in shot forty.
- Give the voice. Pick a preset voice in SceneMixer; members can upload their own sample for a voice.
- Reuse. The actor goes into your library and into any project that follows.
Reuse a cast across projects.

## Your own photo as a reference

Members can upload a reference photo (uploading is a membership feature on the current plan settings). Optimize into a reference sheet reads the face, hair and clothing from the photo and redraws it in the project's look on a clean background, so the identity details come from your photo and the style from the project. Celebrity, brand and franchise images are refused. Using your own photo · The IP check.

## Keeping a cast steady for a whole series

SceneMixer uses the sheet as the reference for every shot, so the face and outfit follow. If a character must change (a disguise, a wound, a new costume), set it as a state rather than redescribing them. Costume changes and states · Why a series stays consistent · Consistent character video.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI create actors for a film?

Yes. On SceneMixer, each character is designed as a full-body sheet from a description, given a voice, and used in every shot.

### Are the actors real people?

No. SceneMixer invents the faces. Real people and famous characters are refused by the IP check.

### Will an actor look the same in every scene?

Yes. The sheet is the reference for every shot and every episode on SceneMixer.

### Can I use my own face?

Members can upload a reference photo; SceneMixer reads the face, hair and clothing and redraws them as a character sheet in the project's look. Celebrity and franchise images are refused.

### Do the actors speak?

Yes. Each has a voice and speaks in the video, in any of 15 languages.

### Is it free?

The first 15 setup images on a SceneMixer account are free; new accounts start with 50 credits.

### Can I reuse a cast in a new project?

Yes. Characters, scenes and voices carry over from your library in SceneMixer.

Updated November 3, 2026


---

# AI Movie Scene Generator: Make One Scene, or a Whole Film From Scenes

URL: https://scenemixer.com/ai-movie-scene-generator/
Generate a movie scene with AI: describe one scene, get a handful of shots with framing, camera move, dialogue and sound, then extend it into a film when ready.

# AI Movie Scene Generator

Quick answer: A movie scene in SceneMixer is a handful of numbered shots that share a place and a cast: where the camera is, who is in frame, what is said, how long each holds. You can make a single scene on its own, a segment of up to 15 or 30 seconds, with the dialogue and sound generated in the same pass, or make it the first scene of a longer film. A scene costs its finished seconds at 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a scene is made of

| Part of the scene | Written as | Example |
| Place and time | A location and a time of day | A diner at 3 a.m., rain on the glass |
| Who is there | Characters from your cast, with positions | She is at the counter; he is by the door |
| Shots | Numbered shots with seconds, framing and camera move | Shot 3, 4 s: close on her hands, slow push-in |
| Lines | Dialogue tagged to the speaker | HE: “You came back.” |
| Sound | Ambience and music in the same pass | Rain, a fridge hum, one low note |

How to write one shot · Everyone stays put · AI scene generator.

## Make one scene first

- Describe the scene in two or three sentences, or paste it from your script.
- Approve the two or three characters and the one place it needs.
- Read the shot list; add a close-up where the line lands, remove a shot that only repeats.
- Render the scene, then use the free 5-second preview first if you want to see the look.

## Scenes that work

- Two people and a table. Dialogue scenes are the steadiest.
- A chase through one corridor. One place, plenty of action lines.
- A reveal. One line, one reaction shot, a held beat.
- A quiet scene. Weather, a look, a door; film pacing allows it.
Dialogue scenes · Scenes with several characters · Fight scenes.

## From a scene to a film

In SceneMixer, scenes are the unit of a film: each is storyboarded and rendered on its own, so you can make the scene you care about most first and add the rest around it. The cast and place you drew for the scene are reused. Add the next part · AI film making.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI generate a movie scene?

Yes. Describe the scene; SceneMixer writes it as numbered shots with framing, lines and sound, and each segment is rendered with dialogue and sound in one pass.

### How long is a scene?

On SceneMixer, a segment is up to 15 or 30 seconds, and a scene is one or a few segments.

### Can I make just one scene?

Yes. A scene is a complete project of its own.

### How much does a scene cost?

Its finished seconds at 10 credits a second on the default tier, plus the join if there is more than one segment.

### Do the characters talk?

Yes. SceneMixer speaks lines in the video, in any of 15 languages.

### Can I add camera moves?

Yes. Camera moves are written into each shot in the storyboard.

### Is there a free preview?

Yes, about 5 seconds, once.

Updated November 3, 2026


---

# AI Movie Creator: Make a Whole Movie as One Person

URL: https://scenemixer.com/ai-movie-creator/
An AI movie creator for solo makers: write it, cast it, read the storyboard and render it yourself. What one person can finish, in what order, and what it costs.

# AI Movie Creator

Quick answer: As a solo movie creator you do four jobs in one place: write the script (or let AI Create draft it), approve the cast and the places, read and fix the storyboard, then render and join the shots. Nothing needs a crew, but everything needs your decisions, because the tool only draws what you approve. Video costs 10 credits a second on the default tier and everything before it is free or small, so a solo creator's real budget is the number of finished seconds.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The four jobs

| Job | What you decide | Where it happens |
| Writer | The story, the ending, who says what | Upload Script or AI Create; the storyboard text |
| Casting and design | Faces, outfits, places, the look | Step 1 and Step 2; the first 15 setup images are free |
| Director | Which shot, how long, what the camera does | The storyboard: numbered shots you can rewrite |
| Editor | What to redo, what to keep | Per-segment version history; joining the episode |

A tour of the editor · Segment versions.

## A sensible order for a first movie

- One page of story. A single place and two or three people. Write the ending first.
- Read the cast and places. Cut anyone who does not change the ending.
- Read the storyboard before you render. Editing it is free on SceneMixer, and it is where a movie is made or lost.
- Render the opening only. The free 5-second preview first; then the first sequence.
- Finish, then judge. Join, watch with the sound on, and fix single shots rather than whole scenes.

## What a solo creator should not do

- Do not plan a feature first. A 90-minute plan is hundreds of paid segments. Make a two-minute version and see what the storyboard teaches you.
- Do not add characters to rescue a weak scene. Every extra character is another sheet to approve and another voice.
- Do not fix a face in the video. Fix it on the character sheet, once.
Spending fewer credits · Choosing script length and cost.

## Cost of finishing one

Video is 10 credits a second on the default tier (about $0.06 a second at the lowest credit price), plus 1 credit a second to join. New accounts start with 50 credits. A ten-episode series of 90-second episodes from a 30,000-character novel comes to about 10,026 credits, about $62 at the lowest credit price. AI movie making · AI film making · AI short movie maker.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can one person make a movie with AI?

Yes. SceneMixer covers the script, cast, places, storyboard, shots and joining; the decisions stay with you.

### Do I need any experience?

No. The storyboard is plain text you can read and rewrite, and SceneMixer's free preview shows the look before you spend.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join. Reading and storyboard editing are free or small.

### How long can the movie be?

A segment is up to 15 or 30 seconds and an episode holds up to 60 segments; the Film type on SceneMixer has no hard length cap.

### Do the characters talk?

Yes, in any of 15 languages, one voice per character.

### Can I fix one bad shot?

Yes. SceneMixer regenerates a single shot on its own, and the earlier version stays in the history.

### Is it free to try?

New accounts start with 50 credits; the first 15 setup images, the first storyboard and a 5-second preview are free.

Updated November 3, 2026


---

# AI Film Production Software: From Script to Finished Cut

URL: https://scenemixer.com/ai-film-production-software/
AI film production software for small teams and solo makers: script breakdown, casting sheets, a storyboard, rendered shots, joins and previs. Each price shown first.

# AI Film Production Software

Quick answer: SceneMixer is AI film production software, not a production company: you bring the script and the decisions, and each stage of a production has a tool. The script is read into a breakdown, casting sheets are drawn, a shot list is written as a storyboard, shots are rendered with dialogue and sound, and the segments are joined into a cut. Every stage shows its price before it runs, and a failed generation is refunded.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Stage by stage

| Production stage | What you get | Cost shape |
| Script breakdown | Characters, locations, props and an outline | 4 credits per 1,000 characters |
| Casting and design | Character sheets, location and prop plates | First 15 setup images free |
| Shot list | A storyboard with numbered shots, seconds, framing, lines | Editing is free; the first episode's storyboard on your first drama is free |
| Shooting | Rendered segments with dialogue and sound | 10 credits a second on the default tier |
| Assembly | Segments joined into one file | 1 credit a second |

AI script breakdown · AI shot list generator · AI character sheet generator.

## Where it fits a real production

| Use | What you make | Start with |
| Previs | A rough, voiced version of a scene to test pacing before a shoot | Previs and animatics |
| Pitching | A pilot or a one-minute trailer from the same cast | A pilot or pitch trailer |
| Short films | A finished short from a page of script | A step-by-step short film |
| Series | A season with one cast and weekly episodes | A multi-episode series |

## Controls a producer cares about

- Price before each stage. SceneMixer shows the cost of every generation before you press the button. Where credits go in one episode.
- Refunds on failures. A generation that fails to deliver is not charged. What a clip fix costs and when you are refunded.
- Version history per segment. A retake never loses the take you liked.
- A reusable cast. Characters, places and voices carry into the next project. Reuse across projects.

## What it is not

It is not a studio or an outsourcing service, it does not record actors, and it will not use a real person's face or a famous character. It produces generated footage you direct through text. If you want a human crew, use this for previs and for pitching. AI film making · AI movie generator.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### What is AI film production software?

Software that generates a film's pictures and sound from your script and decisions: breakdown, casting sheets, storyboard, shots and assembly.

### Is SceneMixer a production company?

No. It is a tool; you write and direct, and it generates the footage.

### Can I use it for previs?

Yes. A rough voiced version of a scene is a common use.

### How is it priced?

Reading is 4 credits per 1,000 characters, English counted lighter; video is 10 credits a second on SceneMixer's default tier; joining is 1 credit a second.

### Are failed generations charged?

No. SceneMixer does not charge for a generation that fails to deliver.

### Can I reuse my cast?

Yes. SceneMixer carries characters, scenes and voices across projects.

### Does it work for a feature?

Yes, as sequences, paid by the finished second on SceneMixer; many teams make a trailer and the opening first.

Updated November 3, 2026


---

# AI Fight Scene Generator: Write and Render a Fight, Shot by Shot

URL: https://scenemixer.com/ai-fight-scene-generator/
Generate an AI fight scene: write numbered shots with framing and camera moves, keep both fighters the same in every shot, and render with impact sound and dialogue.

# AI Fight Scene Generator

Quick answer: A fight scene works in SceneMixer when it is written as shots, not as a paragraph: who is in frame, the camera, the single action, how long it holds. The two fighters are drawn once so they look the same from the first exchange to the last, and each segment of up to 15 or 30 seconds is rendered with its sound in one pass. A short fight is one to three segments at 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A fight in six beats

| Beat | Shot to write | Camera |
| 1 Standoff | Two figures, a held breath | Wide, slow push-in |
| 2 First exchange | A strike and a block | Medium, camera tracks the move |
| 3 Advantage | One fighter presses; the other gives ground | Over the shoulder, then close |
| 4 Reversal | A small mistake changes it | Close on the hands or the face |
| 5 Finish | One clean hit, or a decision not to | Low angle, then still |
| 6 Aftermath | Breath, a look, what it cost | Wide, held |

One action per shot. Fights read clearly when each shot shows one thing. How to write one shot · Anime fight scenes.

## Keep the fighters steady

- Two fighters, two sheets. Each is drawn once, with an outfit that does not change mid-fight.
- One place. A rooftop, a dojo, an alley: drawn once and used as the set.
- Positions carry across cuts. Who stands where stays consistent. Everyone stays put.
Writing fights, crowds and group scenes.

## Looks that suit a fight

Anime and cel-shaded 3D read action cleanly; painterly looks suit duels with weight; realistic film suits grounded brawls. A boxing or fight movie · An action movie ·

Visual Style: Cel-shaded 3D

Visual Style: Painterly Anime
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Sound and dialogue

SceneMixer generates impacts, footsteps and ambience in the same pass, and lines spoken mid-fight are performed by the character they belong to, in any of 15 languages. Content rules apply to what is generated, so write the fight as action and consequence. AI video generator with dialogue.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI generate a fight scene?

Yes. Write it as numbered shots; SceneMixer renders each segment with sound and dialogue in one pass.

### How long is a fight scene?

On SceneMixer, a segment is up to 15 or 30 seconds; a short fight is one to three segments.

### Will the fighters look the same in every shot?

Yes. SceneMixer draws each once, and that sheet is used for every shot.

### How much does it cost?

Its finished seconds at 10 credits a second on the default tier, plus the join if there is more than one segment.

### Which style is best for fights?

Anime and cel-shaded 3D read action cleanly; painterly and realistic looks suit heavier duels.

### Can I regenerate one shot?

Yes. On SceneMixer, a single shot is regenerated on its own.

### Is there a free preview?

Yes, about 5 seconds, once.

Updated November 3, 2026


---

# AI Film Editing Software: Fix, Trim and Join Generated Shots

URL: https://scenemixer.com/ai-film-editing-software/
AI film editing for generated films: one card per segment, trim, redub, enhance or regenerate a clip, then join and download. Where to use a full editor instead.

# AI Film Editing Software

Quick answer: SceneMixer edits generated footage, not filmed footage. The editor has a timeline with one card per segment, and each clip can be regenerated, trimmed, re-dubbed or enhanced, rewritten through AI edit, and rolled back through its version history; when the segments are right, one button joins them into an episode, without re-encoding where it can. It is not a multi-track suite: for layered footage, colour grading or effects, download the full-quality file and finish it in your editor.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What you can do to a clip

| Action | What it does | Good for |
| Regenerate | Renders the segment again from its storyboard | A face, a move or a line that came out wrong |
| AI edit | Rewrites the segment's text from a plain request | Changing what happens without retyping |
| Trim | Shortens the clip | Cutting a held beat |
| Redub | Replaces the voice or line | A line you want said differently |
| Enhance | Raises the quality of a clip | The best shots before export |
| Version history | Switches back to an earlier take | When the first take was better |

A tour of the editor · Segment versions.

## Joining the episode

SceneMixer's composite uses the active version of each segment and refuses to run until every segment is finished, so a half-shot episode cannot be joined by mistake. It is priced at 1 credit a second and the price is shown before you confirm. The download is the full-quality master; the player uses a lighter copy. Compositing and sharing.

## When to use a full editor too

- You want to mix generated shots with your own footage.
- You need colour grading, effects or a multi-track sound mix.
- You are cutting trailers from several episodes.
Download the master from the editor and finish there. Captions can be added after export. AI film making · AI video remake.

## Cost of a fix

On SceneMixer, a fix costs one shot, not one episode, and it tells you the price and the refund rule before you confirm. What each clip fix costs · Edit one shot and regenerate.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Is there AI film editing software?

SceneMixer edits the footage it generates: regenerate, trim, redub, enhance and join clips. For filmed footage or layered edits, use a full editor.

### Can I trim a clip?

Yes, with the Trim option under Edit this clip in SceneMixer.

### Can I change a line after the shot is made?

Yes, with Redub, or by rewriting the shot and regenerating it.

### Can I go back to an earlier take?

Yes. SceneMixer keeps a version history for every segment.

### How do I join the clips?

Press Composite Episode in SceneMixer. It joins the active version of each segment once all are finished.

### What quality is the download?

The full-quality master. The player and shared pages use a lighter copy.

### Does it replace my video editor?

No. SceneMixer finishes the generated episode; layered edits and grading happen in your own editor.

Updated November 3, 2026


---

# AI Movie Pitch Generator: A Pitch Trailer and Cast, Not a Slide Deck

URL: https://scenemixer.com/ai-movie-pitch-generator/
Pitch a movie with AI: a pilot or one-minute trailer, a drawn cast, a promo cover and a mood reel from one project. It makes the pictures for a pitch, not the deck.

# AI Movie Pitch Generator

Quick answer: SceneMixer does not write a pitch deck; it makes the things a pitch deck points at. From one project you get a drawn cast, a one-minute pitch trailer rendered from a second short script of your best five to eight moments, a pilot episode if you want to prove the story works, a promo cover for the title slide, and a mood reel of three or four key scenes. You write the logline and the deck in your own tool and link or embed these.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a pitch needs, and what makes it here

| In the pitch | Made here as | How |
| Proof the story works | A pilot episode | Type TV Series, upload the pilot script, render it |
| The idea in a minute | A pitch trailer | A second short script of five to eight moments, same cast |
| The look | A mood reel | Three or four key scenes in your chosen visual style |
| The title slide | A promo cover | Three poster options per generation |
| The people | Character sheets | Full-body designs from your descriptions |

A pilot or pitch trailer, step by step · AI movie poster generator · AI trailer generator.

## The trailer, in five beats

- The world. A wide shot that says where we are.
- The person. Who we follow, doing something characteristic.
- The problem. The thing that changes.
- The line. One spoken line that sells the premise.
- The title. A held shot and the name.
Use the same characters as the pilot by picking them from your SceneMixer library. Reuse characters.

## Be honest in the room

A pitch trailer made this way is previs: it shows tone, cast and pacing, not a finished broadcast, and it should be presented that way. It will not use a real person's face or a famous character, so casting stays invented. Previs and animatics.

## Cost

A one-minute trailer is sixty seconds of video at 10 credits a second on the default tier plus 1 credit a second to join; the cast is drawn once, with the first 15 setup images free; the price of every generation is shown first. Make a movie trailer.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make a movie pitch?

SceneMixer can make the pictures for one: a pilot, a one-minute trailer, a cast, a promo cover and a mood reel. The slide deck itself you write in your own tool.

### Does it generate a pitch deck?

No. SceneMixer does not produce slides or a written treatment.

### What is a pitch trailer?

A one-minute video of five to eight key moments, made with the same cast as your project.

### Can I make a poster for the deck?

Yes. SceneMixer's promo cover gives three poster options per generation.

### How much does a pitch trailer cost?

Sixty seconds of video at 10 credits a second on the default tier plus 1 credit a second to join.

### Is it a finished film?

No. Present it as previs: tone, cast and pacing.

### Can I use real actors' faces?

No. SceneMixer refuses real people and famous characters; the cast is invented.

Updated November 3, 2026


---

# AI Animation Maker: Animated Stories With a Cast That Stays the Same

URL: https://scenemixer.com/ai-animation-maker/
An AI animation maker for stories: pick a 2D, 3D, clay or watercolor look, draw the cast once, read a free storyboard and render scenes with voices.

# AI Animation Maker

Quick answer: An AI animation maker can mean four different tools, and it helps to know which one you are buying. SceneMixer is the story kind: you give it a script or an idea, choose an animated look, and it draws each character and place once, writes the shots as a storyboard and renders scenes with spoken dialogue. It does not animate a single picture, it is not a frame-by-frame drawing program, and it is not a template video maker; it makes scenes and episodes that keep the same characters.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Four kinds of animation tool

| Kind | You do | Output | SceneMixer is… |
| Frame-by-frame software | Draw or pose every frame yourself | Exact hand-made motion | Not this |
| 3D software | Model, rig and animate | Full control, long learning curve | Not this |
| Picture animators | Upload an image and add motion | A short moving clip | Not this |
| Story generators | Give a script or idea | Scenes and episodes with a steady cast | This |

AI animation generator: all seventeen looks · Script to animation.

## The looks

Choose once in Step 1: Anime, Painterly Anime, Watercolor Anime, Korean Webcomic, Family 3D Animation, 3D Toy Cartoon, Claymation, Watercolor Storybook, Ink Wash Animation and more. Every character, place and prop is drawn in the look you pick.

Visual Style: Family 3D Animation

Visual Style: Claymation

Visual Style: Watercolor Storybook

Visual Style: Ink Wash Animation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

2D · 3D · Stop-motion and clay · Comic to video.

## How a scene is made

- Write or paste the story; one line in AI Create also works.
- Approve the characters and places in your look; the first 15 setup images are free on SceneMixer.
- Read the storyboard and rewrite shots; editing is free.
- Render the segments at 10 credits a second on the default tier and join them at 1 credit a second.

## Free to begin

New accounts start with 50 credits, and the cast drawings, the first storyboard and a 5-second preview are free. Free AI animation maker · What you get for free.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### What is an AI animation maker?

A tool that generates animated video. SceneMixer is a story kind: scripts and ideas become scenes and episodes with a steady cast.

### Does it animate a single picture?

No. SceneMixer generates scenes from a script; it is not a picture animator.

### Which animation styles can I make?

Anime, painterly anime, Korean webcomic, family 3D, 3D toy cartoon, claymation, watercolor storybook, ink wash and more.

### Can I make a series?

Yes. SceneMixer reuses the cast and sets across episodes.

### Do the characters speak?

Yes, in any of 15 languages.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Is there a free version?

You can start free on SceneMixer: 50 credits, the first 15 setup images, the first storyboard and a 5-second preview.

Updated November 3, 2026


---

# AI Cartoon Story Generator: Cartoon Stories for Kids and Channels

URL: https://scenemixer.com/ai-cartoon-story-generator/
Make cartoon stories with AI: a simple story, a cast of animals or kids drawn once in a cartoon look, scenes with voices, and a series of episodes for YouTube or bedtime.

# AI Cartoon Story Generator

Quick answer: A cartoon story is a small cast, one problem and a gentle ending, and SceneMixer makes it as scenes: the characters are drawn once in a cartoon look, each scene is rendered with their voices, and the episode is joined into one file. A story of 2 to 4 minutes is one episode; a run of them is a channel. Pick a look such as 3D Toy Cartoon, Family 3D, Claymation or Watercolor Storybook, and keep the same heroes for every story.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Four story shapes that work

| Shape | Example | Look |
| The lesson | A fox learns to share a berry | Watercolor Storybook |
| The helper | A small robot fixes a town's clock | 3D Toy Cartoon |
| The mix-up | Two siblings swap lunchboxes | Family 3D Animation |
| The journey | A snail crosses the garden | Claymation |

A bedtime story series · Educational cartoons ·

Visual Style: Claymation

Visual Style: Watercolor Storybook
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Make it a series

SceneMixer draws the hero, the friends and the home once and reuses them in every story, so episode ten has the same faces as episode one. Voices are kept too. Posting weekly is a loop of write, read the storyboard, render, join. Run a kids' cartoon YouTube channel · Reuse a cast · AI animated series maker.

## Keep it kind and simple

- One idea per story. Two scenes and a closing image beat six scenes.
- Short lines. Voices carry short lines best.
- Invent your cast. Existing cartoon characters and mascots are refused by the IP check.
Kids story videos · AI cartoon maker.

## Cost

A 3-minute cartoon is 180 seconds of video: that many seconds at 10 credits a second on SceneMixer's default tier plus 1 credit a second to join. Drawing the cast is free for the first 15 images, and a 5-second preview is free. Where credits go in one episode.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make a cartoon story?

Yes. SceneMixer turns a story into scenes with a cast drawn once, spoken in the video, joined into an episode.

### Is it suitable for kids?

You choose a gentle look and write the story; keep it simple and kind. SceneMixer applies content rules to what it generates.

### Can I make a series?

Yes. Reuse the same cast and voices across stories.

### How long is a cartoon episode?

On SceneMixer, a story of 2 to 4 minutes is one episode, built from up to 60 segments.

### Which look is best for toddlers?

3D Toy Cartoon and Watercolor Storybook are soft, bright looks that suit young children.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Can I post it on YouTube?

Yes. Choose 16:9 Landscape in SceneMixer's Step 1.

Updated November 3, 2026


---

# AI Mini Series Maker: A 4 to 8 Episode Story With One Cast

URL: https://scenemixer.com/ai-mini-series-maker/
Make a mini series with AI: plan 4 to 8 episodes with an arc, draw the cast once, storyboard each episode and render them in order with the same faces, voices and places.

# AI Mini Series Maker

Quick answer: A mini series is one story told in four to eight parts, and the plan matters more than the tool: each episode ends on a turn and the last one pays off the first. SceneMixer reads the story into an episode outline, draws the cast once, and lets you make the episodes one at a time, up to 240 to 300 seconds each, at 10 credits a second on the default tier. You can stop after any episode and pay only for what you rendered.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A six-episode arc

| Episode | Job | Ends on |
| 1 | Meet the person and the problem | A decision that cannot be undone |
| 2 | First attempt | It works, but costs something |
| 3 | Complication | A secret comes out |
| 4 | Midpoint turn | The goal changes |
| 5 | Low point | The plan fails |
| 6 | Payoff | A final image that answers episode 1 |

Hooks and cliffhangers · Episode outline generator.

## Make it in order

- Read the outline for all episodes and fix the arc before anything is drawn.
- Approve the cast and the two or three places that recur.
- Make episode one; read its free storyboard first (free on your first drama).
- Watch it, adjust the cast if needed, then continue. Add episodes or continue.

## What six episodes cost

On SceneMixer, six episodes of 90 seconds are 540 seconds of video: that many seconds at 10 credits a second plus 1 credit a second to join each episode. The worked example in the full pipeline is ten 90-second episodes from a 30,000-character novel: about 10,026 credits, about $62 at the lowest credit price. What a micro-drama costs.

Related: AI TV series creator · AI episode generator · AI short drama generator · AI micro drama.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### What is a mini series?

A story told in about four to eight episodes with one arc.

### Can AI make a mini series?

Yes. SceneMixer reads the story into an episode outline, and each episode is storyboarded and rendered with the same cast.

### How long is each episode?

Up to 240 seconds as a Short Drama and up to 300 as a TV Series.

### Can I stop after a few episodes?

Yes. SceneMixer charges only for the episodes you render.

### Will the cast stay the same?

Yes. SceneMixer draws each character once and uses them in every episode.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join each episode.

### Is episode one free?

Its storyboard is free on your first SceneMixer drama, with a 5-second preview; the video is paid.

Updated November 3, 2026


---

# AI Sitcom Generator: Fixed Sets, Running Gags and a Steady Cast

URL: https://scenemixer.com/ai-sitcom-generator/
Make an AI sitcom: a small ensemble in one or two sets, a running gag, spoken dialogue and a punchline every episode, with the same cast and rooms in every episode.

# AI Sitcom Generator

Quick answer: A sitcom is a perfect fit for a tool that draws each character and place once: the cast is small, the sets repeat and the jokes live in dialogue. In SceneMixer you set up the ensemble and the one or two rooms, write each episode as a few scenes of talk, and the lines are spoken in the video by the character they belong to. A 2 to 3 minute episode is a handful of segments at 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The sitcom shape

| Part | In a 3-minute episode | What to write |
| Cold open | About 20 seconds | A visual gag or a bad plan |
| Complication | About 90 seconds | Two people want opposite things in the same room |
| Escalation | About 50 seconds | The plan makes it worse |
| Tag | About 20 seconds | A callback to the cold open |

Make an AI sitcom or comedy skit · AI dialogue generator.

## Why sets and ensembles are cheap and steady

The kitchen, the office, the bar: each is drawn once and used in every episode, so the series feels like a place. Four or five characters with clear voices carry the show, and each has one thing they always do. The same room in every episode · Reuse a cast.

## Writing jokes that render

- Put the joke in the line. SceneMixer speaks the dialogue exactly as written.
- Write the reaction as a shot. A look to camera is its own shot with its own seconds.
- Keep the physical gags simple. One object, one action.
Writing dialogue in the storyboard.

## Cost per episode

A 3-minute episode is 180 seconds of video at 10 credits a second on the default tier plus 1 credit a second to join. The cast and the sets are drawn once for the whole season, with the first 15 setup images free. AI TV show generator.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make a sitcom?

Yes. A small ensemble and one or two sets are drawn once, and each episode is written as scenes of dialogue that are spoken in the video.

### Does it have a laugh track?

SceneMixer generates no laugh track; the jokes live in the lines and the timing.

### How long is an episode?

Typically 2 to 3 minutes, built from a handful of segments.

### Will the cast and sets stay the same?

Yes. SceneMixer draws each once and reuses it in every episode.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Can I make it vertical?

Yes. Choose 9:16 Portrait in Step 1.

### Is there a free preview?

Yes, about 5 seconds, once.

Updated November 3, 2026


---

# AI Story Maker: Write the Story, Then Watch It Play as Video

URL: https://scenemixer.com/ai-story-maker/
An AI story maker that goes past text: write or generate a story, check the cast and places it finds, then turn it into scenes with voices and one video.

# AI Story Maker

Quick answer: Most story makers stop at words. SceneMixer starts with the words and keeps going: type one line in AI Create (or paste your own story), read the draft, fix what you do not like, and the same project draws the characters and places and turns the story into scenes with spoken dialogue. The writing and the editing are free to try; the video is paid by the finished second at 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## From one line to a story you can watch

| Step | What you do | What you get |
| 1 Prompt | Type one line, or paste a story | A draft split into episodes or scenes |
| 2 Read | Fix lines, merge names, cut scenes | A clean cast list and outline |
| 3 Draw | Approve the character and place drawings | A cast that looks the same in every scene |
| 4 Storyboard | Read the shots, rewrite what is wrong | Numbered shots with seconds and lines |
| 5 Render | Generate the segments and join them | One video file |

A movie from a prompt · Story to movie · AI story video generator.

## Make the story easier to film

- Give it a person who wants something. The draft improves immediately.
- Limit the places. Fewer places are cheaper and steadier.
- Turn thoughts into lines or voice-over. A camera cannot show a thought.
Write in your own language.

## What it does not do

SceneMixer does not publish a book or format a manuscript, and it will not write with real people's likenesses or famous characters. It is for stories you intend to see. AI film script generator.

## Free to try

New accounts start with 50 credits; the draft, the cast drawings (the first 15 images), the first storyboard and a 5-second preview are free. What you get for free.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI write a story and turn it into a video?

Yes. SceneMixer's AI Create drafts the story, and the same project turns it into scenes with spoken dialogue.

### Can I use my own story?

Yes. Paste it as prose or a script.

### Can I edit what the AI wrote?

Yes. In SceneMixer, every line and shot is editable, for free.

### Is it free?

Drafting and editing are free to try; video is paid by the finished second.

### How long can the story be?

Up to 300,000 characters of English per project.

### Do the characters talk?

Yes, in any of 15 languages.

### What does the video cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

Updated November 3, 2026


---

# Free AI Movie Maker: What You Can Make Free, and Where Payment Starts

URL: https://scenemixer.com/free-ai-movie-maker/
A free start for an AI movie: 50 credits on signup, the cast drawings, a first storyboard and a short preview are free. Here is exactly where the free part ends.

# Free AI Movie Maker

Quick answer: You can start a movie in SceneMixer without paying, and it is worth knowing where the free part ends. New accounts get 50 credits, the first 15 setup images are free, the first episode's storyboard on your first drama is free, and a preview of about 5 seconds is free. The finished video is paid by the second at 10 credits a second on the default tier, and on a free account every video carries a small AI label in the corner.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What is free, and what is not

| Step | Free? | Notes |
| Sign up | Yes | 50 credits to start |
| Draft or paste a script | Reading costs 4 credits per 1,000 characters | The welcome credits cover a short script |
| Character and place drawings | The first 15 images are free | Per account |
| Storyboard | Free for the first episode of your first drama; editing is free | Read it before you render |
| Preview | About 5 seconds, once | Shows the look and the voices |
| The finished video | Paid | 10 credits a second on the default tier |

What you get for free · The free preview · Free sheets are per account.

## A free first session

- Paste a page of script or type one line in AI Create.
- Approve the cast and places using the free images.
- Read and edit the free storyboard.
- Run the free preview to see and hear a few seconds.
- Decide whether to render the rest. Nothing is charged until you press a paid button, and its price is shown first.

## The AI label

On a free account every video carries a small AI label in the corner. A paid membership turns it off by default; a video shared through a SceneMixer link always shows it. Compositing and sharing.

## If you want to go on

Credits are sold in packs and as membership plans, and the price of every generation is shown before you start. Free AI drama generator · AI movie making · Pricing.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Is there a free AI movie maker?

On SceneMixer, you can start free: 50 credits, the first 15 setup images, a free first storyboard and a 5-second preview. The finished video is paid.

### Can I make a whole movie for free?

No. The video itself is paid by the finished second, so a full movie needs credits.

### Does the free version have a watermark?

On SceneMixer, videos on a free account carry a small AI label in the corner; a paid membership turns it off by default.

### Where do I see the price before I pay?

Every paid button shows its price first; nothing is charged until you confirm.

### How much is a paid second?

A paid second on SceneMixer is 10 credits a second on the default tier, about $0.06 a second at the lowest credit price.

### Is the free preview long?

SceneMixer's free preview is about 5 seconds, once, enough to see the look and hear the voices.

### Do I need a card to start?

No card is needed to read the script, draw the cast and read the storyboard.

Updated November 3, 2026


---

# Free AI Animation Maker: Test a Look and a Cast Before You Pay

URL: https://scenemixer.com/free-ai-animation-maker/
Try an AI animation maker free: pick a look, draw the cast with the free images, read a free storyboard and see a short preview. The video itself is paid.

# Free AI Animation Maker

Quick answer: The free part of SceneMixer is exactly what an animator needs to decide: does the look suit my story, and do my characters read well? You get 50 credits on signup, the first 15 setup images free (enough to draw a small cast in the animated look you choose), a free first storyboard, and a preview of about 5 seconds. The finished animation is paid by the second at 10 credits a second on the default tier.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## A free test of a look

- Pick a look in Step 1: anime, painterly, 3D toy, family 3D, claymation, watercolor storybook.
- Paste a page of story. Reading a short script is covered by the welcome credits.
- Draw the cast. The first 15 images are free; if a look does not suit the story, you find out here.
- Read the storyboard and the free 5-second preview.
- Switch looks in a new project and compare the casts before you render a second of video.

Visual Style: Family 3D Animation

Visual Style: Claymation

Visual Style: Watercolor Storybook
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## What stays paid

| Free | Paid |
| Welcome credits, cast drawings (first 15), first storyboard, 5-second preview | Video segments at 10 credits a second on the default tier |
| Editing the storyboard | Joining the episode at 1 credit a second |
| Rewriting a shot's text | Regenerating a shot |

What you get for free · The free preview.

## Looks worth testing first

For kids' stories try 3D Toy Cartoon and Watercolor Storybook; for series try Anime or Family 3D Animation; for a small world try Claymation. AI animation maker · All animated looks · AI cartoon maker.

## The label on a free account

On SceneMixer, videos on a free account carry a small AI label in the corner; a paid membership turns it off by default. Compositing and sharing.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Is there a free AI animation maker?

You can start SceneMixer free: 50 credits, the first 15 setup images, a free first storyboard and a 5-second preview. The animated video is paid.

### Can I make a full animation free?

No. Video is paid by the finished second.

### What can I test for free?

The look, the cast, the storyboard and a short preview.

### Do I need a card?

No card is needed to draw the cast and read the storyboard.

### Which look should I test first?

The one that suits your genre; kids' stories suit 3D Toy Cartoon or Watercolor Storybook.

### How much is a paid second?

A paid second on SceneMixer is 10 credits a second on the default tier, about $0.06 a second at the lowest credit price.

### Does the free version have a watermark?

Videos on a free account carry a small AI label; a paid membership turns it off by default.

Updated November 3, 2026


---

# AI Movie From a Photo: What a Photo Can Do in a Movie Here

URL: https://scenemixer.com/ai-movie-from-photo/
Make an AI movie from a photo: use it as a character reference for a story, or up to four product photos for an ad. A photo does not become a slideshow or one clip.

# AI Movie From a Photo

Quick answer: People search for an AI movie from a photo and mean three different things, so here is what SceneMixer does with each. A photo of a person can be the reference for a character who then stars in a story: members upload it, and the face, hair and clothing are read and redrawn in your project's look. Up to four photos of a product become an ad. A photo does not become a slideshow or a single animated clip: the movie is a story with a script, scenes and spoken dialogue, and the photo feeds it.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Three things people mean

| You mean… | Here, that is… | Start with |
| Put me (or someone I know) in a movie | A character reference sheet made from the photo; members only | Using your own photo |
| Sell this product | An ad made from one to four photos | An ad from one photo |
| Make this picture move | Not what this does: scenes come from a script, not from a single still | AI video with dialogue |

Photos are accepted as JPEG, PNG or other common formats and are re-encoded in your browser before upload. What you can feed in.

## Starring someone: how a photo becomes a character

In Step 2, open a character card and press Upload & Replace. Optimize into a reference sheet reads the face, hair and clothing from the photo and redraws them in the project's look on a clean background, so the identity details come from your photo and the style from the project. That sheet is then used in every shot. Uploading is a membership feature on the current plan settings. Celebrity, brand and franchise images are refused. The IP check.

## From the photo to a movie

- Upload the reference and approve the sheet in your chosen look.
- Write or generate a script around that character (AI Create is a good start).
- Read the storyboard, then render and join the segments at 10 credits a second on the default tier.
A movie from a prompt · AI actor generator · A birthday story video.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make a movie from a photo?

A photo can be the reference for a character who stars in a story, or product photos can become an ad. SceneMixer does not turn a photo into a slideshow or a one-clip animation.

### Can I put myself in a movie?

Members can upload a photo, and SceneMixer redraws the face, hair and clothing as a character sheet in the project's look.

### Is photo upload free?

On SceneMixer, uploading a reference photo is a membership feature on the current plan settings; photos for an ad are the exception.

### Can I upload a celebrity's photo?

No. Celebrity, brand and franchise images are refused.

### Can I animate a single picture?

Not on SceneMixer: movies are made from a script with scenes and dialogue.

### How many photos can I use for an ad?

Up to four photos of what you sell.

### What formats work?

JPEG, PNG and other common formats; SceneMixer does not accept SVG.

Updated November 3, 2026


---

# AI Interactive Story Generator: Linear Stories, Not Branching Games

URL: https://scenemixer.com/ai-interactive-story-generator/
Looking for an AI interactive story generator? SceneMixer makes linear animated stories with voices, not branching games. How to make several versions of one story.

# AI Interactive Story Generator

Quick answer: An interactive story lets the reader choose; SceneMixer does not. It makes a story that plays from start to finish, as scenes with characters, voices and sound. If what you want is a story that comes alive on screen, that is exactly the job. If you want choices and branches, the closest thing here is making two or three versions of the same story with the same cast and offering them as separate videos.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What you get, and what you do not

| Interactive fiction tool | SceneMixer |
| The reader picks a path | The story is fixed; the viewer watches it |
| Text on screen | Spoken dialogue and sound in the video |
| Branches in a tree | One timeline; versions are separate videos |
| No pictures needed | Characters and places are drawn once and reused |

AI story video generator · AI story maker.

## Making versions of one story

- Write the shared opening and make it once.
- Write two or three endings as separate short episodes in the same project, with the same cast.
- Render each ending and post them as separate videos, or as a poll followed by the chosen ending.
Because the cast and places are drawn once, every version looks like the same story. Add episodes or continue · Reuse a cast.

## If you need real branching

A game or interactive-fiction engine is the right tool for choice logic. SceneMixer can supply the characters, the scenes and the voiced cut-scenes that you drop into one. Scenes for games · A game trailer or cutscene.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Is SceneMixer an interactive story generator?

No. It makes linear animated stories with voices; the viewer watches rather than chooses.

### Can I make a choose-your-own-adventure?

SceneMixer does not make a branching player. You can make separate videos for each ending with the same cast.

### Can it write the story for me?

Yes. One line in SceneMixer's AI Create drafts a story you can edit.

### Are there voices?

Yes, in any of 15 languages, one voice per character.

### Can I use the scenes in a game?

You can download the finished video and use it as a cut-scene in your own project.

### Is it free to try?

New accounts on SceneMixer start with 50 credits; the first 15 setup images and a 5-second preview are free.

### How much does a video cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

Updated November 3, 2026


---

# AI Picture Book to Video: Turn a Children's Book Into an Animation

URL: https://scenemixer.com/ai-picture-book-to-video/
Turn a picture book into a video with AI: paste the text, pick a storybook look such as watercolor or claymation, draw the characters once, and add voices and sound.

# AI Picture Book to Video

Quick answer: A picture book is short text and a lot of pictures, and SceneMixer works from the text. Paste the words of a book you wrote or have the rights to, choose a look such as Watercolor Storybook or Claymation, and the characters and places are drawn once and every page becomes a scene with narration or spoken lines. It does not read the book's own illustrations: scanned pages are not used, so the pictures are new, in the look you choose.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Page by page

| In the book | In the video |
| A page of text | A shot or two, with the narrator as voice-over |
| The hero in every picture | One character sheet used in every shot |
| The recurring place (the garden, the house) | A location drawn once and reused |
| Repeated phrases (“And then…”) | Spoken lines that return the same way |

Narration is kept as voice-over lines you can edit.

Visual Style: Claymation

Visual Style: Family 3D Animation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Rights and illustrations

Use a book you wrote or hold the rights to, or a public-domain text. SceneMixer draws the pictures fresh from the text. What you get to keep · Book to video.

## Length and cost

A short picture book is usually a few minutes of video: those seconds at 10 credits a second on the default tier plus 1 credit a second to join. The first 15 setup images are free, so you can draw the cast and look before you pay. A bedtime story series · Kids story videos · AI animated story generator.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI turn a picture book into a video?

Yes, from its text: SceneMixer draws the characters and places once in a storybook look, and each page becomes a scene.

### Does it use the book's illustrations?

No. The pictures are drawn new from the text; scanned pages are not used.

### Can it narrate the book?

Yes. The narrator can be kept as voice-over lines.

### Which look suits a picture book?

Watercolor Storybook, Claymation and 3D Toy Cartoon are common choices.

### Can I adapt someone else's book?

Only with the rights to the text; SceneMixer refuses well-known characters.

### How long is the video?

Usually a few minutes for a short picture book.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

Updated November 3, 2026


---

# AI Audiobook to Video: Turn a Story You Narrated Into Scenes

URL: https://scenemixer.com/ai-audiobook-to-video/
Turn an audiobook or spoken story into video with AI: SceneMixer takes the text, not an audio file, and makes scenes with voices. How to go from recording to transcript.

# AI Audiobook to Video

Quick answer: SceneMixer takes three kinds of input: text, up to four photos for an ad, or a video to remake. It does not take an audio file, so an audiobook or a recorded story goes in as its transcript. From the text, the characters and places are drawn once, the narrator stays as voice-over lines, and the people in the story speak in the video. The voices are generated; the recording is not reused.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## From a recording to a video

- Get the text. Use the book's manuscript, or a transcript of your recording from a speech-to-text tool.
- Tidy it into scenes. Cut ads, intros and housekeeping; keep the story and the narrator's best lines.
- Upload as .txt, .docx or .md. Read the cast and places.
- Choose the look and voices. Narration becomes voice-over; characters get their own voices.
- Render and join. Priced by the finished second.
Podcast episode into a drama · Which files work.

## What changes from the audio

| Your audio | The video |
| A narrator reading | Voice-over lines, in a generated voice |
| Different voices for characters | A voice per character, chosen or matched |
| Sound effects and music | Ambience and a score generated for the episode |
| Pauses and pacing | Set by the storyboard's shot seconds |

If your own voice matters, SceneMixer lets members upload a voice sample for a character. Picking voices.

## Rights and cost

Use text you wrote or have the rights to. Reading is 4 credits per 1,000 characters; video is 10 credits a second on the default tier plus 1 credit a second to join. Novel to video · Book to movie.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can I upload an audiobook?

No. SceneMixer takes text, up to four photos for an ad, or a video to remake. An audiobook goes in as its text.

### How do I get the text?

Use the manuscript, or transcribe the recording with a speech-to-text tool and tidy it.

### Will it use my narration?

No. SceneMixer turns the narrator into generated voice-over lines; members can upload a voice sample for a character.

### Can it add sound effects?

SceneMixer generates ambience and a score for the episode.

### Which files can I upload?

.txt, .docx and .md.

### How much does it cost?

Reading is 4 credits per 1,000 characters, English counted lighter; video is 10 credits a second on SceneMixer's default tier.

### Is it free to try?

New accounts on SceneMixer start with 50 credits, with the first storyboard and a 5-second preview free.

Updated November 3, 2026


---

# AI Fan Film Maker: Make Your Own Universe, Not Someone Else's

URL: https://scenemixer.com/ai-fan-film-maker/
Make a fan film or fan fiction video with AI using original characters in your own world: draw the cast once, write scenes, add voices. Franchise characters are refused.

# AI Fan Film Maker

Quick answer: A fan film here is a film in a world you love, made with characters you invent. SceneMixer refuses well-known characters, franchises and brand mascots, so you cannot cast an existing hero; what you can do is set an original character in the same kind of world, tone and genre. Describe the look and the setting, draw the cast once, write the scenes, and render with voices.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What works, what is refused

| You want | Works? | Do this instead |
| Your own characters in a world like one you love | Yes | Describe the genre, tone and setting in your own words |
| A character from a film, game or book | No: refused by the IP check | Invent a hero with the same role and a different name and design |
| Fan fiction you wrote | Yes, with original characters | Paste it as prose; swap named characters for your own |
| Your friends as the cast | Members, via reference photos | Using your own photo |

The IP check · AI actor generator.

## Make a fan-style film in six steps

- Write the premise: a world, a rule that bends, a person who breaks it.
- Invent the hero and one rival with their own look and voice.
- Pick a look: anime, painterly, claymation or live-action.
- Read the storyboard and cut anything that leans on a copyrighted detail.
- Render the opening scene first.
- Add episodes with the same cast.

## Genres fans make

Space-opera crews, school-of-magic stories, post-apocalypse road films, detective duos, monster hunters. Each can be done with original names and designs. A monster movie · Isekai anime · AI animated series maker.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can I make a fan film with AI?

Yes, with original characters. Existing franchise characters are refused.

### Can I use a character from my favourite show?

No. SceneMixer's IP check refuses famous characters, franchises and brand mascots.

### Can I adapt my fan fiction?

Yes, if you swap existing characters for ones you invent.

### Can my friends be in it?

Members can upload reference photos, and SceneMixer redraws them as characters.

### Which style can I use?

Anime, painterly anime, claymation, 3D cartoon, live-action looks and more.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Is there a free try?

New accounts on SceneMixer start with 50 credits; the first 15 setup images and a 5-second preview are free.

Updated November 3, 2026


---

# Free AI Story Video Generator: What Is Free and Where Payment Starts

URL: https://scenemixer.com/free-ai-story-video-generator/
Start a story video free: 50 credits on signup, free character drawings, a free first storyboard and a short preview. The finished video is paid by the second.

# Free AI Story Video Generator

Quick answer: You can begin a story video without paying: new accounts get 50 credits, the first 15 setup images are free, the first episode's storyboard on your first drama is free and editing storyboards is free, and a preview of about 5 seconds is free. The finished video is paid by the second at 10 credits a second on the default tier, and on a free account every video carries a small AI label.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Free first, in order

- Paste a story or type one line in AI Create.
- Read the cast and the places it finds.
- Draw the characters with the free images.
- Read and edit the storyboard.
- Watch the free preview. Then decide whether to render the rest.
What you get for free · The free preview.

## A free story video, exactly

| Part | Free? |
| Welcome credits | Yes: 50 |
| Character and place drawings | The first 15 images |
| Storyboard for the first episode (first drama) | Yes; editing text is free |
| Video preview | About 5 seconds, once |
| The whole story video | Paid by the finished second |

Free AI movie maker · AI story video generator · Free AI drama generator.

## The label and the price

Videos on a free account carry a small AI label; a paid membership turns it off by default, and a shared link always shows it. On SceneMixer, each paid second is 10 credits a second on the default tier, about $0.06 at the lowest credit price. Compositing and sharing.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Is there a free AI story video generator?

On SceneMixer, you can start free: 50 credits, the first 15 setup images, a free first storyboard and a 5-second preview. The finished video is paid.

### Can I make a whole story video for free?

No. The video is paid by the finished second.

### Does it have a watermark?

Videos on a free account carry a small AI label; a paid membership turns it off by default.

### Is the story editor free?

Yes. Editing the script and storyboard text in SceneMixer is free.

### Do I need a card?

No card is needed to read the story, draw the cast and read the storyboard.

### How long is the free preview?

About 5 seconds, once.

### How much is a paid second?

A paid second is 10 credits a second on the default tier.

Updated November 3, 2026


---

# Free AI Short Film Maker: Start Free, Pay Only to Render

URL: https://scenemixer.com/free-ai-short-film-maker/
Start an AI short film free: 50 credits on signup, a free cast and storyboard, a short preview, then the render paid by the second. See where the free part ends.

# Free AI Short Film Maker

Quick answer: A short film starts free in SceneMixer and costs money only when you render video. New accounts get 50 credits, the first 15 setup images are free, the first storyboard on your first drama is free, and you can watch a preview of about 5 seconds. A 3-minute short is 180 seconds of video at 10 credits a second on the default tier plus the join.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a free short film session gives you

- A script, written by you or drafted by AI Create.
- A cast and places, drawn with the free images, in the look you chose.
- A numbered storyboard you can rewrite shot by shot.
- A free preview showing the look and the voices.
That is usually enough to decide whether the film works before you spend. A step-by-step short film · AI short film maker.

## Where payment starts, in numbers

| Length | Seconds of video | Segments (15 s / 30 s) |
| 1 minute | 60 | 4 / 2 |
| 3 minutes | 180 | 12 / 6 |
| 5 minutes | 300 | 20 / 10 |
| 10 minutes | 600 | 40 / 20 |

On SceneMixer, each second is 10 credits a second on the default tier, plus 1 credit a second to join. Where credits go in one episode.

## Make it cheap

- One place and two or three characters.
- Fix a bad shot alone rather than redoing a scene.
- Use the preview before the first render.
Spending fewer credits · Free AI movie maker.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Is there a free AI short film maker?

On SceneMixer, you can start free: 50 credits, free cast drawings, a free first storyboard and a 5-second preview. The finished video is paid.

### Can I finish a short film for free?

Not entirely; the video is paid by the finished second.

### How much is a 3-minute short?

180 seconds of SceneMixer video at 10 credits a second on the default tier plus 1 credit a second to join.

### Does a free account add a label?

Yes, SceneMixer adds a small AI label on every video; a paid membership turns it off by default.

### Can I enter a festival?

Check the festival's rules on AI-assisted work and the label.

### Can I edit the storyboard free?

Yes. Editing text in SceneMixer is free.

### Do I need a card?

No card is needed to start.

Updated November 3, 2026


---

# AI Video Generator for Storytelling: Stories, Not Clips

URL: https://scenemixer.com/ai-video-generator-for-storytelling/
An AI video generator for storytelling: characters and places drawn once, an editable storyboard, scenes with spoken dialogue and episodes joined into one video.

# AI Video Generator for Storytelling

Quick answer: Storytelling needs three things a clip generator does not give: people who stay the same, places that stay put, and an order. SceneMixer starts from the story. The cast and the places are drawn once, the storyboard lays out the shots in order and can be edited, and scenes are rendered with the dialogue spoken. For short stories, serial stories and long ones, it joins the segments into one video.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Clips vs stories

| | A clip generator | A storytelling generator |
| Unit | One prompt, one clip | A script, scenes and episodes |
| Characters | Re-described every time | Drawn once, reused |
| Order | You stitch clips together | A storyboard sets the order |
| Dialogue | Often added afterwards | Spoken in the generation |
| Fixing | Re-prompt and hope | Edit one shot and regenerate it |

AI story video generator · AI video generator with dialogue.

## Story shapes it handles

- Short story: one episode, 2 to 5 minutes. Animated story generator.
- Serial: a cast across many episodes. TV series creator.
- Long story: episodes joined into parts. Long story video.
- True or personal story: invented faces, or members' reference photos. A birthday story.

## Try it small

Write one page, approve the cast, read the free storyboard and watch the free 5-second preview. Video is 10 credits a second on the default tier plus 1 credit a second to join. New accounts start with 50 credits. What you get for free.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### What is the best AI video generator for storytelling?

One that keeps characters and places the same across scenes and lets you edit the order; SceneMixer draws the cast once and uses a storyboard.

### Does it write the story?

SceneMixer's AI Create can draft a story from one line; you can also paste your own.

### Do the characters talk?

Yes, in any of 15 languages.

### Can I make a long story?

Yes, as episodes or sequences with the same cast.

### Can I fix one bad shot?

Yes. Edit and regenerate a single shot in SceneMixer.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Is there a free try?

Yes: SceneMixer gives 50 credits, the first 15 setup images, a first storyboard and a 5-second preview.

Updated November 3, 2026


---

# ChatGPT to Video: Turn a ChatGPT Script Into a Voiced Video

URL: https://scenemixer.com/chatgpt-to-video/
Turn what ChatGPT wrote into a video: paste the script, and SceneMixer reads the cast and places, draws them once, storyboards the scenes and renders them with voices.

# ChatGPT to Video

Quick answer: A chatbot writes text; it does not make a consistent video. The route that works is to let the chatbot draft the story, then bring the text to SceneMixer: paste it, and the characters and places are read out, drawn once, laid out as a storyboard and rendered with spoken dialogue. Nothing connects the two services; you copy the text across. SceneMixer also has its own writer, AI Create, for when you want the draft and the video in one place.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The hand-off

- Ask for a script, not a poem. Characters with names, a few places, dialogue with speaker names, a clear ending.
- Copy it into a .txt or paste it. No link is read; the text itself goes in. What you can feed in.
- Read the cast and places. Merge names the chatbot varied.
- Choose the look, approve the drawings, read the storyboard.
- Render and join. Video is 10 credits a second on the default tier.

## A prompt that gives usable scripts

Write a three-scene script for a 2-minute animated short. Three named characters, two locations, dialogue with speaker names in capitals, one visual gag and a closing line. Keep each scene to six lines.

Short scenes and named speakers are what the read can use best. Writing dialogue so it is picked up · Scene headings and character cues.

## Why not stay in the chatbot

- A chat reply is not built to keep a face the same from one picture to the next.
- It does not give you a storyboard and a cast list to edit in place.
- It does not join clips into an episode.
That is the gap SceneMixer fills. A movie from a prompt · Script to video · AI script generator.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can ChatGPT make a movie?

It can write the script. A consistent video with a drawn cast, a storyboard and joined scenes is what you make by bringing the text to SceneMixer.

### How do I move a ChatGPT script into SceneMixer?

Copy the text and paste it, or save it as .txt, .docx or .md and upload it.

### Is it connected to ChatGPT?

No. You copy the text across.

### Can SceneMixer write the script itself?

Yes, with AI Create from one line.

### What should I ask the chatbot for?

Named characters, a few locations, dialogue with speaker names and a clear ending.

### How much does the video cost?

On SceneMixer, reading is 4 credits per 1,000 characters, English counted lighter; video is 10 credits a second on the default tier plus 1 credit a second to join.

### Is there a free try?

New SceneMixer accounts start with 50 credits; the first storyboard and a 5-second preview are free.

Updated November 3, 2026


---

# Midjourney to Video: Turn Your Image Style Into a Story With Voices

URL: https://scenemixer.com/midjourney-to-video/
Take a character or style you made in an image tool and build a story around it: upload it as a reference (members), draw the cast once, and render scenes with voices.

# Midjourney to Video

Quick answer: An image tool makes beautiful single pictures; a story needs the same character in forty of them. In SceneMixer you can bring a picture of a character you made in an image tool as a reference (uploading is a membership feature on the current plan settings): it is read for face, hair and clothing and redrawn as a reference sheet in your project's look, and that sheet is used in every shot. Then write the story, read the storyboard and render scenes with spoken dialogue.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What carries over, and what does not

| From your image tool | Carries over? | Notes |
| The character's face, hair, clothing | Yes, read into a reference sheet | Redrawn in the project's look |
| The exact art style of the image | Chosen from the project's looks instead | Pick the closest look in Step 1 |
| The background and pose | No, by default | Optimize drops them so they do not follow into every shot |
| Images of real people, celebrities, franchises | No | Refused by the IP check |

Using your own image as a reference · The IP check.

## The route

- Write the story, or draft it with AI Create.
- In SceneMixer's Step 2, press Upload & Replace on the character and choose Optimize into a reference sheet.
- Repeat for the other principals; draw the rest from text.
- Read the storyboard, render and join.
Use Use this image as is only for art you already drew as a full figure on a plain background.

## Why a story tool, not another prompt

Re-prompting an image tool for the same character in each scene drifts. A reference sheet used for every shot does not. Consistent character video · AI actor generator · AI comic to video.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can I turn Midjourney images into a video?

You can use an image of a character as a reference (members) and build a story around it; in SceneMixer, scenes are generated from a script, not by animating the image.

### Is it free to upload?

Uploading a reference is a membership feature on the current plan settings.

### Will the style match my image?

The face, hair and clothing carry over; the art style is chosen from the SceneMixer project's looks.

### Can I upload a celebrity image?

No. In SceneMixer, celebrity, brand and franchise images are refused.

### What formats work?

JPEG, PNG and other common formats; SceneMixer does not accept SVG.

### Do the characters talk?

Yes, in any of 15 languages.

### How much does the video cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

Updated November 3, 2026


---

# AI Pilot Episode Generator: Make the Pilot That Proves Your Series

URL: https://scenemixer.com/ai-pilot-episode-generator/
Generate a pilot episode with AI: set the type to TV Series, upload the pilot script, draw the cast once and render a voiced pilot of up to 300 seconds.

# AI Pilot Episode Generator

Quick answer: A pilot has one job: show that the series works. In SceneMixer, set the type to TV Series, give it the pilot script (or let AI Create draft one), and it draws the cast and the main places once, storyboards the episode and renders it with spoken dialogue, up to 300 seconds. Because the cast is reusable, the pilot's characters carry straight into episode two, a trailer or a pitch.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What a good pilot does

| Pilot job | How it shows up |
| Introduce the world | One establishing shot, one place drawn once |
| Make us follow someone | A lead with a want, in the first scene |
| Show the engine | The thing that will repeat every week |
| End on a question | A closing beat that makes episode two necessary |

A pilot or pitch trailer, step by step · Hooks and cliffhangers.

## From script to pilot

- Upload the pilot script and set the type to TV Series.
- Approve the lead cast and the two or three places that will recur.
- Read the storyboard; the first episode's storyboard on your first drama is free.
- Render and join. Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second.

## After the pilot

Use the same cast for a one-minute trailer, a mood reel, and episode two. AI movie pitch generator · AI TV series creator · Add episodes.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can AI make a pilot episode?

Yes. Set the type to TV Series, give it the pilot script, and SceneMixer renders the episode with the cast and dialogue.

### How long can a pilot be?

On SceneMixer, TV Series episodes run up to 300 seconds.

### Can I reuse the cast?

Yes. SceneMixer carries the characters and places into later episodes and trailers.

### Is it a finished broadcast?

No. Treat it as previs: tone, cast and pacing.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Is the storyboard free?

The first episode's storyboard on your first SceneMixer drama is free; editing text is free.

### Can I make a trailer too?

Yes, from a second short script with the same cast.

Updated November 3, 2026


---

# AI Movie Maker App: Make a Movie From Your Phone's Browser

URL: https://scenemixer.com/ai-movie-maker-app/
Looking for an AI movie maker app? SceneMixer runs in your phone's browser: same account, 9:16 in Step 1, a one-hand editor. No store download is linked.

# AI Movie Maker App

Quick answer: People who search for an AI movie maker app, an AI film maker app or an AI movie app want to make a movie with the phone in their hand. On November 2, 2026 the way to do that with SceneMixer is the website in your phone’s browser: sign in with Google, Apple or an email code and you see the same projects, credits and renders as on a computer. The site does not link an App Store or Google Play download, and its own install-as-an-app option is switched off, so this page does not send you to one. What works on a phone is the whole job, from a one-line idea to a downloaded video, with vertical 9:16 Portrait one choice away in Step 1.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## What an app does, and how the browser covers it

Most people who want an “app” want five things: to start it from the home screen, to stay signed in, to make vertical video, to use photos from the phone, and to carry on later from somewhere else. Here is where each one stands today.

| You expect from an app | On SceneMixer today | Good to know |
| Install it from a store | Not linked from the site. Open scenemixer.com in your phone’s browser. | Checked November 2, 2026: the home page and the footer carry no store link. If that changes, this page will say so. |
| An icon on the home screen | Your browser’s own Add to Home Screen makes a shortcut that opens the website. | It is a shortcut to the site, not a separate app. The site’s install-as-an-app option is off. |
| Stay signed in | Google, Apple or an email code. There is no password to remember. | Sign up and log in |
| Make vertical video | A new SceneMixer project starts as 16:9 Landscape on phones, tablets and computers alike; pick 9:16 Portrait in Step 1 for a vertical feed. | Vertical or landscape |
| Use photos from the phone | Ad photos (up to four) and, on a paid plan, reference photos come from your camera roll or files app. | Using your own photo |
| Carry on from another device | Same account: projects, credits and renders in progress are shared. | What follows your account |
| Share the result | Share lists the usual platforms and adds the phone’s own share menu; Download saves the original-quality file. | Original vs play version |

## A first movie on the phone, in order

- Sign in. New accounts start with 50 credits.
- Give it a story. Type one line in AI Create and let it draft the episodes, or paste a script from any app. A .docx or .txt from your files app works too: which files work.
- Step 1. Choose Short Drama, TV Series or Film, choose 9:16 Portrait for a phone-first video, pick the three looks, the region and the dialogue language.
- Step 2. Look at the cast and the places. The first 15 setup images on an account are free, so this is the time to fix a face.
- Read the storyboard. The first episode’s storyboard on your first drama is free, and a 5-second preview is free too. Reading before rendering is the habit that saves credits.
- Render, then join. Video is 10 credits a second on the default tier, joining is 1 credit a second. Each button shows its price before you press it. Rendering runs on the server, not in your phone, so you can leave and come back.
- Share. Download the file or use the share menu. Posting to TikTok, YouTube Shorts and Reels.

## The editor on a small screen

The episode editor is three columns on a computer. On a phone it stacks into three views you swipe between: the preview (play, download, share), the timeline (one card per segment, with its price and state) and the storyboard text of the segment you selected. The cast and scene rail sits behind a button. The model picker and the AI-badge switch live in a bottom sheet: tap the model name to open it. Everything the desktop editor does is reachable; it is simply arranged for one hand. Using the site on a phone · Editor layout for vertical and landscape projects.

## What is easier at a desk

Nothing is phone-only or desktop-only, and prices are the same everywhere. A few jobs are just faster with a keyboard and a wide screen:

- Pasting and reading a long script, or a step 1 summary with many episodes.
- Editing several storyboard segments in a row; a line or two is fine on a phone, a rewrite of a scene is not.
- Comparing two versions of a clip side by side.
A common split is to write and review at the desk, and check progress, watch the result and share from the phone.

## Five questions to ask any AI movie app

Whatever you end up using, these decide whether you get a movie or a pile of clips.

| Ask | Why it matters | How SceneMixer answers |
| Does the hero look the same in scene two? | Most disappointment comes from faces that drift between clips. | Each character, place and prop is drawn once and attached to every shot that uses it. |
| Is the price shown before I press the button? | Video is the expensive part. | Every price is shown first. Video is 10 credits a second on the default tier. |
| Do the characters speak? | Silent clips need a separate voice step. | Dialogue is performed in the video itself, in any of 15 languages. |
| Can I fix one shot? | Re-doing a whole episode for one bad shot costs real money. | Edit that shot’s text and regenerate the segment alone. |
| What is on the file I download? | A badge or a watermark can rule a clip out for your use. | Free accounts carry a small AI badge on every video; on a paid plan it is off by default. |

Why shared videos carry the badge.

## Where to go next

Related: AI movie maker · Vertical video generator · AI drama for TikTok and Reels · AI movie making · Free AI movie maker.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Is there a SceneMixer app for iPhone or Android?

Checked on November 2, 2026: the site does not link an App Store or Google Play download. The way to use SceneMixer on a phone is the website in the phone’s browser, with the same account as on a computer.

### Can I make a whole movie on my phone?

Yes. Write or generate the story, pick the look, read the storyboard, render the segments, join them and download or share, all in the phone’s browser. Long edits are easier with a keyboard.

### Can I add SceneMixer to my home screen?

Your browser’s Add to Home Screen makes a shortcut that opens the website. The site’s own install-as-an-app option is switched off at the moment.

### Will it make vertical video on a phone?

Yes. A new SceneMixer project starts as 16:9 Landscape on every device, and you choose 9:16 Portrait in Step 1; the choice is fixed once you continue.

### Is it free to try?

New accounts start with 50 credits, the first 15 setup images are free, the first episode’s storyboard on your first drama is free, and there is a free preview of about 5 seconds.

### Do I need to keep the page open while a video renders?

No. Generation runs on the server, not on your phone, so you can start a render and check the result later from the phone or a computer.

### Can I upload photos from my camera roll?

Yes. Up to four ad photos, and on a paid plan reference photos for characters and places, come from the camera roll or the files app.

### Does the phone cost more credits than a computer?

No. Prices are the same everywhere. Video is 10 credits a second on the default tier plus 1 credit a second to join.

Updated November 2, 2026


---

# AI Music Video Generator: What It Can and Cannot Do With a Song

URL: https://scenemixer.com/ai-music-video-generator/
AI music video and lyrics video generator, honestly: SceneMixer does not take your song or sync to a beat. It makes story-led visuals you can set to music in an editor.

# AI Music Video Generator

Quick answer: SceneMixer is not a tool that takes your song and cuts pictures to it. It does not accept a song as the thing a video is built around, it does not find the beat, and it does not time lyrics to a vocal. What it makes is the picture side: a story told in scenes, with characters who look the same in every shot, in segments of 15 or 30 seconds. You can write those scenes from your lyrics, download the video with the built-in score switched off, and lay your own track under it in any video editor. So it suits a narrative music video, where verses become scenes; it does not suit a lyric video with timed words or a beat-reactive visualiser.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Four things people mean, and what happens to each

| You mean… | Here | Why |
| Upload my song and get a video timed to it | Not possible | The inputs are text, up to four ad photos and one video. A song is not an input, and nothing detects a beat or a verse. |
| Turn my lyrics into a story with pictures | Yes | Write the story the lyrics tell as text. It is read into a cast, places and a numbered storyboard, then rendered. |
| Words on screen, timed to the singing (a lyric video) | Not timed | A short phrase can be written into a shot as an on-screen caption. Word-by-word timing against a vocal is done in an editor, where the song is. |
| An artist performing the song | Not your recording | A character can be shown singing in a shot, but the sound is not your vocal and nothing lines the mouth up to your track. A real artist’s face is not accepted: the IP check refuses celebrity images. |

If the song must drive the picture, you need a tool that reads audio. If the picture must tell a story, read on. What you can feed in.

## What SceneMixer does with sound

Four kinds of sound exist here, and none of them is a song you supply.

- The built-in score. Once an episode’s first shot is generated, the site plans music from the storyboard (which mood each stretch needs, where it turns), picks a piece for each stretch and mixes it under the voices. There is no track picker, no volume control and no request box. A Music switch beside the player turns it off, and a download with it off is the dialogue version without music. How the score works.
- Spoken lines. Each character has a voice, in any of 15 languages, and a shot with no dialogue has no voice at all; what you hear is the ambience written into that shot. Dialogue and sound in one generation.
- A character’s voice sample. Members can upload a short clip as the reference for how a character speaks. It is a speaking-voice reference, not a track to lay under pictures. How character voices work.
- The sound of a video you upload. In a remake, the clip’s speech is transcribed and the music is described, and that description sets the tone of the new episode’s own score. The source track itself is never used. How a remake reads dialogue and music.

## A narrative music video in six steps

- Mark the song. Write down where the verses, choruses and bridge start, in seconds. That list is your structure.
- Write the story the song tells. One place and one beat per segment. Give the lyrics as the idea; do not paste them as spoken lines, because spoken lines are voiced.
- Pick the type. A Short Drama episode runs up to 240 seconds, a TV Series episode up to 300, and a Film has no cap, so most songs fit one episode.
- Choose the segment length. Step 3 has a segment-length switch: 15 seconds as standard, or up to 30 for a long take on the tiers that allow it. A 30-second verse is two short segments or one long one.
- Write silent shots. Leave out dialogue, add an Ambient: line, and read the storyboard before you render. Editing text is free.
- Render, download with Music off, finish in an editor. Put the song under the picture, trim each segment’s start or end to land on the beats, and add timed lyrics there. Trimming a clip.

## From song sections to segments

| Song part | Typical length | Segments of 15 s | What to put in the storyboard |
| Intro | 10–15 s | 1 | The place, empty or just waking. No lines. |
| Verse | 30–45 s | 2–3 | One story beat per segment; keep the same place within a segment. |
| Chorus | 20–30 s | 2 | The image that repeats. Return to the same location each time so the chorus reads as a refrain. |
| Bridge | 15–30 s | 1–2 | The turn: a new place, a new light, a change in who is on screen. |
| Outro | 10–15 s | 1 | The last image. A held look, not a new event. |

The typical lengths are ordinary song structure, not a product rule. SceneMixer's only fixed numbers are the segment sizes and the episode caps above. Shot seconds and timing.

## A lyric turned into shots

Take an invented chorus line: “The porch light stays on for the one who left.” As a story beat it becomes images, not a sung line. Three shots, twelve seconds, no spoken words:

| Beat | Shot |
| The porch | [Shot 1] 4s: Wide shot, a porch at dusk, the light on, an empty chair. Ambient: crickets, a screen door ticking in the wind. |
| The waiting | [Shot 2] 4s: Close-up on @Nora at the window, watching the road. Ambient: a clock. |
| The return of the chorus | [Shot 3] 4s: Wide shot, the same porch at dawn, the light still on. |
The same porch comes back each time the chorus does, which is how a repeated lyric becomes a repeated image. The shot format · A stage-and-studio story.

## Length and cost

Video is charged per finished second: 10 credits a second on the default tier (about $0.06 a second at the lowest credit price), plus 1 credit a second to join. A three-minute video is 180 of those seconds. New accounts start with 50 credits, the first 15 setup images are free, the first storyboard on your first drama is free and a 5-second preview is free, so you can draw the cast and read the whole storyboard before spending on video. Where credits go in one episode.

## Where to go next

Related: AI video with dialogue · AI trailer generator · AI storyboard generator · Vertical video generator · AI video remake.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Can I upload my own song?

No. A song is not an input: SceneMixer takes text, up to four ad photos or one video. To use your own track, download the finished video with the score off and lay the song under it in an editor.

### Does it sync the video to the beat?

No. SceneMixer does not analyse a song for its beat, so cuts are not placed on it. You trim segments to the beat in an editor after download.

### Can it make a lyric video?

Not with words timed to a vocal. Short phrases can be written into a shot as on-screen captions; timed lyrics are added in an editor.

### Can it write the song?

No. SceneMixer plans and mixes a background score for each episode from the storyboard, with no track picker and no request box.

### Can a character sing my song?

A character can be shown singing in a shot, but the sound is not your recording and no lip-sync to your track is done.

### How long can a music video be?

On SceneMixer, a Short Drama episode runs up to 240 seconds, a TV Series episode up to 300 and a Film has no cap. Each segment is 15 or 30 seconds.

### Can I make it vertical for TikTok or Reels?

Yes. Choose 9:16 Portrait in SceneMixer's Step 1. On a phone it is the default.

### How much does a music video cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join. The first 15 setup images, the first storyboard and a 5-second preview are free.

Updated November 3, 2026


---

# AI Script Coverage: A Read-Back and Outline, Not a Verdict

URL: https://scenemixer.com/ai-script-coverage/
AI script coverage and analysis, plainly: SceneMixer reads your script back as a summary, cast and episode outline to check. It gives no score or recommendation.

# AI Script Coverage

Quick answer: Script coverage, in the trade sense, is a reader’s report: a logline, a synopsis, comments on story, character and dialogue, and a verdict such as pass, consider or recommend. SceneMixer does not write that report and gives no grade or recommendation. What it does when you paste or upload a script is read it back: a title, protagonist, era and logline, the characters, places and props it found, and an outline for every episode with its conflict, its emotional high point and its closing hook. Treat that as a cold read. If the machine’s version of your story is not the story you meant, you have learned something before a single picture is drawn.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Coverage, section by section: what you get here

| A coverage report has… | Here? | Where it shows up |
| Logline and premise | Yes | The summary after the read: title, protagonist, era, logline and a one-line premise, all editable. |
| Synopsis | Yes, per episode | An outline for each episode of three to five sentences, with a conflict, an emotional peak and an unresolved ending. |
| Character notes | Yes, as descriptions | Each named character with an appearance specific enough to draw, plus a role and a voice. Not an opinion of the writing. |
| Structure and pacing | You judge it | The episode list shows how many episodes, and each entry’s conflict, peak and hook. The reading does not say whether they work. |
| Dialogue | Carried through | Your lines go into the storyboard with the speaker named and are spoken as written. Nothing is rated. |
| Marketability, comps, genre fit | No | No comparison to other titles and no market view. |
| Grade or pass / consider / recommend | No | There is no score, star rating or verdict anywhere in the product. |

Everything SceneMixer lists points back at your own text: each episode entry quotes the line of the script it came from, character for character, so you can check it. For the cast, places and props as production objects, see AI script breakdown.

## The cold-read test

A human reader’s first value is that they know only your pages. The read gives you a cheap version of that. Before you press the button, write down three lines for yourself: who the protagonist is, what they want, how episode one ends. Then compare.

| What you see | Worth checking | Where to fix it |
| A different protagonist from the one you meant | Does the script give someone else more of the action early on? | Roles in Step 2, or the script itself |
| Ten episodes for a story you saw as three | The type and the split were chosen before the read. | A new project with the type you want |
| A character who matters is missing | A person who speaks once and is never touched or named again may not make the cast list. | Add them in Step 2 |
| An episode ends on an answer, not a question | The closing hook in that entry. | Rewrite the last beat of that episode |
| The last quoted line is halfway through your file | The read stopped short. | Check the length ceiling below, then start a new project and read it again |

A mismatch is a prompt to look, not a finding about your script: the reading is a model, and it can be wrong. Checking the summary, and what re-parse means.

## Reading the episode list for rhythm

The list is the closest thing to a structural overview. Read the entries one under another and ask the questions a story analyst would ask, answering them yourself:

- Does every entry have a conflict? An episode whose outline is a sequence of events with no one wanting something is flat.
- Do the peaks rise? Skim the emotional high points in order. A series that peaks in episode two and never again will feel it.
- Are the hooks specific? “She discovers the truth” is not a hook; “she opens the second envelope” is.
- Who is on the page? Count how many entries each lead appears in. The person you think of as the lead should be most of them.
Outlines cannot be edited on the site, but storyboard text can, for free, so a rewrite costs nothing until you render video. AI episode outline generator.

## Hear it before you decide

The read is a summary. The storyboard is the script as shots: numbered, with seconds, framing, the speaker and the line. The first episode’s storyboard on your first drama is free, and a 5-second preview is free. A quick dialogue check you can run on any script: a spoken line takes about three words a second in English, so a twelve-word line needs a four-second shot. Lines that do not fit their shot are easy to spot in the storyboard. How dialogue is written in a shot · Shot seconds and timing.

## What goes in, what it costs, what it will not do

Formats: paste, or upload .txt, .docx or .md. SceneMixer unpacks a .docx to plain text in your browser. PDF and .fdx are not read directly; export to text first. Screenplay layout with scene headings and character cues is understood: using a screenplay.

Length: a floor of about 330 characters of English, and a ceiling of 300,000 characters in one project. Long scripts are read in two passes so the episode list is not quietly cut short.

Cost: reading is 4 credits per 1,000 characters, English counted lighter. If a read fails, the credits come back. New accounts start with 50 credits.

- It will not give a score, a rating or a recommendation.
- It will not compare your script with other films or judge its market.
- It will not rewrite your plot: scene order, dialogue and beats follow the script.
- If you need a recommendation to hand a producer, you need a human reader.

## Where to go next

Related: AI script breakdown · AI script generator · AI film script generator · Script to movie · AI storyboard generator.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Does SceneMixer do script coverage?

Not in the trade sense. It reads your script back as a summary, a cast and an outline for every episode. It does not write comments or give a score, a rating or a recommendation.

### Can AI tell me whether my script is good?

Not here. SceneMixer's read shows you what a machine understood: protagonist, era, characters, and each episode’s conflict, peak and hook. Whether that is good is your call, or a human reader’s.

### What does the read give me?

A title, protagonist, era, adaptation strategy and logline, counts of characters, scenes, props and episodes, and an outline for each episode with a verbatim quote from the script.

### How is this different from AI script breakdown?

The breakdown page covers the cast, places and props as production objects. This page is about using the same read as a cold read of your story.

### Which files can I use?

Paste text, or upload .txt, .docx or .md. SceneMixer does not read PDF or .fdx directly.

### How long can the script be?

From about 330 characters of English up to 300,000 characters in one project.

### How much does it cost?

On SceneMixer, reading is 4 credits per 1,000 characters, English counted lighter, and a failed read is refunded. The first episode’s storyboard on your first drama is free.

### What happens if the read fails?

SceneMixer's Step 1 shows the reason and the credits are back. Fix what it names, usually a script that is too long, too short or just a link, and press Next again.

Updated November 3, 2026


---

# AI Moodboard Generator: A Look Board From Your Story's Own Art

URL: https://scenemixer.com/ai-moodboard-generator/
AI moodboard and lookbook generator: draw a cast board, location plates, prop sheets and key art in your chosen look. Sheets stay neutral; mood comes from style cards.

# AI Moodboard Generator

Quick answer: A moodboard fixes how a project should look and feel before anyone shoots it. SceneMixer does not make a collage from a prompt; it draws the pieces a look board is made of, from your story: a casting board of character sheets, a location board of empty scene plates, a prop board, and three key-art posters, all in the three looks you choose in Step 1. The sheets are plain on purpose, with a white background, even light and no mood, so the atmosphere on your board comes from the style cards, the genre mood on the poster and the light you write into the storyboard. Download the images and arrange them in any slide or board tool.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## The boards you can build

| Board | Made from | What it shows | Where |
| Look board | Three style cards in Step 1: Visual, Scene and Costume | The medium and finish, the architecture and materials, the clothing. Members can type their own description instead of a preset. | Step 1 |
| Cast board | Character sheets | Face, build and one outfit each, full length on white with a plain side light. | Step 2, Characters |
| Location board | Scene images | An empty place with three layers of depth and a readable light source. No people, no effects. | Step 2, Scenes |
| Prop board | Prop sheets | The object alone on white, lit only to show its material. | Step 2, Items |
| Key art | The cover art tab | Three posters with your title painted in, from your leads’ sheets, with twelve genre moods to choose from. | Step 2, Cover art |
| Motion reference | The free 5-second preview and finished clips | How the light and camera behave in motion. | Episode editor |

Character sheets · Scene images · Props · Posters.

## Why the sheets are neutral, and where the mood lives

A sheet is a reference: every shot that uses a character, place or prop is generated from its sheet, and the model copies the light and mood as faithfully as the face. A character sheet lit like a poster brings that lighting into a sunny scene. So sheets are deliberately plain, and the cinematic look is set elsewhere:

- The three style cards set the medium (Visual), the spatial language and materials (Scene) and the clothing (Costume) for everything that gets drawn.
- The genre mood on the cover art (Romance, Suspense, Epic, Family, Feel-good, Workplace, Horror, Comedy, Teen, Survival, Sci-fi, Action) sets the poster’s colours and atmosphere.
- The storyboard names the light of each shot, its source, direction and time of day.
So for a board that feels like something, use the key art and the preview frames; for a board that says what things look like, use the sheets. Why sheets look plain.

## Compare three looks cheaply

SceneMixer fixes the visual style once you leave Step 1, so a look is tried by starting a short project in it. The free image allowance is counted per account, not per project, so this stays cheap:

- Write or generate a 200-word story with one lead and one main place.
- Start three projects from it, each with a different set of cards in Step 1. Reading is priced by length, so a short script is cheap.
- In each, draw only the lead and the main place: two images per look, six in all, inside the first 15 free setup images.
- Put the three pairs side by side and keep the one that feels like your film.

Visual Style: Shot on Film

Visual Style: Soft Glow

Visual Style: Family 3D Animation

Visual Style: Watercolor Storybook

Scene Style: Neon Nights

Scene Style: Urban Opulence

Scene Style: American Small Town

Costume Style: Streetwear

Costume Style: Soft Everyday
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
Free sheets are per account · Visual styles one by one · Scene and costume styles one by one · Writing a custom style.

## A cast board, from a real sample

This is what a cast board looks like in practice: one reference image per character, on a plain background, in the look the project chose.
Character sheets from the sample The Veil at Dawn: one reference image per character, used in every shot they appear in. The same images are what every shot of the drama is generated from.

## A lookbook for a story, not for garments

A lookbook in this sense is the set of outfits your characters wear. In SceneMixer, each character sheet carries one outfit. A full change of clothes is a state of the same character, with its own image, so a board of five outfits for one lead is five state cards. A change that only needs words, such as rain, dust or a different time of day, is written in the storyboard and needs no new image. Costume changes and scene states · One outfit per character sheet · Editing a sheet: outfit, hair, expression.

It does not model real garments: you cannot photograph a jacket and have it worn in a still. Photos of what you sell become an ad, which is a video, not a set of stills. AI ad video generator.

## Getting the images out

Open Assets in the sidebar: the Characters, Scenes and Items tabs hold every sheet from every project, and each entry has a Download button. SceneMixer does not count generated images against your storage quota; only uploads do. Images after the first 15 are 7 credits each on the default image model. Arrange them in a slide deck or any board tool, add your own references beside them, and share the board. The asset library · Which image tier for sheets · AI movie pitch generator.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Is SceneMixer a moodboard generator?

It draws the pieces a look board is made of, from your story: character sheets, scene images, prop sheets and key-art posters, in the style you choose. It does not make a collage from a prompt.

### Why do the images look so plain?

Sheets are references for every shot, and the model copies their light and mood. SceneMixer keeps them neutral on purpose; the atmosphere is set by the style cards, the poster’s genre mood and the storyboard.

### Can I make a lookbook?

For characters in a story, yes: outfits are states of a character, each with its own image. SceneMixer does not model real garments from photos.

### Can I change the look after I start?

The visual style is fixed once you leave Step 1. To compare looks, start a short project in each; the free image allowance is per account.

### Can I download the images?

Yes. Every sheet in the asset library has a Download button, and generated images do not count against your storage quota.

### How many images are free?

The first 15 setup images on an account, then 7 credits each on the default image model. The first project’s first set of three posters is free.

### Can I use my own reference images?

Members can upload a reference and have it redrawn in the project’s look. SceneMixer refuses celebrity, brand and franchise images.

### Can I write my own style?

On SceneMixer, members can type a custom visual style description of up to 200 characters in Step 1. It replaces the preset for the whole project.

Updated November 3, 2026


---

# Watch AI Short Dramas: Free Sample Episodes in 15 Languages

URL: https://scenemixer.com/watch-ai-short-dramas/
Watch AI short dramas free: sample episodes of 2 to 3½ minutes in English and 13 other languages, each with its cast and outline. Make your own if you like one.

# Watch AI Short Dramas

Quick answer: SceneMixer’s shelf of AI short dramas is its sample library: twenty-three episodes that were each made from a script, ten with English dialogue (one of them vertical) and thirteen with native dialogue in other languages. Each is one episode, between 1:46 and 3:30 long, and plays on its own page with no account. It is a fixed shelf, not a streaming service or a feed with new episodes every day. Start with the ten English episodes below, then try one in your own language. If one makes you want to try, every episode began as text, and the steps are at the bottom.

| SceneMixer at a glance | What you get |
| Video price | From $0.06 a second (10 credits) on the default tier, up to $0.47 a second on the top tier; every price is shown before you press the button |
| Clip length | Up to 30 seconds per generated segment; an episode is composited from its segments at 1 credit a second |
| One episode, worked out | A 60-second episode on the default tier is about 660 credits (video plus compositing), about $4.11 at the lowest credit price; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 |
| Dialogue | Spoken in the video itself, one voice per character, in any of 15 languages |
| Formats | 9:16 vertical or 16:9 landscape; 480P to 1080P, plus a 2K tier |
| Free to start | 50 credits on signup, 15 free setup images, a free storyboard for the first episode and a free 5-second video preview |
| Who uses it | 2,900+ creators have generated with SceneMixer (accounts with at least one generation, counted November 2026) |
| Input | A script, a novel or a one-line idea; .txt, .docx and .md uploads |

## Ten English episodes, with one line each

| Episode | What it is | Length | Watch |
| The Glass Floor | Corporate thriller. A firing on a glass-walled executive floor turns into a public accusation, and the fired executive is offered an impossible counter-move. | 2:22 | Play |
| The Eagle’s Vow | War drama at a Roman frontier camp and the ambush road beyond it. | 2:16 | Play |
| Rocket Pup | 3D animated family adventure. A dog on a wobbling rocket, a vain mayor, a letter mystery and a giant metal cat. | 2:20 | Play |
| Neon Debt | Cyberpunk chase thriller. A courier runs a data case through a rain-slicked night market with a kill-team behind her. | 2:32 | Play |
| Crimson Vow | Rain-soaked urban thriller that turns into a gothic bloodline mystery within minutes. | 2:23 | Play |
| Rust City Knockout | Painterly anime action. One propulsive night chase through a rain-soaked harbour district. The longest on the shelf. | 3:30 | Play |
| Starfall Command | Sci-fi military thriller on an orbital station under siege; a former fleet commander steals the enemy’s command core. | 2:14 | Play |
| The Last Envelope | Wartime espionage thriller in an alternate Blitz-era London, played as noir. | 2:20 | Play |
| The Veil at Dawn | A woman wakes with memories of a past life and walks out on a toxic wedding. The shortest English episode. | 1:46 | Play |
| The Violin on Mercer St | Vertical (9:16) mystery. A violinist returns to the neighbourhood she fled seven years earlier. | 2:28 | Play |

These are the English-language episodes from the sample library. Nine are 16:9 landscape. The Violin on Mercer St is 9:16, the shape of a phone held upright. Every page is labelled as AI-generated and made end to end with SceneMixer.

## The same shelf in thirteen more languages

The native-language episodes were written and performed in the language itself; the site’s own sample pages say there is no dubbing pass. They run between 1:52 and 2:44. Native dialogue vs dubbing.

| Language | Episode | Length |
| Portuguese | A Carta de Lisboa | 1:52 |
| Spanish | El Secreto de Costa | 1:57 |
| Japanese | 浪人の誓い | 1:53 |
| Korean | 윈터 프라미스 | 2:04 |
| Indonesian | Ikrar Jakarta | 2:22 |
| Russian | Зимняя корона | 1:54 |
| French | L’Héritier du Défilé | 2:10 |
| Italian | Il Tavolo dell’Oliva | 2:25 |
| Dutch | De Vuurtoren van de Fjord | 2:15 |
| German | Die Istanbul-Täuschung | 2:05 |
| Turkish | Çantasındaki Sözleşme | 2:24 |
| Arabic | العهد الصحراوي | 2:44 |
| Ukrainian | Останній сигнал | 2:22 |
The fifteenth language, Traditional Chinese, is on the languages page with the others: all 15 languages.

## What each sample page shows

Open any episode and you get more than a video. The page lists the numbers behind it (episodes, characters, scenes, aspect ratio and look), the episode title and length, a written synopsis with the design direction of the look, and the character sheets. In SceneMixer, those sheets are the reference images the video was generated from, so you can compare a face on the sheet with the same face in the episode. If you are thinking of making your own, that comparison is the quickest way to judge how well the cast holds together.

## How to watch an AI drama with a critical eye

An AI drama is easy to enjoy and easy to judge unfairly in both directions. Five things worth noticing in a two-minute episode:

- Do the faces stay the same? Compare a character in the first shot and the last. Drift between shots is a common failure in AI video.
- Do the mouths match the words? In SceneMixer, dialogue is performed in the video, not laid over it, so the lips should move with the line.
- Is it one place? A good segment stays in one room or street and the background does not change behind the person.
- Is there a hook? Short dramas usually end on a question. Notice whether the last shot makes you want the next one.
- What is still hard? Crowds, hands doing fine work and on-screen text remain the weakest areas. When does AI video start to look real?

## From watching to making

Every episode above began as text. To make one of your own:

- Give it a story. Type one line in AI Create, or paste a script.
- Pick the look and the language. SceneMixer's Step 1 sets the type, the format (16:9 or 9:16), the three looks and the dialogue language out of 15.
- Approve the cast. The first 15 setup images are free, so you can see your characters before you pay for video.
- Read the storyboard. The first episode’s storyboard on your first drama is free; a 5-second preview is free as well.
- Render and join. Video is 10 credits a second on the default tier, joining is 1 credit a second. New accounts start with 50 credits.

| What you watched | What you control |
| The look of the episode | Three style cards in Step 1: Visual, Scene and Costume. |
| The faces | The character sheets in Step 2, drawn once and used in every shot. |
| The language | The dialogue language, set per project. |
| Landscape or vertical | 9:16 Portrait or 16:9 Landscape in Step 1. |
| The length | Segments of 15 or 30 seconds, up to 60 to an episode. |
Posting an episode to TikTok, YouTube Shorts and Reels · What you can feed in.

## Where to go next

Related: AI drama generator · AI short drama generator · AI micro drama · Free AI drama generator · AI TV series creator.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## In SceneMixer vs. stitching tools together

Most AI drama workflows are five or six tools in a row. This is where each step happens here, and what it replaces.

| Step | In SceneMixer | Doing it by hand with separate tools |
| Script and episode plan | Paste a script or a novel, or type one line: it is read into a cast, locations, props and an episode outline you can edit | A chat model writes the script; you split it into episodes and carry the text from tool to tool |
| Character faces and places | Every character, location and prop is drawn once as a reference sheet and attached to every shot it appears in | An image model, one sheet at a time; you upload the right reference to every clip yourself |
| Shots | A storyboard per episode with numbered shots and seconds; video is generated segment by segment in 9:16 or 16:9 against the same sheets | A video model, one clip per prompt, describing the same people again each time |
| Voices and dialogue | Lines are performed in the video itself, one voice per character, in any of 15 languages | A text-to-speech tool for each line, then a lip-sync pass on every clip |
| Fixing one shot | Edit that shot's text and regenerate the segment alone; the rest of the episode is untouched | Regenerate the clip, then re-cut the timeline |
| Finishing | Segments are composited into one episode file with its score; members download without the AI badge | A video editor to join the clips, add music and captions, then export |
| Publishing | You upload the finished episode to TikTok, YouTube, Instagram or anywhere else; nothing is posted for you | Upload by hand, the same as here |

## FAQ

### Where can I watch AI short dramas?

SceneMixer’s own shelf is the sample library at /drama/: twenty-three episodes that each play on their own page, with no account needed.

### Are the episodes free to watch?

Yes. Each plays on its page without signing in.

### How long is an episode?

Each sample is one episode. The English ones run 1:46 to 3:30 and the native-language ones 1:52 to 2:44.

### Are there AI dramas in languages other than English?

Yes. Thirteen episodes have native dialogue in Portuguese, Spanish, Japanese, Korean, Indonesian, Russian, French, Italian, Dutch, German, Turkish, Arabic and Ukrainian, and SceneMixer's languages page adds Traditional Chinese.

### Is there a vertical one for my phone?

Yes. The Violin on Mercer St is 9:16. The rest are 16:9.

### Are these real actors?

No. The faces are AI-generated from character sheets, and each page is labelled as an AI-generated short drama.

### Is this an app with new episodes every day?

No. It is a fixed set of sample episodes on the website, not a feed or a streaming service.

### Can I make my own?

Yes. Type one line or paste a script, draw the cast, read the free storyboard and render. New accounts start with 50 credits.

Updated November 3, 2026


---

# Free AI Video Generator: What You Get Free, and Where Paying Starts

URL: https://scenemixer.com/free-ai-video-generator/
Start free with signup credits, free setup images, a free first storyboard and a free short preview. Video is paid by the second; here is exactly where free ends.

# Free AI Video Generator

Quick answer: SceneMixer is free to start and paid to finish. A new account gets 50 credits, the first 15 setup images, the storyboard for episode 1 of its first drama, one free promo cover and one free preview clip of up to 5 seconds. Video itself is charged from the first second, at 10 credits a second on the default tier, and there is no free video allowance. So the free part is enough to see your own script as a cast and a storyboard, and to see one real shot of it. It is not enough to finish an episode. This page lists exactly what is free, what is not, and how to use the free part to find out whether the rest is worth paying for.

## What is free on SceneMixer

| What | How much | The catch |
| Signup credits | 50 credits | Spent on anything that is charged: reading a script, an image, video. |
| Setup images | 15 images, one pool for the whole account | Characters, locations and props. Only on a project whose script has been read and paid for, and not on ads or remakes. Whatever you do not use is still there for the next drama. |
| Storyboard | Episode 1 of your first drama | Decided the moment that project is created. Regenerating it, and every later episode, is charged. |
| Promo cover | One batch of three poster candidates | Your first drama only, on the same conditions as the setup images. |
| Preview clip | One clip of up to 5 seconds | One per account, at the lowest tier, made from your own cast and script. |
| Reading an ad photo | Up to four photos, read once | The read is free and tells you what it thinks the product is. The ad itself is charged. |
| Editing | Storyboard text, switching between takes, choosing which setup image a segment uses | Always free. Only generating video is charged. |

The full list, with the conditions in words, is in What you get for free; the setup images are explained in free sheets are per account.

## What is not free on SceneMixer

- Video. It is charged from the first second, once per attempt. There is no free video allowance and no “try again free”: each generation is a separate paid call to a model service.
- Reading a script. : 4 credits per 1,000 characters, on every project. Writing a script with AI is charged too; each length option shows its own price before you press the button.
- Joining an episode. : 1 credit a second, priced by total length and shown before you confirm.
- Regenerating a storyboard. Charged, and it also discards the finished video in that episode.

## A first session that costs almost nothing

- Write a small brief. On SceneMixer’s AI Create tab the shortest length is Mini, about a thousand characters: one scene, a test. Or paste a page of your own script. Either way, reading a page is a few credits.
- Read the cast. Step 2 shows the characters, locations and props the read found, with the first 15 images free.
- Open episode 1’s storyboard. Free on your first drama. Read every shot; change anything by editing the text, which costs nothing.
- Use the free preview. One shot of your own episode, about 5 seconds, at the lowest tier. How the preview works.
- Decide. If the shot is what you imagined, go on a segment at a time. If it is not, you found out for the price of reading a page.

## If you go on: the cheap way to spend

Most credits in a SceneMixer drama go on video, and most wasted credits go on regenerating video that the text could have fixed. So: fix the storyboard text first; draft on a low tier and finish on a higher one; generate one segment before the whole episode. Spend fewer credits on an AI drama.

For scale: a 60-second episode on the default tier is about 660 credits including the join, about $4.11 at the lowest credit price, before retakes. Where credits go in one episode · AI video cost calculator.

## Why there is no unlimited free video

Every clip is rendered by a model service that charges SceneMixer for each call, whether or not you like the result. A service that gave unlimited free video would be paying for it from somewhere, and the usual somewhere is the quality of the result or the life of the product. So the free part here is the part that costs little to give (reading, images, text editing, one storyboard, one short preview), and video is priced per second with the price on the button. Why unlimited AI video plans disappeared.

## Free accounts and the badge

Videos made on a free account carry a small AI badge in the bottom-right corner. Turning it off is a membership benefit, and a link you share always carries the badge. AI video without a watermark has the table of who gets a clean file.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### Is there really a free AI video generator here?

SceneMixer is free to start: signup credits, free setup images, the first episode's storyboard and one short preview. Video is charged from the first second, so finishing an episode is paid.

### Can I make a whole video for free?

No. Apart from one preview clip of a few seconds, SceneMixer has no free video allowance. The signup credits can be spent on video like any other credits.

### How long is the free preview?

SceneMixer’s free preview runs up to 5 seconds, one per account, at the lowest tier, made from your own cast and script. It is a taste of one shot, not a cut of your project.

### Do free videos have a watermark?

They carry a small AI badge in the bottom-right corner. Turning it off is a membership benefit.

### Can I try my own script before paying?

Yes. SceneMixer charges a few credits per page to read a script, the first drama's episode-1 storyboard is free, and the free preview renders one shot of it.

### What happens to my free setup images if I do not use them?

They belong to the account, not to one project, so whatever you do not use on your first drama is still there for the next one.

### What does it cost after the free part?

SceneMixer charges 10 credits a second for video on the default tier, and joining is 1 credit a second; the price is shown on the button before you press it.

### Is there another way to earn credits?

Sharing your first finished video on SceneMixer can earn credits; the guide has the current terms.

Updated November 3, 2026


---

# AI Video Maker: Make an Episode, an Ad or a Remake Without Editing

URL: https://scenemixer.com/ai-video-maker/
An AI video maker for stories, ads and remakes: bring a script, up to four photos or a clip, and get a finished video with voices, music and a consistent cast.

# AI Video Maker

Quick answer: SceneMixer makes three kinds of video, and each starts from a different thing you already have. A script, a novel or a one-line idea becomes a drama with a consistent cast, voices and music. Up to four photos of a product or a place become an ad of 15, 30 or 60 seconds. A clip you shot becomes a remake with new characters in a new look. You do not cut anything: the cast, the locations, the storyboard, the voices, the music and the final join are made for you, and you approve the cast once and read the storyboard before any video is paid for.

## Three things it makes

| You bring | You get | How long | Start here |
| A script, a novel or a one-line idea (.txt, .docx, .md) | A drama: characters, locations, episodes, a storyboard for each, then video with spoken dialogue | Segments of 15 or 30 seconds joined into episodes; a series as long as the script | Upload a script |
| One to four photos of the same subject and one line about it | An ad with a script, a cast, locations, sound and an end card carrying your brand | 15, 30 or 60 seconds (one, two or four segments) | An ad from one photo |
| A video you shot, within the length and size on the upload box | A remake: the clip read shot by shot, then shot again with new characters in a new look and language | Priced by the clip’s length | AI video remake |

SceneMixer runs all three through the same steps behind the scenes: read the material, make the characters, locations and props, split into episodes, write a storyboard for each, shoot each segment, join. What you can feed in.

## What “without editing” means

SceneMixer does not ask you to place clips on a timeline. What it asks instead is three decisions at the right moments, each before money is spent:

- Approve the cast. After the read you see every character, location and prop with a picture. The first 15 images on an account are free.
- Read the storyboard. It is plain text, shot by shot, with seconds. Editing it costs nothing, and it is what the video is made from.
- Choose a tier and a length. Each segment shows its price on the button. Draft on a low tier, finish on a higher one.
Everything else is automatic: the voices (one per character), the music (scored for you, with a switch to turn it off), the join, and subtitles on accounts that have bought credits at least once. The built-in score · Subtitles.

## What comes out of SceneMixer

| Property | Options |
| Shape | 9:16 vertical or 16:9 landscape, chosen in Step 1 |
| Quality | 480P up to 1080P, and a 2K tier on some models |
| Dialogue | Spoken inside the video itself, one voice per character, in any of 15 languages |
| Segment length | Standard 15 seconds or Long take 30 seconds; the price per second is the same |
| Music | Scored for each episode; no track picker |
| Consistency | One cast, one set of locations and props for the whole series, fixed when the script is read |

Why a whole series stays consistent · 15 or 30 second segments · Which quality tier.

## What one video costs

On SceneMixer’s default tier, video is 10 credits a second, and joining the segments into an episode is 1 credit a second. A 60-second episode comes to about 660 credits including the join, about $4.11 at the lowest credit price, before retakes. Reading the script is 4 credits per 1,000 characters. A shot that misses is regenerated alone, not the episode. Where credits go in one episode.

## A first video in six steps

- Sign up. A new SceneMixer account starts with 50 credits.
- Write the brief. One line on AI Create: who, where, what goes wrong. Or upload a script.
- Step 1. Pick the type, the shape, the three style choices and the dialogue language.
- Step 2. Read the cast. Replace a picture with your own photo if you want a specific face. Use your own photo.
- Storyboard. Read episode 1, edit what you want, use the free 5-second preview.
- Video. Generate one segment, check it, then Generate all and join the episode.

## What it is not for

SceneMixer is built for stories, ads and remakes with people in them. It does not do screen recordings, slide narration or a single still brought to life on its own. It does not keep a real person’s face from a photo, and it does not use famous characters or brand mascots. Why a photo gets flagged. And a generation takes minutes, not seconds, because the model takes minutes. Why a generation takes time.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### What can an AI video maker like this make?

SceneMixer makes dramas and series from a script, ads from one to four photos, and remakes of a clip you shot, each with spoken dialogue, music and a consistent cast.

### Do I need editing skills?

No. With SceneMixer you approve a cast, read a storyboard and choose a tier; the voices, music and join are made for you.

### How long can a video be?

A SceneMixer segment is 15 or 30 seconds and segments are joined into an episode. A series is as long as the script; an ad is 15, 30 or 60 seconds.

### Can it make a video from a photo?

Yes, as an ad: SceneMixer takes up to four photos of the same subject plus one line. Photos can also replace a character, location or prop picture.

### Which languages can the characters speak?

SceneMixer speaks dialogue in the video itself in any of 15 languages, one voice per character.

### Can I post the video to TikTok or YouTube?

Yes. Download the original-quality file or share a link; the badge rules and the platforms' own AI labels apply.

### How much does one video cost?

SceneMixer charges 10 credits a second for video on the default tier plus 1 credit a second to join; a 60-second episode is about 660 credits before retakes.

### Is it free?

SceneMixer is free to start, not free to finish. The free part is described in the free AI video generator page.

Updated November 3, 2026


---

# Create an AI Video: The Five-Step Workflow, With a Real Brief to Copy

URL: https://scenemixer.com/create-ai-video/
Create an AI video in five steps: brief, cast, storyboard, shots, join. A brief to copy, how to write a shot, what each step costs and where to look when one misses.

# How to Create an AI Video

Quick answer: Creating an AI video that holds together takes five steps, in this order: write a brief, check the cast, read the storyboard, shoot the segments, join them. The order matters because the cheap steps come first: a brief and a storyboard are text and cost almost nothing to change, while video is charged every time it is generated. So you do the thinking in text, shoot one segment to check it, and only then shoot the rest. The workflow below is the one SceneMixer uses end to end, with a brief you can copy and the price of each step.

## The five steps in SceneMixer and what each one costs

| Step | You do | It costs |
| 1. Brief | Write one line, or paste a script. One line becomes a script on the AI Create tab | Writing with AI is priced by the length you pick, shown on the button; pasting your own is free until it is read |
| 2. Cast | Read the characters, locations and props the read found; replace a picture if it is wrong | Reading is 4 credits per 1,000 characters; the first 15 images on an account are free |
| 3. Storyboard | Read each shot; edit the text | Episode 1 of your first drama is free; editing is always free |
| 4. Shots | Generate one segment, check it, then the rest | 10 credits a second on the default tier, per attempt |
| 5. Join | Join the segments into one episode, download or share | 1 credit a second |

## Step 1: a brief to copy

The one-line brief matters more than the length you pick. It needs a person, a place, and one thing that goes wrong. Copy this and change the nouns:
A one-minute short: at 11:58 on New Year’s Eve a courier delivers a letter to the wrong flat and stays for the countdown; one stairwell, one balcony, fireworks off screen.
On SceneMixer’s AI Create tab the length options are Mini (about a thousand characters), Brief (2,500), Short (5,000), Medium (10,000) and Long (20,000); each option shows its own price. Mini is a test, Short is the usual first choice. If you paste your own text instead, it needs at least about a paragraph, roughly 330 characters of English. Choosing script length and cost · Which script files work.

## Step 2: check the cast before you spend anything on video

After the read SceneMixer gives you a cast, a set of locations and a prop list, each with a description and a picture. They are fixed once, and every shot of every episode is generated from the same set, which is why faces, outfits and rooms stay the same from shot to shot. Check three things: the age and look of each character, the number of places (fewer is easier), and any prop that matters to the plot. If a face is wrong, fix the description first, then the picture. Why a whole series stays consistent · A character looks the wrong age or ethnicity.

## Step 3: write each shot like a viewer would see it

A SceneMixer shot is a marker with its seconds, then a few sentences saying what the camera sees: the framing, whether it moves, who is in the picture, the one thing that happens, and the line if there is one. One action per shot, and enough seconds for it to happen. Here is the shape:

| Weak | Strong |
| Mara arrives and is upset and talks to Teo. | [Shot 1] 4s: Wide shot, a stairwell at night. @Mara, in a wet jacket, checks a door number and knocks. |
| There is a big emotional moment. | [Shot 3] 3s: Close-up on @Mara. She looks at the letter, then at the paper hat. @Mara: Thirty seconds. Can I watch from your balcony? |
| Fireworks fill the sky. | [Shot 4] 4s: Two-shot on a small balcony. Both count the last three numbers aloud. The sky flashes off screen. |
Editing this text costs nothing, and the seconds are real: they add up to the length of the segment. How to write one shot · Write dialogue in the storyboard.

## Step 4: shoot one segment first

A SceneMixer segment is one generation of 15 or 30 seconds. Open the first one, look at the price on the button, and generate only that. Check the faces, the room and whether the line is spoken. If it is right, use Generate all: it lists how many segments and how many seconds it will submit, and each is charged at its own price and refunded if it fails. Generate all · The free preview.

Draft on a low tier and finish on a higher one. A generation takes minutes, and segments beyond your account’s concurrency limit wait in line and start by themselves.

## When a shot misses: where to look

| What you see | Look here |
| The picture does not match what you wrote | When a shot comes out wrong: edit the text, then regenerate that segment only |
| A character’s face changed between shots | My character’s face changed: a checklist |
| Text appears on screen or the dialogue is wrong | Text on screen and wrong dialogue: fixes |
| Characters swap sides between cuts | Why characters swap sides |
| The shot is fine but you want to spend less | Spend fewer credits |

Every attempt is a separate paid call, and SceneMixer keeps every take, so you can switch back to an earlier one for free.

## Step 5: join, download, post

Once every segment is finished, SceneMixer joins them into one episode for 1 credit a second, shown before you confirm. Download the original-quality file or share a link. Composite and share an episode · Posting to TikTok, Shorts and Reels.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### How do I create an AI video?

In SceneMixer you write a brief, check the cast, read the storyboard, shoot one segment, then the rest, and join them. The steps that are text come first because they cost almost nothing to change.

### What should I write in the brief?

A person, a place and one thing that goes wrong, in one line. Pick a length from Mini to Long; SceneMixer shows the price on the button.

### How long does it take?

Writing and reading take minutes. Each generated segment takes minutes too, because the model takes minutes; segments beyond your limit wait in line.

### How much does it cost to create an AI video?

SceneMixer charges 10 credits a second for video on the default tier, and joining is 1 credit a second; a 60-second episode is about 660 credits before retakes.

### What if a shot comes out wrong?

Edit the shot's text, which is free, and regenerate that one segment. SceneMixer keeps earlier takes, and switching back is free.

### Can I create an AI video from just one sentence?

Yes. One line on SceneMixer’s AI Create tab becomes a script, a cast and a storyboard; you read and edit them before any video is paid for.

### Can the video have speech and music?

Yes. SceneMixer speaks dialogue in the video itself in any of 15 languages, and scores each episode for you.

### Do I need to write shots myself?

No. SceneMixer writes the storyboard for you; reading it and editing a line is enough. Writing a shot yourself is for the shots you want to control.

Updated November 3, 2026


---

# Image to Video AI: Using Photos as Ads, Characters and Locations

URL: https://scenemixer.com/image-to-video-ai/
How photos become video in SceneMixer: up to four photos make an ad, and your own photo can replace a character, location or prop sheet that every shot then follows.

# Image to Video AI

Quick answer: SceneMixer does not have a one-button “animate this photo” tool. Photos go in two ways, and both end in video. Up to four photos of one product or place, plus one line, become an ad of 15, 30 or 60 seconds. Or a photo replaces the picture of a character, location or prop in a drama, and every video generated for that character then looks at your photo instead of the old sheet. If you want a specific person, room or object to move, the second route is the one: put the photo on the card, write what happens in the shot, and generate the segment.

## Two ways a photo becomes video

| Route | What you upload | What moves | Start here |
| An ad | One to four photos of the same subject, up to ten megabytes each and sixteen in total, plus one line about it | A made-up spot built around your product or place, with a script, a cast, sound and an end card | An ad from one photo |
| A character, location or prop | One photo per card in Step 2 of a drama (JPG or PNG) | Whatever the storyboard says happens to that person, in that room, with that object | Use your own photo |

The browser reads WebP, HEIC, PNG or JPEG and sends JPEG or PNG; files your browser cannot open (SVG, camera RAW, HEIC outside Safari) are refused by SceneMixer the moment you pick them. What you can feed in.

## Route one: photos into an ad

SceneMixer reads the photos once, for free, before anything is charged. The read says what it thinks the product is, the category, the product name and any brand name legible on the label, so you can correct it. Your one line wins when the line and the photo disagree. A product is redrawn as a clean studio shot; a shop front or a car or a house is kept as it is. People in the photo are removed first. Which photos work · Length, format and quality.

## Route two: a photo as the face, the room or the object

In SceneMixer’s Step 2, open a card and press Upload & Replace. A dialog asks how to use the image. Optimize into a reference sheet reads the face, hair and clothing from your photo and redraws it in the project’s style on a clean background, priced like an image. Use this image as is puts the photo on the card untouched. Whichever you choose, every video generated for that character from then on looks at that image. Uploading is a membership feature on the current plan settings.

The description under the card does not change, and it still steers how the shot is written, so bring it in line with the picture. Clips you made earlier are untouched. Your own reference images: what changes.

## Prepare the photo like a sheet

| For | Use | Why |
| A person | One person, full figure, plain background, 9:16 | A snapshot’s background, light and pose would otherwise follow the character into every shot, so optimise it |
| A place | The room or street, no people, 16:9 | It becomes the plate every shot in that location is built on |
| A prop | The object alone on a plain background, 16:9 | It is the reference whenever the prop is in frame |
| Artwork you made as a sheet | Use as is | It is already in the right shape |

## Making the photo move: write the shot

The photo sets how it looks; the shot text sets what it does. A short example, once the person is on the card as Ana:
[Shot 1] 5s: Medium shot, a kitchen in morning light. @Ana puts down a cup and turns to the window. @Ana: It stopped raining.
Generate that one segment, look at it, and edit the text if the action is wrong. Editing is free in SceneMixer; generating is charged per attempt. How to write one shot.

## What it will not do

It will not animate a single still on its own, and it will not keep a real person’s likeness: a photo that reads as a real person’s face, a celebrity, or a brand mascot is flagged and refused. Why a photo is flagged. Use photos of your own product, your own place, or a face you made, and keep the rights question in mind: the first time you upload, a window asks you to confirm you have the right to use the material and that anyone whose face or voice is in it consented.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### Can SceneMixer turn one photo into a video?

As an ad: up to four photos of one subject plus a line make a 15, 30 or 60 second spot. SceneMixer has no tool that animates a single still on its own.

### Can I use a photo of myself as a character?

Upload a photo on the character's card in Step 2 and choose to optimise it or use it as is. SceneMixer may flag photos that read as a real person's face; the guide explains why.

### What file types work?

Your browser reads WebP, HEIC, PNG or JPEG and sends JPEG or PNG. SceneMixer refuses SVG and camera RAW when you pick them.

### How many photos can an ad use?

A SceneMixer ad takes up to four photos of the same subject, ten megabytes each and sixteen in total.

### Does the photo cost anything?

SceneMixer reads ad photos for free. Optimising a photo into a reference sheet is priced like an image; using it as is is not an image.

### Will the video look exactly like my photo?

SceneMixer redraws a product as a clean studio shot and keeps a place as it is. A character sheet made from a photo keeps the face, hair and clothing, redrawn in the project's style.

### Do earlier clips change if I replace a photo?

No. SceneMixer leaves clips you generated earlier untouched; new videos for that character look at the new image.

### Is uploading my own photo free?

Uploading your own photo is a SceneMixer membership feature on the current plan settings.

Updated November 3, 2026


---

# AI Video Editor: Fix One Shot Instead of Re-Cutting the Whole Video

URL: https://scenemixer.com/ai-video-editor/
Edit an AI video by changing the text of one shot and regenerating only that segment. What the editor changes, what each fix costs, and when to export to another editor.

# AI Video Editor

Quick answer: SceneMixer edits a video by editing the words behind it, not by cutting footage. Open the episode, click the segment you want to change, edit its text, save, and regenerate that one segment. Editing is free; only the regeneration is charged, priced by that segment’s length and shown on the button, and the old take is kept so you can switch back. It is not a frame-accurate timeline: for colour grading, titles over footage or cutting between exact frames, download the original-quality file and use an editor made for that.

## What the SceneMixer editor changes, and what each change costs

| You want to change | How | It costs |
| What happens in a shot, or who says what | Edit the shot’s text and save | Nothing |
| Which take of a segment is current | Switch to an earlier take | Nothing, and instant |
| Which setup image a segment uses | Pick another image for that segment | Nothing |
| The picture of one segment | Regenerate that segment | Priced by its length, shown on the button |
| A line of dialogue | Redub it rather than regenerating the picture | Cheaper than regeneration |
| A few extra seconds at the end | Trim | Cheaper than regeneration |
| Who plays a character, or a voice | Recast a character or replace a voice in a finished clip (members) | See the guide |
| Words on screen | Subtitles switch next to Music; on accounts that have bought credits at least once | Nothing |
| The music | On or off with the Music switch | Nothing |

Spend fewer credits · Swap a character in a clip · Change a voice · Subtitles.

## The editor in one picture

The SceneMixer editor has three panes. On the left, the asset library for this episode: the characters and locations it draws on. In the middle, the storyboard for the selected segment: shot-by-shot text with durations. On the right, the video for that segment, or the button to generate it. Along the bottom, the timeline of every segment in the episode. The storyboard text is not a log. It is what gets turned into the video prompt, so editing it is how you change the picture. Fixing one shot without redoing everything.

## A repair, step by step

- Find the shot. Click the segment on the timeline.
- Edit the text. Change the action or the line. The seconds are real: they add up to the length of the segment, so cutting a line without changing the seconds leaves dead air.
- Save. Nothing is charged and no model is called; only the shots you touched are re-read.
- Regenerate that segment. The price is on the button, and SceneMixer touches no other segment.
- Compare. Every take is kept. If the new one is worse, switch back for free.
Changing the script does not change a video you have already made; the shot you have came from the old text until you regenerate it. Switch back to an earlier version.

## The one button that throws work away

In SceneMixer, Regenerate the storyboard rebuilds the episode from scratch, and every video you already paid for in that episode is discarded with the old segments, without a refund. The editor makes you confirm. Regenerating one segment is safe; regenerating the storyboard is not. What you lose · When a shot comes out wrong.

## When to export and use another editor

If you need frame-accurate cuts, colour grading, titles over footage or a mix with real footage, a finished SceneMixer episode is a normal video file. Join the segments, then download the original-quality file rather than the play version, and open it in an editor made for that. Original versus play version · Composite and share an episode. Joining is 1 credit a second.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### Can I edit an AI video after it is made?

Yes. In SceneMixer you edit the text of the shot you want to change and regenerate that one segment. Editing is free; the regeneration is charged by that segment's length.

### Does editing one shot cost me the whole episode?

No. SceneMixer charges only the segment you regenerate. The rest of the episode is untouched.

### Can I go back to the earlier version?

Yes. SceneMixer keeps every take, and switching between takes is free and instant.

### Is it a timeline editor?

SceneMixer’s bottom timeline lists every segment in the episode, but changes are made by editing each segment's text. For frame-accurate cuts, export the file and use another editor.

### Can I change a line of dialogue without redoing the picture?

Yes, with SceneMixer’s Redub, which is cheaper than regenerating the segment.

### Can I add subtitles?

Yes. Turn on Subtitles next to the Music switch; SceneMixer charges no credits for them and offers them on accounts that have bought credits at least once.

### What is the button I should avoid?

Regenerate the storyboard. SceneMixer rebuilds the episode and discards every finished video in it without a refund.

### Can I download the edited video?

Yes. Join the segments in SceneMixer and download the original-quality file or share a link.

Updated November 3, 2026


---

# AI Video Generator: What Goes In, What Comes Out, What It Costs

URL: https://scenemixer.com/ai-video-generator/
What an AI video generator takes in (script, idea, photos or a clip), what happens in between, what comes out, and what one real minute of finished video costs.

# AI Video Generator

Quick answer: An AI video generator takes material in, has a model turn it into pictures and sound, and hands a file back. SceneMixer takes four kinds of material: a script, a one-line idea, up to four photos, or a clip you are allowed to use. It hands back video made of segments of up to 15 or 30 seconds, joined into episodes, in 9:16 or 16:9, with spoken dialogue and music. Between the two ends it does what most generators skip: it reads your material into a fixed cast, a set of locations and props, and a shot-by-shot storyboard that you can read and edit before any video is paid for. This page follows the path from input to file, with the limits at each end and a worked cost for one minute of finished video.

## What goes in

There are four ways in, one per tab on the home page. They differ in what you hand over and in what the first step costs; after that they share one path.

| Material | How it goes in | Limits | First step |
| A script or story | Upload Script tab: paste it, or upload TXT, DOCX or Markdown | At least about a paragraph (roughly 330 characters of English). A hard ceiling that refuses rather than silently cutting. Links and PDFs are not accepted: paste the text | The read: 4 credits per 1,000 characters, where 3,333 characters of English count as 1,000 |
| One line of idea | AI Create tab; Inspire Me fills the box with an example to edit | Five lengths from Mini (about 1,000 characters) to Long (20,000 characters) | Writing, priced by the length you pick and shown on the button; then the read |
| Photos | Ad Video tab: one to four photos of the same subject, plus an optional line | 10 MB each and 16 MB in all; your browser converts them to JPEG or PNG | Reading the photos is free; the ad then has one price shown up front |
| A clip | Remake a video tab: MP4 or MOV | 300 MB, and up to 5 minutes. You confirm you hold the rights the first time you upload | The read, priced by the clip’s minutes |

SceneMixer reads the clip, it does not reshoot it: the video model never receives your upload. It is cut into shots and described, and a new video is generated from that description with your cast. What you can feed in has every format and limit; Four ways to start a project helps you pick a tab.

## What happens in between

Whichever tab you used, SceneMixer takes the project through the same five stages. Each one produces something you can read before the next one is paid for.

- The read. The text is read for characters, scenes and props, split into episodes, and given a visual style. It takes minutes and you can leave the page. What comes back is a summary (title, protagonist, era, how the story was adapted), an episode list, and a description for every character, location and prop.
- The cast. Each description becomes a picture: characters at 9:16, locations and props at 16:9. The first 15 images on an account are free. These pictures are the reference every later shot is made from.
- The storyboard. Each episode becomes segments, and each segment becomes three to five numbered shots with their seconds, the people and props named, and the lines with a voice. It takes a few minutes. Editing it costs nothing.
- The segments. A segment is one generation. You press the button with the price on it, the credits are held, and the clip arrives in typically ten to twenty minutes.
- The join. The segments are put together into one file per episode, at 1 credit a second.
The order puts the cheap, readable stages first on purpose: a wrong name or a missing second is a free edit in stage 3 and a paid regeneration in stage 4. What is inside a project shows how the pieces nest.

## What comes out of SceneMixer

| Property | What you get |
| Shape | 9:16 portrait or 16:9 landscape, chosen once in step 1 (a new project starts at 16:9). It cannot be changed after you continue |
| Segment length | Up to 15 seconds on every tier, up to 30 seconds on the long-take tiers. A ceiling, not a target: the story decides where a segment ends |
| Resolution | Set by the tier: from 480P to 1080P, and a 2K tier on some models. The default is a cheap tier on purpose |
| Sound | Dialogue is spoken inside the video, one voice per character, in any of 15 languages; each new episode is scored |
| Files | The site plays a lighter copy; Download gives the full-quality file. Each episode is its own file |
| Mark | A small AI badge in the bottom-right corner and an AI-generated tag in the file’s metadata. Members can turn them off for their own downloads; a shared link always carries the badge |

The badge rules are in AI video without a watermark. For the numbers behind each row, see What comes out.

## A real minute, costed

Take a one-minute episode on SceneMixer’s default tier. Video is 60 seconds at 10 credits a second, and joining is 60 seconds at 1 credit a second. Together that is 660 credits, about $4.11 at the lowest credit price, before the read, and before any retake. The read of a short script is a few credits more, and the storyboard of your first episode is free.

The bigger a project, the more of the bill is video. The setup, the storyboard and all the editing together are a small share of a series; the lever that moves the total is how many seconds you render and at which tier. The AI video cost calculator itemises an eight-episode series.

For a sense of scale, the sample projects on this site are single episodes made by this same path, and they run from about a minute and three quarters to three and a half minutes, with casts of four to eleven characters. The Last Envelope has six characters and two sets and runs 140 seconds in 16:9; The Violin on Mercer St is the vertical one, at 148 seconds in 9:16.

## Where the time goes

Almost all of it is the model. The stages above take minutes, not seconds, and the video stage is the longest. A reading usually takes a few minutes; a setup image about a minute and a half; a segment typically ten to twenty minutes. An internal SceneMixer measurement in August 2026 averaged around nine and a half minutes for a segment, so treat the card’s own estimate as the live number.

How many generations run at once is a property of your plan: 3 on a free account, 5 on the first paid plan, more above that. Start more and the extra ones wait, credits held, and start by themselves. You can close the tab; the work continues on the server. Why a generation takes time.

## What it will not take in or give back

- A link instead of text. Pages are not fetched. Paste the text.
- A PDF. Refused when you pick it. Export to text or Word first.
- One button that animates a single still. Photos become an ad, or the face, room or object that a drama’s shots follow. Image to video AI has the two routes.
- Readable words drawn into the picture. Video models do not render text reliably, so what must be read goes on an end card or is spoken.
- A recognisable public figure or a franchise character. The check refuses it at generation.
- Free video without limit. SceneMixer charges video from the first second; the free part is the read, the first images, one storyboard and one short preview. What is free.

## What decides how it looks

Four things in SceneMixer, none of them a magic word. The tier sets resolution and the longest segment. The visual style you pick in step 1 sets the look, and it cannot be changed once you continue. The setup images set who everyone is. And the storyboard text sets what happens, which is why most odd shots are fixed by editing a sentence, not by paying for a higher tier. No magic words and Describe the cause, not the result cover the writing.

Start on the cheap tier. Both takes stay in the segment’s history, so shooting the keepers again on a higher tier costs one more generation and nothing else. Draft at the lowest tier, finish at the highest.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### What does an AI video generator actually do?

It takes your material, has a model generate pictures and sound from it, and returns a video file. SceneMixer adds a reading step first, so you see the cast and the storyboard before any video is paid for.

### What can I put into the generator?

SceneMixer takes a script or story as text, a one-line idea, one to four photos of a product or place, or a clip you have the rights to. Links and PDFs are not accepted.

### How long is the video it makes?

Each SceneMixer generation is a segment of up to 15 seconds, or 30 on the long-take tiers. Segments are joined into episodes, and a series is as long as the script.

### How long does it take?

On SceneMixer, a reading and a storyboard take minutes; a segment typically takes ten to twenty minutes. You can close the page while it runs.

### How much does one minute of video cost?

On SceneMixer’s default tier a 60-second episode is about 660 credits including joining, about $4.11 at the lowest credit price, before the read and before retakes.

### Can it make vertical video?

Yes. Choose 9:16 in SceneMixer’s step 1 for a vertical feed, or leave the default 16:9. The shape is fixed once you continue.

### Does the video have sound?

Yes. SceneMixer speaks dialogue inside the video, one voice per character, in any of 15 languages, and each new episode is scored.

### Is there a free AI video generator here?

SceneMixer’s start is free: signup credits, the first 15 setup images, one storyboard and one preview clip of up to 5 seconds. Video itself is charged by the second.

Updated November 2, 2026


---

# AI Story Generator: From One Sentence to a Script You Can Shoot

URL: https://scenemixer.com/ai-story-generator/
An AI story generator that goes past the text: one sentence becomes a script, a cast, an episode list and a shot-by-shot storyboard to read and edit before any video.

# AI Story Generator

Quick answer: Most story generators hand you a block of prose and stop. In SceneMixer the story is the first of four documents: a script written from your sentence (or from the draft or novel you bring), a summary with a cast of characters, locations and props, an outline for each episode, and a storyboard of numbered shots with seconds. Each one can be read and edited before the next is made, and everything up to the video costs little or nothing to change. This page shows what each stage produces, with a real sample, and which edits are free.

## The four documents a story turns into

| Stage | You give | You get | It costs |
| 1. Script | One or two sentences on AI Create and a length from Mini to Long, or your own draft | A full script: episodes, scenes and dialogue, in the language your interface is in. Every draft stays in Writing History | Priced by the length you pick, shown on each option; a writing that fails is not charged |
| 2. Read | The script, plus the Step 1 choices: type, three styles, shape, region, dialogue language | A summary (title, protagonist, era, adaptation strategy, counts, an outline per episode) and a written description of every character, location and prop | 4 credits per 1,000 characters, English counted lighter |
| 3. Storyboard | One episode | Segments made of numbered shots, each with its seconds, framing, who is in frame and the lines | 1 credit per 100 characters of that episode’s share of the script, English counted lighter; episode 1 of your first drama is free |
| 4. Video | One segment | The shot itself, with the lines spoken | 10 credits a second on the default tier, shown on the button |

The first three are text, which is the point of the order. A story that has gone wrong shows up as a wrong summary or a wrong shot, and fixing either is typing, not rendering. SceneMixer charges writing, reading and storyboards by the length of the text; video is charged by the second and is where most of the credits go. After the script is written · Uploading a script and parsing it.

## A real example: what the read made of one story

The sample Starfall Command was made end to end from a script in SceneMixer, so its summary is real output, not a mock-up. Here is what the read produced for it:

| Level | What the read produced |
| Setting | Realistic HD, 16:9. A sci-fi military thriller on an orbital station under siege |
| Cast | 7 characters. Two leads, Kira Vane and Captain Rook; then Admiral Vale, Lieutenant Mara, a deck chief, an Imperial soldier and a rebel crewman |
| Places | 2 sets. The summary describes a burning docking bay and a metal catwalk by the command core |
| Episodes | 1 episode, The Core and the Helmet, 134 seconds |
| Story decisions | Opens in the middle of the action with no exposition; backstory arrives through confrontation; the central twist lands in the final beat |
| Visual plan | Cold grey with one hot accent colour, handheld in the chase, steadier and more symmetrical when the admiral arrives |
The last two rows are the part a generator that stops at prose does not give you. The summary is a plan, not a recap: it decides where the twist lands, how the camera behaves in each part and which colour carries meaning, and the storyboard is then written under that plan. That is why Step 1 is the cheapest place to catch a misunderstanding, before a single shot exists. Checking the summary.

## Where the story comes from

Three starting points feed the same four documents. A premise goes into SceneMixer’s AI Create: one or two sentences carrying the genre, the lead and what they want, the conflict, the ending you want and the tone. Vague briefs get vague scripts; this one gets a lighthouse keeper instead of a lighthouse:
A quiet mystery: a lighthouse keeper on his last night finds a letter addressed to whoever comes last; the harbour master wants it burned; he must decide whether to read it; wintry, unhurried, ends on the light going out by itself.
The length options are Mini (about a thousand characters), Brief (2,500), Short (5,000), Medium (10,000) and Long (20,000), and a longer script means more episodes, not longer ones. Writing the one-line brief · Choosing the length.

A draft is pasted or uploaded as .txt, .docx or .md, and your dialogue is kept when it is written as lines. A novel up to 300,000 English characters is read in two parts: first the cast, locations and props, then the episode list. Long scripts, read in two passes.

## What you can change afterwards, and what needs a new read

| Something is wrong with | Fix it here | It costs |
| A name, a role, a description, a voice | Step 2, on the card. Every shot written afterwards follows it | Nothing (a regenerated picture is priced like an image) |
| A character or place that is missing, or one the story does not need | Step 2: add or delete it | Nothing |
| One episode’s storyboard | Step 3: re-parse that episode, or edit its text | Editing is free; re-parsing is the storyboard fee and discards that episode’s videos |
| The way episodes were cut, the region, the type | A new project with the same script, settings chosen before the read | The read fee for the script’s length; your free images are counted per account |

Episode outlines cannot be edited in SceneMixer; you adjust a story at the storyboard, where every shot is plain text. What you can change in a storyboard.

## Taking the story further

A story that ends where the script ends can keep going in the same SceneMixer project. At the bottom of Step 3, Add an episode takes the next part pasted in, or lets AI write it from where the story stands, with a sentence of direction from you and a draft you read and edit before it is added. Only the new pages are read, and the earlier cast is reused. A second season that should stand alone is a new project, where Pick existing brings each returning character across. Making the next episode.

## What it does not do

SceneMixer writes stories to be filmed, so the output is a script of scenes and dialogue, not a prose novel or a formatted manuscript. Famous characters and real people’s likenesses are caught later by the IP check, which refuses a picture or a clip that resembles a real person, a brand mark or a known character, so invent your own cast. Why something is refused. And it is not instant: reading and generating take minutes. The free part covers the first 15 setup images, the first episode’s storyboard and one short preview of up to 5 seconds, enough to judge a story before video is paid. What you get for free · AI script generator · AI story video generator.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### What is an AI story generator?

A tool that writes a story from a short idea. In SceneMixer the story does not stop as text: it is turned into a script, a cast, an episode list and a storyboard you can edit, then into video.

### Does it write the whole story or only the idea?

SceneMixer writes a whole script: episodes, scenes and dialogue, at a length you pick from about a thousand to twenty thousand characters. Each draft is kept in Writing History.

### Can I bring my own story instead?

Yes. Paste it into SceneMixer or upload a .txt, .docx or .md file. A paragraph is the minimum, and a novel up to 300,000 English characters is read in two parts.

### How do I get a better story out of one sentence?

Put five things in it: the genre, the lead and what they want, the conflict, the ending you want and the tone. If the first draft is off, write again with a sharper brief.

### Can I change the story after it is written?

Not inside SceneMixer’s writing window. Write again with a better brief, or copy the text, edit it and paste it into Upload Script. After the read, the summary, the cast and every storyboard are editable for free.

### Which language is the story written in?

SceneMixer writes in the language your interface is in when you press Start Writing. The language the cast speaks in the video is a separate choice in Step 1.

### Is it free?

SceneMixer prices writing by the length you pick, and reading is 4 credits per 1,000 characters. The first drama gets 15 free setup images and a free episode 1 storyboard, and each account gets one free preview of up to 5 seconds.

### Does it make the video too?

Yes, segment by segment, with the lines spoken. Video is the biggest cost on SceneMixer, at 10 credits a second on the default tier, so the story is checked in text first.

Updated November 2, 2026


---

# 15-Second or 30-Second Segments: What Changes | SceneMixer

URL: https://scenemixer.com/guides/15-or-30-second-segments/
Segment length cuts an episode into 15-second segments or 30-second long takes. What each changes in the storyboard, the price, and which tiers can render them.

# 15-Second or 30-Second Segments

Quick answer: Under the episode list in step 3, Segment length offers Standard 15s and Long take 30s. It sets how long each segment of the storyboard may run — and a segment is one video generation. Fifteen-second segments mean more, shorter generations and a finer grain for editing; thirty-second long takes hold a longer stretch of action in one generation but can only be rendered by the tiers that support 30 seconds. The choice applies to episodes you parse from then on; the price per second is the same either way.

Open a finished episode and read the lengths on the timeline cards.

## Where it is

Step 3, below the episode cards: Segment length — Standard 15s / Long take 30s. The hint beside it says the limit: 30s segments can only be rendered by models supporting 30s. Episodes you have already parsed keep the length they were cut to; the setting applies to the next Parse Storyboard or Re-parse. An episode cut into long takes carries a Long take 30s badge on its card.

## What a segment is

One generation: a few shots that are made together and land as one clip. The shots inside it add up to its length. An episode is a row of segments on the timeline; the composite stitches them. Shot seconds.

## Standard 15s

Segments of up to fifteen seconds: three to five shots each, a beat of a scene. More segments per episode, each cheaper on its own, each independently regenerable. If one comes out wrong you redo fifteen seconds, not thirty. Every video tier can render it. This is the default.

## Long take 30s

Segments of up to thirty seconds: a longer stretch of a scene in one generation, with the camera cutting inside it. Good for a conversation that should not be interrupted by a generation boundary, or an action beat that needs the room. Fewer segments per episode. On SceneMixer, only the tiers that support 30-second clips can render them — the Wan 3.0 and Seedance 2.5 families — and a 30-second segment on a 15-second tier is refused as over cap until you switch tier. Tiers.

## Price

Video is priced per second of segment, so an episode costs the same whether it is cut into fifteen- or thirty-second pieces. What changes is the size of each regeneration: redoing a long take costs twice a standard segment.

## Changing your mind

Switch the setting and Re-parse the episode: it is cut again at the new length, and its existing segments and videos are discarded. Do this before you shoot, not after. What re-parsing loses. Within an episode you can also split one segment by hand. Inserting a segment.

| | Standard 15s | Long take 30s |
| Segment length | Up to 15 seconds | Up to 30 seconds |
| Segments per episode | More | Fewer |
| Cost of redoing one | Smaller | Larger |
| Tiers that can render it | All | Wan 3.0 and Seedance 2.5 families |
| Best for | Fine control, drafts | Uninterrupted conversations, longer beats |

## FAQ

### Which should I use?

Standard 15s unless you have a reason: it is the default, every tier renders it, and mistakes are cheaper to redo.

### Does 30s cost more?

Not per second. An episode costs the same; each regeneration is larger.

### My 30s segment says over cap.

The tier you selected in SceneMixer renders up to 15 seconds. Pick a tier from the Wan 3.0 or Seedance 2.5 families.

### Can I mix lengths in one episode?

An episode is cut at one length, but you can split a segment by hand afterwards.

### Does changing the setting change episodes I already parsed?

No. They keep their length until you Re-parse them. Re-parsing replaces that episode's storyboard and its videos are discarded.

### I picked Long take 30s. Why is a segment only a few seconds?

Thirty seconds is the most a segment can hold, not a length every segment is filled to. The story decides where segments end; raise the seconds of a segment's shots in the editor to lengthen it. Set 30 seconds but got short segments.

Updated November 3, 2026


---

# Which AI Video Tier to Pick: Resolution, Price and Length | SceneMixer

URL: https://scenemixer.com/guides/480p-720p-or-1080p-which-tier/
The model picker sits beside the generate button. What each tier changes, what it costs per second, which is the default, and when it is worth paying for more.

# 480P, 720P or 1080P: Which Tier to Pick

Quick answer: Every video segment is generated on one tier, chosen in the picker beside the generate button. The default is Wan 3.0 · 480P at 10 credits a second — about $0.06 a second at the lowest credit price. Higher tiers give more resolution or a different look, cost more per second, and some allow longer segments than others. Draft on the default, regenerate the shots you will deliver on a higher tier; every version is kept, so nothing is thrown away.

Open a finished episode. The picker is above the preview pane on the right; the price on the generate button changes as you switch tiers.

## Where the picker is

In the episode editor, above the video preview on the right, next to the generate button. Click Select Model, pick a tier, and the price on the generate button updates. The choice applies to the segments you generate from then on; segments you already generated keep the tier they were made on, and you can regenerate any of them on a different tier later.

## What a tier is

A tier is a video engine at a resolution. The picker groups them by family: Wan 3.0 at 480P, 720P and 1080P; Seedance 2.5 (Pro, Lite and 480P); Seedance 2.0 (standard, Fast and 480P); MiniMax H3 (standard and Fast). Each row in SceneMixer shows its price per second. The one marked recommended is the tier we would deliver on; the default is the cheapest of the family.

Two things differ between tiers besides price. Resolution: 480P is fine on a phone and for drafts, 720P is what most people deliver, 1080P is for a big screen. And the maximum length of one segment: the Wan 3.0 and Seedance 2.5 tiers allow longer segments than the Seedance 2.0 and MiniMax tiers. If a segment’s shots add up past the cap of the tier you picked, the segment card says over cap and generation is refused until you shorten it or switch tier. Shot seconds.

## What it costs

Video is priced per second of segment, and the per-second price is the tier’s. SceneMixer's default tier is 10 credits a second; the dearest tier works out to about $0.47 a second at the lowest credit price. The price is on the button before you press it, and it is charged only when the segment is delivered. Where credits go in one episode.

## A way to work

Write and draft on the default tier: generate, look, fix the storyboard, regenerate. When a segment is right, regenerate it once more on the tier you will deliver on. Because every version is kept, the draft and the final sit side by side in the segment’s version history and you can put either in the episode. Versions.

Do not upgrade a segment whose storyboard is still wrong. Resolution does not fix a stranger’s face or a missing prop; the text does.

## The badge and the tier

The tier does not change whether the finished video carries the AI badge; that is a separate switch next to the picker. The AI badge switch.

| Tier family | Resolutions in the picker | Good for |
| Wan 3.0 | 480P (default) · 720P (recommended) · 1080P | Drafts, delivery, big screens; longer segments |
| Seedance 2.5 | Pro · Lite · 480P | A different look; longer segments |
| Seedance 2.0 | Standard · Fast · 480P | Shorter segments |
| MiniMax H3 | Standard · Fast | Shorter segments |

## FAQ

### Which tier should I start with?

The default. Draft there, then regenerate the segments you will deliver on the recommended tier.

### Does a higher tier fix a bad shot?

No. It changes resolution and look, not what the shot shows. Fix the storyboard, then choose the tier.

### Can I mix tiers in one episode?

Yes. Each segment is generated on whatever tier was selected at the time, and the episode composites them together.

### Why is generation refused with an 'over cap' message?

The segment is longer than the selected tier allows. Shorten it, split it, or pick a tier with a higher cap.

### Does the free preview use a tier?

The free preview on SceneMixer is a short opening shot generated at no cost; you do not pick a tier for it.

Updated November 3, 2026


---

# “Shot Is Somewhere Else”: One Place and One Timeline per Segment

URL: https://scenemixer.com/guides/a-segment-stays-in-one-place-and-one-timeline/
Why a segment is blocked when a shot is set in another place or time, what the note says, and three fixes: rewrite it, give it its own segment, or show it on a screen.

# “Shot 3 is somewhere else”: a segment stays in one place and one timeline

Quick answer: A segment is one continuous stretch of one place. If the system reads a shot in it as happening somewhere else or at another time — a cut to the other end of a phone call, a narration over a different place, a flashback dropped in — a note appears under the segment: Shot 3 is somewhere else, and pressing generate is refused with A segment stays in one place and one timeline. Rewrite it to happen here and now, or move it to its own segment. No credits were charged. Fix it one of three ways: rewrite the shot so it happens in this place and moment, move it into a segment of its own with its own location, or show the other place as a screen inside this one.

Need the new segment? Insert a segment · How a segment is cut into pieces: why a segment was split in two.

## Why a segment can only be in one place

Each segment is rendered as one clip, from one set of scene references. That is what keeps the room, the people and the light the same from the first shot to the last. A shot set in another room would be drawn against the wrong room, with the wrong people in it, so the site stops it before you pay rather than after.

The same is true of time. A segment is one continuous timeline: no three days later in the middle, no jump back to last year. Both rules are the storyboard side of spatial consistency and the reason the header has a single Location. The segment header.

## What you see

- A note under the storyboard card, before you press anything: Shot 3 is somewhere else (several shots are listed together, for example Shots 2, 4), then A segment stays in one place and one timeline, so this shot can’t be generated. Rewrite it to happen here and now, or move it to its own segment. It sits with the other notes below the text. What each note means.
- A refusal when you press generate anyway: Shot 3 of this segment shows another place or time — a segment stays in one place and one timeline. Rewrite it to happen here and now, or move it to its own segment. No credits were charged.

Nothing is lost and nothing is charged. Saving the text is never blocked; only making the video is. The note only appears on segments written in the current layout, not on older segments or in ad and remake projects.

## What counts as “somewhere else”

- Cross-cutting. The detective in the car, then a shot of the suspect in his kitchen, then back to the car.
- The other end of a call. A close-up of the person who is speaking on the phone, in her own room.
- Narration over another place. A voice-over about the harbour while the picture is the harbour and the segment is in the office.
- A flashback or a time jump. Ten years earlier, the same house, a different light and different faces.
- A different place reached by walking. She leaves the office and arrives at the street: that is a move, so the street needs its own segment.

What does not count: the room staying the same while the camera moves, people leaving or entering the room, a window onto the street, and a car with its passengers (a moving car is one place; see scenes inside a car, bus, train or plane).

## Fix 1: rewrite the shot so it happens here and now

Usually the cheapest. Keep the beat, change where the camera is. The cross-cut to the suspect becomes what the detective sees on his phone; the flashback becomes an object in the room.

| Before (blocked) | After (stays in the segment) |
| [Shot 3] 3s: Cut to the suspect in his kitchen, pouring coffee, unaware. | [Shot 3] 3s: Close-up on the phone in @Mara’s hand: a photo of the suspect pouring coffee, timestamped this morning. |
| [Shot 4] 3s: Ten years earlier, the same dining table, @Mara as a girl laughing. | [Shot 4] 3s: @Mara turns a framed photo face-down on the table; in the photo, a girl laughing at this same table. |
| [Shot 2] 4s: The caller, @Leo, at the airport, shouting over the announcements. | [Shot 2] 4s: Close-up on @Mara, the phone to her ear, flinching at the noise from the speaker. @Leo (off-screen, shouting over announcements): I land at nine! (voice: Leo) |

Write the dialogue as an off-screen voice, not as a face in the picture. Off-screen voices.

## Fix 2: give the shot a segment of its own

When the other place deserves screen time, make it a segment. Press Insert on the segment before it, write the new header with its own Location and the shots that happen there, and cut those shots out of the old segment. Each segment is generated and priced on its own, and the composite joins them as cuts. Insert or delete a segment.

- Copy the blocked shot or shots to a safe place, and delete them from the segment.
- Select the segment, press Insert, and open the blank card.
- Either write the shots yourself, or use AI edit on the blank card and say what happens and where.
- Set the new segment’s location with the @ menu (Scenes tab). If the place has no scene card yet, add one in step 2 first.
- If the story returns to the first place, insert a third segment for it.

## Fix 3: show the other place as a screen

If the other place is only information — a camera feed, a video call, a TV news report, a photo — keep the camera in this place and put the other place on a screen in the frame, as in the table above. The segment stays in one place; the audience still sees the other end.

## If the note says a shot is somewhere else and you think it is not

The note comes from how the system read the text. If a shot is flagged and you believe it stays in the room, look for a word that names another place or time — outside the hospital, later that night, back at the villa — and take it out or tie it to the room: through the window, the hospital car park. Press Save and look at the note again; it is worked out from the saved text. If you are unsure, press Generate: a shot that is still somewhere else is refused again, and nothing is charged. If it stays, split the shot out as in fix 2 rather than fighting it.

## FAQ

### What does “Shot 3 is somewhere else” mean?

SceneMixer read that shot as happening in another place or at another time than the rest of the segment. A segment is one place and one timeline, so that shot cannot be generated.

### Am I charged when the video is refused?

No. The refusal in SceneMixer says no credits were charged. Saving text is never blocked.

### How do I fix it?

Rewrite the shot to happen in this place and moment, move it into its own segment with its own location, or show the other place as a screen inside this one.

### Can I have a flashback in an episode?

Yes, as its own segment, with its own location and time. It cannot sit inside a segment that is set in the present.

### Is a phone call to another place a problem?

Only if a shot shows the other person in their own place. Keep the camera with the person you are following and write the other voice as off-screen.

### Does a moving car count as two places?

No. A car with its passengers is one place.

### Why do I not see the note on an older segment?

The note belongs to the current storyboard layout. Segments written before late October 2026, and ad and remake projects, are not checked this way.

### Can I switch the check off?

No. Rewrite the shot or split it into its own segment.

Updated November 3, 2026


---

# A Stranger Plays My Character: On Screen or Voice Only

URL: https://scenemixer.com/guides/a-stranger-plays-my-character-on-screen-or-voice-only/
When a stranger plays your lead or a narrator turns up on screen, check the character's Appears as setting. Where to change it, what changes, and the no-image warning.

# A Stranger Plays My Character, or a Voice-Over Got a Face

Quick answer: Two opposite symptoms come from one setting. Every character in step 2 is either On screen or Voice only (off-screen); the script reading picks one, and it can pick wrong. A lead marked Voice only gets no image and its face is never used, so the video invents a stranger whenever the storyboard puts them in the picture. A narrator marked On screen can end up drawn into shots. Since October 2026 you can change it: step 2 → open the character → Edit → Appears as → On screen or Voice only (off-screen) → Save. It is free. A character switched to On screen needs an image; segments already generated keep their old faces until you regenerate them.

Open step 2 of a project, click a character card, then Edit: the Appears as row is under the name. A card that says Off-screen and Voice only — no image generated is a voice-only character.

## Two symptoms, one setting

| What you see | Likely cause | Fix |
| A character the story clearly shows is played by someone else, a different face in every clip | The character is set to Voice only, or is On screen but has no image yet | Set On screen, generate their image, regenerate the segments |
| A narrator, a voice on the phone or an announcer appears in the picture | The character is On screen, or their lines lack an off-screen note | Set Voice only (off-screen); write their lines with (voice-over) or (off-screen); regenerate |
| The editor says In this segment but no character image yet: … | A character in the picture has no image | Press Generate images, make the image in step 2, then generate the segment |
If the face is the right person but drifts from shot to shot, that is a different problem: My character’s face changed: a ten-point checklist.

## How the reading decides, and how it gets it wrong

When SceneMixer reads your script, everyone with an action or a line becomes a character card. Someone who only ever speaks and is never seen — a narrator, the other end of a phone call, a voice on the radio — is marked off-screen: the card shows Off-screen and Voice only — no image generated, Generate All skips them, and their face is never used.

The guess can go wrong. A typical case: a main character who does not appear in the first episodes’ text, or who is only introduced in a character list at the end of the script, can be read as a voice. Then, when a later storyboard puts them in the picture, the video has nothing to draw them from and invents a face.

Two safety nets help. If you never touched the setting and a generated storyboard clearly shows the character in the picture, the setting is switched to On screen for you. And the editor warns you, segment by segment, when someone in the picture has no image (below). A choice you make yourself is kept as you made it.

## Change it: On screen or Voice only

- Open step 2 of the project and SceneMixer’s Characters tab.
- Click the character’s card, then Edit. The edit dialog opens.
- Find Appears as, under the name. Two buttons: On screen and Voice only (off-screen). Until you change it, it shows what the reading decided.
- Pick one. The line under the buttons says what it means:
- On screen: “When this character appears on screen, the video uses their character image.”
- Voice only (off-screen): “Heard but never seen (narrator, voice on the phone): no image is generated automatically and their face is never used in videos.”
- Press Save. Changing the setting costs nothing.
Only characters have this setting; scenes and items do not.

## After you switch to On screen

- Give them an image. Generate it from the card, or with Generate All on the Characters tab, which now includes them. Images are priced on the button; the first 15 setup images on a SceneMixer account are free. Generate All in step 2.
- Check the description describes how they look, not only how they sound; the image is drawn from it. Describe a face so it is not a generic face.
- Name them in the picture. In the storyboard, a visible character is an @ chip in the shot text. Who is in frame.
- Regenerate the segments they are in. Clips already made keep the stranger’s face; a new generation uses the image. Each regeneration is charged as usual; the old take stays in Version History.

## After you switch to Voice only

- No image is made for them, and SceneMixer’s Generate All skips them. An image they already have stays on the card but is not used in videos.
- Their face is never used. In storyboards written before you switched that still name them in the picture, they are treated as off-screen when a video is generated. The storyboard text you see does not change.
- Their voice still works. Keep a voice for them, and write their lines as off-screen: start the delivery note with (voice-over) or (off-screen), keep the voice chip, and do not put their @ chip in the picture text.
- Regenerate any segment where they were drawn in.

| Line | How to write it |
| A narrator | @Narrator (voice-over): She never told anyone what was in the letter. (voice: 🔊Narrator) |
| A voice on the phone | @Dad (off-screen, through the phone, tired): Come home, kiddo. (voice: 🔊Dad) |
Leads, supporting cast and voice-only characters · Lines, voice-over and captions.

## The warning: in this segment but no character image yet

Since October 2026 SceneMixer’s editor tells you before a face goes wrong. Above a segment whose picture names a character who has no image, a line reads:

In this segment but no character image yet: {names}. The video can only go by the text description, so the face won’t match the character.

with a Generate images button that takes you to step 2, where the character’s image is made (priced on the button, free while your free images last). When you press Generate all in the editor, the confirmation says the same for the whole batch: In these segments but no character image yet: {names}. The faces in the videos won’t match these characters.

It is a warning, not a block: you can still generate. The usual reasons a character has no image:

- The reading marked them Voice only, so Generate All skipped them, and later they were put in the picture.
- They arrived with new episodes added to the project after the images were made.
- Their image failed and was not retried.
Once the image exists, the warning disappears on its own.

## Fixing a project where it already happened

- Watch the episode and note every segment where the wrong face appears.
- In step 2, set the character to On screen and have SceneMixer generate the image. Check the image before going on.
- In the editor, open one of those segments: the shot text should name the character with an @ chip. If it names someone else or no one, fix the text. Shots with no linked characters.
- Regenerate that one segment and check the face before regenerating the rest.
- Regenerate the others, then composite again. Old takes stay in Version History.

| Setting | Image | Generate All | Face in videos | Voice |
| On screen | Needed; generated like any character | Included | Used whenever the character is in the picture | Yes |
| Voice only (off-screen) | Not generated automatically | Skipped | Never used | Yes |

## FAQ

### Why is my main character played by a stranger?

Most often the character is set to Voice only in SceneMixer, or is On screen but has no image yet. Without an image the video can only follow the text and invents a face. Set On screen in step 2, generate the image and regenerate the segment.

### Where is the On screen / Voice only setting?

Step 2, SceneMixer’s Characters tab: click the character, press Edit, and look for Appears as under the name. Pick On screen or Voice only (off-screen) and press Save.

### Does changing it cost anything?

No. Changing the setting in SceneMixer is free. Generating an image for the character and regenerating segments are charged as usual.

### Will the clips I already made change?

No. A finished clip on SceneMixer keeps whatever face it was made with. Regenerate the segments the character appears in; the old takes stay in Version History.

### My narrator keeps appearing in the picture. What do I do?

Set the narrator to Voice only (off-screen) in SceneMixer, write their lines with (voice-over) at the start of the delivery note, keep the voice chip, and leave their @ chip out of the picture text. Then regenerate the segment.

### Does a voice-only character still have a voice?

Yes. Voice only means heard but never seen. Match or pick a voice for them in SceneMixer's step 2 like any other character.

### Can I generate a video while the no-image warning is showing?

Yes. SceneMixer shows a warning, not a block, but the face of that character will not match. Generating the image first is the fix.

### Why did the setting change by itself?

If you never set it yourself and a generated storyboard clearly shows the character in the picture, SceneMixer switches the setting to On screen. A setting you chose yourself stays as you chose it.

### Can a scene or a prop be voice only?

No. In SceneMixer, only characters have the Appears as setting.

Updated November 3, 2026


---

# Ad and Remake Projects: What Is Different in the Wizard and Editor

URL: https://scenemixer.com/guides/ad-and-remake-projects-what-is-different-in-the-editor/
Ads and remakes run through the same steps as a script project, with a few controls hidden, renamed or fixed. The exact differences for each, so nothing looks missing.

# Ad Projects and Remake Projects: What Looks Different in the Wizard and the Editor, and Why

Quick answer: An ad and a remake are ordinary projects underneath: the same cast, scenes, storyboards and clips, priced the same way. What differs is what the wizard shows. Ads: the steps are named for ads (ad script; people, scenes, props; shots and final cut), the script can be edited or rewritten by AI, content type, dialogue language and region are not offered, the product carries a badge and is only ever redrawn from your original photo, segment length is fixed at 15 seconds, there is no delete-episode, the composite asks for the end card first, and the onboarding tour does not run. Remakes: there is no parse button (the read replaced it), region and styles are set in the style step and applied when the descriptions are tidied, dialogue language is not offered, the read produces the episode outline, and the storyboard is generated from Step 3 at the usual price.

Open an ad in advanced mode and a remake side by side: count the controls in Step 1; the missing ones are the ones listed here.

## Ads: the same three steps, renamed

Step 1 ‘Ad script’: the creative script, editable, with a button to let AI rewrite it; no content type, dialogue language or region (an ad is written in Chinese on a Chinese interface and in English otherwise). Step 2 ‘People, scenes, props’: the product card carries a badge; its image is redrawn from your original photo only, never from text; screenshots are used as they are. Step 3 ‘Shots and final cut’: one episode, segments fixed at 15 seconds, no delete-episode; the composite button first asks for the end card (brand, product name, tagline, call to action). Hand It to AI or Advanced Mode.

## Ads: what is fixed

Segment length (15 s), the number of segments (from the ad duration), the resolution tier chosen at the start, and the one-price quote if you chose to hand it to AI. Regenerating a shot, a sheet or the composite works as in any project and is priced per action. The End Card.

## Remakes: what replaced the parse

The read of your clip produces the cast, scenes, props and an episode outline written like a parsed script. So Step 1 has no parse button; the style step holds visual style, scene style, costume style and region, applied when you confirm and the descriptions are tidied to match. Dialogue language is not offered in Step 1: it is chosen on the upload screen (defaulting to your interface language) and fixed once the read starts. After the Read.

## Remakes: the storyboard step

Step 3 shows one episode with its outline and a generate-storyboard button at the usual storyboard price; the storyboard is written to the site’s own pacing, so the result runs longer than the source. Segments, clips and composite then work exactly as in a script project. After the Read, Generate the Storyboard.

## What is the same everywhere

Sheets and states, the segment editor, versions, the model selector, the badge switch, sharing, downloads, the library, credits and refunds. If a control is missing in an ad or a remake, it is one of the items above; nothing else is hidden.

| Control | Script project | Ad | Remake |
| Parse button | Yes | Yes (from the ad script) | No; the read replaced it |
| Content type / dialogue language | Step 1 | Hidden | Hidden |
| Region | Step 1 | Hidden | Style step |
| Segment length | 15 or 30 s | Fixed 15 s | 15 or 30 s |
| Delete episode | Yes | No | Yes |
| End card before composite | No | Yes | No |
| Onboarding tour | Yes | No | Yes |

## FAQ

### Why can't I change the ad's segment length?

Ad shots in SceneMixer are paced for 15-second segments; the count comes from the ad duration.

### Why is there no dialogue language in a remake's Step 1?

It is chosen on SceneMixer's upload screen and fixed once the read starts; it cannot be changed afterwards.

### Can I redraw the product from text?

No. The product image is always redrawn from your original photo so it stays your product.

### Does an ad get the first-project free allowance?

No; SceneMixer ads are never the free first project.

### Can a remake use the 30-second segment length?

Yes, like any script project; set it in Step 3.

Updated November 3, 2026


---

# Ad Blocked by Content Review: Edit the Shots Before You Retry

URL: https://scenemixer.com/guides/ad-blocked-by-moderation-edit-before-you-retry/
When the video provider's content check refuses an ad segment, the ad pauses and remembers it. Why retry is refused until those shots change, what triggers it.

# An Ad Blocked by Content Review: Why Retry Is Refused Until You Edit, and What to Change

Quick answer: If the provider’s content check refuses one or more segments of your ad, the ad stops at the video stage, the refused segments are refunded (within the daily limit below), and the progress page names them. The ad also remembers the exact text of those segments: pressing retry without changing them is refused, because resubmitting the same content would only be refused again and would count toward the account-level pause. Edit the refused segments in the shots view — the picture text, the lines, sometimes the product description — then retry; the other segments are not regenerated. Typical ad triggers: a named real person, a celebrity or brand lookalike, medical or financial claims, weapons or violence for a joke, revealing clothing. Superlatives are already cleaned automatically.

Open a paused ad’s progress page: the refused segments are listed with the provider’s reason; open one, change the text, and the retry button becomes available.

## What the progress page shows

The video stage marks each segment as done, refused or waiting. A refused segment shows the reason the provider gave. The retry button under the page is refused with a message while any refused segment still has the text it was refused with. The Ad Progress Page.

## Why retry is refused without an edit

The same text goes to the same check and gets the same answer; a block is normally refunded, but every block counts against your account’s daily moderation count, and enough of them pause all video generation for a while. Requiring an edit protects you from looping into that pause. Video Generation Paused.

## What to change

Read the provider’s reason. A person: the ad’s people are generated, and a real face in your photo is not carried over (product photos are redrawn as the product alone; people are removed from place and vehicle photos first), so look at how the script names or describes its people. Lookalike: a celebrity name, a franchise character, a famous logo in the script or on the product; rename or describe generically. Claims: cures, guaranteed returns, medical effects; write what is visible instead. Violence or weapons as a gag: cut it. Clothing: less skin in the shot text. The rules: Content Rules and Moderation.

## If the account-level pause is on

The progress page says so separately: video submissions for your whole account are paused for the stated time. Edit now, retry when the pause ends; the ad keeps its place. Repeated blocks across ads are reviewed by people; see the content rules for what leads to restrictions.

## What was charged

Refused segments are refunded automatically when the provider confirms the block, up to 12 blocks per account in 24 hours; a block past that, or on a day the site’s own daily refund limit is reached, is held for review instead of refunded. Segments that were delivered are charged and kept; the retry only regenerates the edited ones, priced by their seconds. The one-price cap still applies to the whole. The One Price Is a Cap.

| Reason shown | Likely cause in an ad | Change |
| Person / face | A named or recognisable person in the script | Describe a generic person |
| Lookalike / IP | Celebrity, character, famous logo | Rename, describe generically |
| Claim | Cure, guarantee, medical effect | Show, don't claim |
| Violence | A gag with a weapon | Cut it |
| Adult | Revealing clothing in the shot text | Describe the outfit plainly |

## FAQ

### Why can't I just retry?

The same text gets the same answer and each block counts toward the account pause.

### Do I lose the segments that passed?

No. On SceneMixer, only the edited segments are regenerated.

### Was the blocked segment charged?

No. Confirmed blocks on SceneMixer are refunded, up to 12 a day per account; past that the credits are held for review.

### Can the retry change the product photo?

No. SceneMixer fixes the photo when the ad is created; start a new ad with a different photo.

### Does a block affect my account?

One block, no. Repeated blocks trigger the temporary pause and are reviewed.

Updated November 3, 2026


---

# AI Ad Options: Vibe, Length, Format, Quality and Price | SceneMixer

URL: https://scenemixer.com/guides/ad-type-length-aspect-and-resolution/
The four pickers under the images: eight ad vibes, 15, 30 or 60 seconds, landscape or portrait, 480P to 1080P. What each changes, and how the all-in price follows.

# Ad Vibe, Length, Format and Quality

Quick answer: Under the images in the Ad Video composer sit four pickers. Ad vibe chooses the storytelling — Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life. Length is 15, 30 or 60 seconds. Format is landscape or portrait. Quality is 480P, 720P or 1080P, and the menu shows the all-in price of each. Length and quality move the price; vibe and format do not. Then one optional line about what it is and who it is for, and Make my ad.

The whole ad flow in one guide; this page is the four pickers.

## Ad vibe

How the ad tells its story. Epic: scale, spectacle, a big reveal. Funny: a joke with the product as the punchline. Warm: people, feeling, a small moment. Minimal: the product, light, silence, one line. Twist: it looks like one thing, then it is another. Mouth-watering close-ups: food and drink up close — steam, pour, cut. Show how it works: use it, see the result; good for apps, services, gadgets. A day in the life: someone living with the product, in the shop, in the clothes.

The vibe also sets the look of the ad and the colours of the end card. It does not change the price. The ad’s music is written for the ad itself: Where the Music in an Ad Video Comes From.

## Length

15, 30 or 60 seconds. On SceneMixer, an ad is made in fifteen-second segments — one, two or four — each shot on its own and joined in the final cut with the end card at the end. Longer means more segments, more shots, and a higher price; it also means more room for cast and sets, which the reading uses. Fifteen seconds is a single beat: the product and one idea. Sixty is a small story.

## Format

Landscape for sites, screens and pre-roll; Portrait for feeds and stories. It is the shape of every shot and of the end card. Product and place images are used at their own shape either way. Aspect ratio.

## Quality

480P, 720P or 1080P. The menu lists the all-in price beside each, so you can see what the step up costs before you pick. 480P is for drafts and small screens; 720P is what most people ship; 1080P is for big screens. If the quality you picked is temporarily unavailable when shooting starts, the ad tells you and offers to continue with one that is.

## The one line

Under the pickers: one line about what it is and who it is for. Optional. It wins over the photo when they disagree — write our family restaurant and the reading treats the photo as a place, not a product — and it seeds the story. Leave it blank and the reading builds the story from the image alone.

## The price

The button on SceneMixer reads All-in ✦ N for one-click generation: a cap, not an estimate — the actual charge is at most that. Length and quality change it; vibe, format and the line do not. Expert mode shows a smaller price to start and then prices each step on its button. One-click or expert.

## Changing your mind

Before Make my ad, change anything. After it, the ad is made with the choices you gave it; a different length or format is a new ad. Rebuild final cut on a finished ad re-joins the same shots, for example after editing the end card, without reshooting. The progress page.

| Picker | Options | Changes the price? |
| Ad vibe | Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life | No |
| Length | 15s, 30s, 60s | Yes |
| Format | Landscape, Portrait | No |
| Quality | 480P, 720P, 1080P | Yes — shown in the menu |

## FAQ

### Which vibe for a restaurant?

Mouth-watering close-ups for the food, A day in the life for the room and the people in it.

### Does a longer ad cost more?

Yes. It is made of more fifteen-second segments; the all-in price on the button follows.

### Does portrait cost more than landscape?

No. Format does not change the price on SceneMixer.

### Can I change the length after starting?

No. A different length is a new ad.

### Is the all-in price exact?

It is a cap. The actual charge for a one-click ad is at most that amount.

Updated November 3, 2026


---

# Add a Character, Scene or Prop to This Episode | SceneMixer

URL: https://scenemixer.com/guides/add-a-character-scene-or-prop-to-this-episode/
The editor's left panel: bringing a character, scene or prop into an episode with Add assets, what Edit (this episode) changes, and what needs a paid plan.

# Adding an Asset to This Episode, and Episode-Only Edits

Quick answer: The left panel of the editor lists what this episode uses. Add assets opens a picker over this project’s characters, scenes and items and adds the one you choose to this episode so it can be named in shots; it lists this project’s entries only. Edit (this episode) on an entry changes its image for this episode only; the note under it says so. Adding and episode-only edits cost no credits; uploading your own image and creating a brand-new entry need a paid plan.

Open a finished episode and click a character in the left panel to see the episode-scoped controls.

## What the panel shows

Three groups — Characters, Scenes, Items — holding the entries the storyboard names. An entry that is in the project but not in this episode is not shown, which is why the @ menu may not offer it: bring it in first.

## Add assets

Press Add assets. The picker shows this project’s entries by type with a search box; Added marks the ones already in. Pick one and it joins the panel. It does not list other projects: to bring a character, scene or item over from another project, open its card in step 2 and press Pick existing — the whole entry moves over, voice included, and it can be restored. Reusing entries across projects. Then name it in a shot with @. The @ menu.

An entry that does not exist anywhere yet is created in step 2, not here: Add on the step-2 tab, describe it, generate its image, then add it to the episode. Creating a new entry that the script did not produce needs a paid plan on SceneMixer. What a paid plan changes.

## Edit (this episode)

Click an entry in the panel and choose Edit (this episode). Replace its image — upload one, or make one with AI Fine-tune — and the change applies to this episode only; the note reads Changes apply to this episode only. The project’s entry in step 2 and every other episode keep the original. Both ways to replace the image need a paid plan: an upload costs no credits, while AI Fine-tune is priced like an image.

Use it for a look that belongs to one episode: the wedding dress, the bandaged hand, the room decorated for one night. A look that recurs across episodes is better as a state. AI Fine-tune · States.

## States in the panel

An entry with states shows them, and a segment in SceneMixer can be told which one to use: Use this state in this shot, Use this state in this episode, or Use main image. A dashed AI pick marks the state the storyboard chose for a shot on its own; your choice overrides it.

## Removing

An entry that the episode no longer names can be left in the panel; it does nothing until a shot names it. Delete the project entry in step 2 if it should not exist at all.

| Action | Scope | Cost |
| Add assets | This episode gains an existing entry | Free |
| Edit (this episode) — upload | This episode’s image for that entry | No credits; paid plan |
| Edit (this episode) — AI Fine-tune | This episode’s image for that entry | Priced like an image; paid plan |
| Use this state in this shot / episode | One segment, or this episode | Free (the state image itself is priced when made) |

## FAQ

### Why is a character missing from the @ menu?

It is not in this episode yet. Add it with Add assets in SceneMixer, then name it.

### Does Edit (this episode) change the character everywhere?

No. Only this episode. SceneMixer’s step 2 and other episodes keep the original.

### Can I bring in a character from another project?

Not from this picker — SceneMixer’s picker lists this project’s entries only. In step 2, open the card and press Pick existing: the whole entry comes over, voice included, and you can restore the original. Reusing entries across projects.

### Where did the crowd image line go?

SceneMixer no longer offers crowd images. Background people come from the one line in the segment header. Background people.

### Do I need a paid plan to upload an image for one episode?

Yes. Uploading on SceneMixer is a paid-plan feature. It does not cost credits.

### Does adding an asset cost credits?

No. Adding an asset to SceneMixer is free; the shots that use it are priced as usual.

Updated November 3, 2026


---

# Background Music for AI Short Drama: The Built-in Score

URL: https://scenemixer.com/guides/add-background-music-to-an-ai-drama/
Every episode is scored for you: music chosen for the episode and mixed under the voices. Where to hear it, how to switch it off, and how downloads and shares carry it.

# Background Music in an AI Short Drama: The Built-in Score

Quick answer: An episode on SceneMixer is scored for you. Once the first shot of an episode is generated, the site writes a plan for its music — which mood each stretch needs and where it turns — chooses a piece for each stretch (writing a new one when none fits), and mixes them in under the voices, shot by shot; the finished episode gets a version with the score as well. You add nothing. Playback and downloads carry the music by default; a Music switch beside the player turns it off and gives you the dialogue-enhanced version instead. The music is included with the videos and costs nothing extra. There is no track picker, no volume control, and no button to make the music again.

Open a finished episode and listen for the score: it comes in under the scene, steps back under a line, and returns after it.

## What an episode gets

The score is chosen for the episode by the site, not picked by you from a list. The site reads the storyboard — the scenes, the mood of each stretch, where the lines are — and plans the music: a handful of pieces per episode, each playing for at least half a minute and changing at scene boundaries, with a short crossfade from one piece into the next. The music runs through the episode instead of leaving long silences, and it sits lower under every line. Everything is mixed under the voices; under a line of dialogue the music steps further back. A series keeps one sound: the instrumentation follows the show’s genre, setting and era, so the music stays in one family across the episodes.

## When it appears

SceneMixer starts scoring by itself once the episode’s first shot is generated. While an episode is being scored, the Music switch beside the player shows a small in progress note, and each shot plays without music until its scored version is ready — usually within minutes, and sooner for the later shots. New shots are mixed as they arrive, so a fresh episode fills in piece by piece rather than all at once.

## Where you hear it, and the two switches

A shot, in the editor: the Music switch sits next to Play continuously under the player. On — the default — the shot plays and downloads with the score. Off: both give the dialogue-enhanced version without music. The setting is saved to your account, so it follows you between projects and devices (on a guest session it applies to this browser).

The composite: each finished composite in the composite records has its own Music switch, shown once the episode has a scored version of the composite. On, playback and download both use the version with the score; off, playback is the plain composite and the download is the dialogue-enhanced version without music. While that version is still being made, the switch shows in progress and the download gives the dialogue version.

The asset library: no separate switch there — the library follows the setting from the editor.

Shared links: a shared link always plays the version with the episode’s music, whatever your switch says. The switch is about what you play and download here, not about what your viewers get.

## Downloads

SceneMixer’s download button always gives the full-quality file. With the switch on it carries the score; with the switch off you get the dialogue-enhanced version without music — that is how you take the episode into an editor without the site’s music under it. Composite and share an episode covers the composite side.

## Music the characters can hear

A separate thing from the score: if the music is something the characters could hear — a radio, a busker, a piano next door — write it as sound in the shot: Ambient: a radio playing an old love song in the kitchen, Ambient: the café radio, quiet. SceneMixer generates it with the clip, like rain or traffic, and it lives inside the scene. Keep it short and specific, and say that it sits under the dialogue: quiet, far off.

## Your own music instead

- Turn the Music switch off for the shots and the composite; the downloads now come without the score.
- Download the composite (or a single shot) and open it in any video editor, or in the editing tools of the platform you post to.
- Lay your licensed track under the whole episode, so the music runs across the cuts, well below the voices.

If a clip happens to carry music of its own — older clips, or a take where the video model added some — the site’s score keeps out of those seconds, fading away and coming back, so you never hear two tracks fighting.

## When the episode changes

Regenerating a shot, re-editing it or lip-syncing it produces a new version, and SceneMixer mixes the score under it automatically; when you change lines, the music is fitted around the new timing. Regenerating the storyboard is the one thing that makes the music itself anew: the plan follows the new shots and the episode is scored from scratch. Episodes that were scored before October 23, 2026 keep the music they have; only newly scored episodes use the newer method. Shots inserted after the plan was made can be left without music, because the plan covers the shots that existed when it ran.

## Ads and remakes

Ad projects and remakes on SceneMixer are scored by the same system, but not by the short-drama method of choosing a piece for each stretch. Ads differ in two ways: the music is written for that single ad, and it is carried through to the end of the final cut. Where the Music in an Ad Video Comes From · Remake: how dialogue and music are read from the clip.

| You want | Where it comes from | How |
| The episode scored | The episode’s own score | Automatic; Music switch on |
| The episode without music | The dialogue-enhanced version | Turn the Music switch off before downloading |
| Your own track instead | Your edit | Download the version without music, add one track under it |
| Music the characters hear | The clip | Write it in the shot’s Ambient line |
| An ad’s music | Written for the ad | Comes with the finished ad |

## FAQ

### Do I need to add music to my episode?

No. Every episode on SceneMixer is scored for you once its first shot is generated; the music is chosen for the episode and mixed in under the voices.

### Does the music cost extra?

No. The music comes with the videos on SceneMixer; nothing is charged for writing, generating or mixing it.

### Can I choose the track, or change its volume?

No. The score is chosen for the episode and mixed to sit under the voices. SceneMixer has no track picker and no volume control.

### Can I download the episode without music?

Yes. Turn the Music switch off first; SceneMixer’s downloads then give the dialogue-enhanced version without music, for both a shot and the composite. Turn it back on for the scored version.

### Why does one shot have no music?

The score in SceneMixer is mixed in as shots arrive, so a fresh shot can be waiting for its scored version, and the score steps aside for the seconds where the clip already carries music of its own. A shot added after the episode was planned can also be left without music; regenerating the storyboard makes the plan follow the new shots.

### Can I get different music?

Not from a list, and there is no button in SceneMixer to make the music again. The score follows the episode’s storyboard; regenerating the storyboard is what makes a new plan and new music.

### My episode was made before October 2026 and has no music. Will it get some?

No. Scoring on SceneMixer applies to episodes whose first clips were generated on or after October 3, 2026; older episodes are not scored later. Download and add your own track if they need music.

### A shot carries music I did not ask for. Can I remove it?

The clip’s own music is part of its sound and cannot be taken out on its own, but SceneMixer’s score stays out of those seconds, so the two never play at once. Regenerating the shot is the way to a take without it.

### Does the music carry into downloads, the library and shared links?

Yes. Downloads and SceneMixer’s asset library follow your Music switch; a shared link always plays the version with the episode’s music.

Updated November 3, 2026


---

# Add Subtitles to an AI Drama: Burned-In Captions and SRT

URL: https://scenemixer.com/guides/add-subtitles-to-your-ai-drama/
Turn on Subtitles in the SceneMixer editor to caption every line from your script, download the video with subtitles burned in, or export an SRT for the whole episode.

# Add subtitles to your AI drama: on-screen, burned in, or as an SRT file

Quick answer: In the episode editor, turn on Subtitles next to the Music switch. SceneMixer listens to each clip, matches what is spoken to the lines in your script, and shows them over the video; downloads then come with the subtitles burned in. For a whole episode, tick Add subtitles when you export it, and the export gets Download with subtitles and Download SRT. The words are your script’s, so names and spellings come out exactly as you wrote them; the timing comes from the voices. Subtitles cost no credits. They are available on accounts that have made at least one purchase.

Writing lines so they are spoken: write dialogue in the storyboard.

## Who has it

Subtitles are switched on for projects of accounts that have bought credits or a plan at least once; an order that was fully refunded does not count. If you do not see a Subtitles switch next to Music in the editor, your account is not in that group yet.

## Subtitles on one clip

- Open an episode in the editor and select a segment that has a finished video.
- Next to Music, turn Subtitles on. The hint says it plainly: On: show subtitles over the video and download the subtitled version. Off: play and download as-is.
- The first time, you will see Making subtitles while every finished segment of the episode is listened to. It takes a moment; you can keep working.
- Play the clip: each line appears as it is spoken.
- Press Download. The subtitled version is prepared on request — it will download automatically when ready (keep this page open).

The switch is remembered for your account, so the next episode you open starts with subtitles on.

## Subtitles on the whole episode

- Press Export episode once every segment is done.
- In the export dialog, tick Add subtitles: SceneMixer transcribes the whole episode once and aligns it to the script and adds a minute or two.
- When the export is ready, open Exports. With subtitles on, the export offers Download with subtitles and Download SRT.

The subtitled download follows the Music switch: with music on you get the version with the episode’s background music, with it off the version without. The SRT file is plain text with times, ready to upload to YouTube or any player.

Exported without ticking it? Turn the subtitles switch on in Exports: the exports listed there are subtitled the same way.

## Where the words come from

The text is always your script’s lines, not a transcription. SceneMixer recognises the speech only to find when each line is said. That is why a character’s invented name, a place in another language or an unusual spelling comes out exactly as you wrote it.

- A line that was not spoken is left out of the subtitles, and the editor tells you: Line 3 was not spoken. See when the video says or skips a line.
- Something said that is not in the script — a stray word — is not subtitled.
- A clip with no dialogue shows No dialogue audio in this clip and simply has no subtitles.

## When you change the lines afterwards

If you edit a segment’s lines after its subtitles were made, the editor says Subtitles still match the script at generation time and offers Realign to new lines. That re-times the new text against the same recording; it costs nothing and does not count toward the daily limit.

For a whole-episode export, if segments were regenerated or deleted after you exported, its lines no longer match: export the episode again and tick Add subtitles.

## What the subtitles look like

White bold text with a soft shadow, no grey box, at most two lines, placed low in the frame and sized to the video: a vertical 9:16 clip gets a wider line than a landscape one, and in landscape the subtitles sit above the corner badge. Chinese, Japanese and Korean use their own typefaces; punctuation follows subtitle conventions. The editor overlay and the burned-in download use the same layout.

## Cost and limits

Subtitles cost no credits: listening to a clip, making the subtitled download and exporting the SRT are all free. To keep it that way there is a daily limit per account on how many segments and episodes are subtitled and how many subtitled downloads are made; ordinary use of a series does not come near it. Realign to new lines does not count.

## Messages you may see

| Message | What to do |
| Subtitles failed · turn off and on to retry | Turn the switch off and on again. |
| The subtitled version is taking longer than usual | Click Download again in a few minutes. |
| Couldn’t make the subtitled version this time. Download the version without subtitles? | Download without, or try again later. |
| The storyboard changed after this export, so its lines no longer match | Export the episode again with Add subtitles ticked. |
| No dialogue audio in this clip | Nothing; a clip without lines has no subtitles. |

## FAQ

### Does SceneMixer add subtitles automatically?

Yes, once you turn the Subtitles switch on in the editor. Every finished segment is listened to and each script line is timed to the voice.

### Can I download the video with subtitles burned in?

Yes. With Subtitles on, Download gives the subtitled version, for a clip or for a whole exported episode.

### Can I get an SRT file?

Yes, for a whole episode: tick Add subtitles when you export, then use Download SRT in Exports.

### Do subtitles cost credits?

No. They are free on SceneMixer, with a daily limit per account.

### Why don't I see the Subtitles switch?

Subtitles are available on SceneMixer accounts that have bought credits or a plan at least once.

### Are the subtitles a transcription?

No. The text is your script's lines; speech recognition is only used for the timing. Names come out spelled as you wrote them.

### What if the video skipped a line?

That line is left out of the subtitles and the editor notes which line was not spoken. Regenerate the segment if you need the line said.

### I changed the lines. Do I pay again?

No. Use Realign to new lines; it is free and does not count toward the daily limit.

### Can I have subtitles in a different language from the dialogue?

The subtitles follow the dialogue's own script. For a second language, make a second-language version of the drama; see make a second-language version.

Updated November 3, 2026


---

# Affiliate Earnings Statuses Explained: Received, Under Review, Voided

URL: https://scenemixer.com/guides/affiliate-earnings-statuses-explained/
Received, Under review, Clearing, Withdrawable, Paid, Converted, Voided, clawed back, capped, Refund adjustment: what each mark in your affiliate earnings means.

# Every Status in Your Affiliate Earnings, Explained

Quick answer: Each row of Earnings is one payment by someone you referred: the date, who, the type (credits or membership), what they paid, the rate, a status and what you earn. Credit commissions are normally Received: already in your balance. Under review and Voided rows are greyed because you do not have that money. A red clawed back note means the payment was refunded; (capped) means that referral has reached the credit limit. Cash rows move through Clearing, Withdrawable, Being paid and Paid.

If a number looks too low, look under it: a red 'clawed back' amount is a refund of the referred payment, not a mistake.

## Credit commission statuses

- Received — the credits were added to your balance when the payment cleared.
- Under review — held for a check, for example when the referred account was created on the same device as yours. It is either released (becomes Received) or voided. The card above the table adds up how many credits are under review.
- Voided — will not be paid: a self-referral, a commission from your own extra account, a review that was not passed, or a payment made while the program was paused.

## Cash commission statuses (partners)

- Clearing — the payment to SceneMixer is less than 30 days old.
- Withdrawable — cleared; waiting for the monthly payout.
- Being paid — included in the payout being sent.
- Paid — sent to your payout method.
- Converted — turned into credits at your request.

How payouts work.

## Marks next to the amount

- −N clawed back (red) — the referred person was refunded or charged back, so the matching share of your commission was taken back.
- (capped) — your credit commission from this one referral has reached the program’s limit per person; further payments by them earn nothing more. Your other referrals are unaffected.
- Refund adjustment (type column) — a cash commission that had already been paid out or converted was refunded afterwards; this negative row is netted off your next payout.

## Credits still owed

When a credit commission is clawed back but your balance cannot cover it, SceneMixer’s page says N credits from refunded orders are still owed. Future credit commissions cover that first. Nothing else is touched: the owed amount is simply taken from your next credit commissions before they reach your balance.

## What the other columns mean

- From — the referred account, partly hidden.
- Type — Credits for a top-up, Membership for a plan payment, Refund adjustment as above.
- They paid / Rate / You earn — what the person paid, your percentage, and the result.

How the affiliate program works.

| You see | It means |
| Received | Credits are in your balance |
| Under review | Held for a check; released or voided |
| Voided (grey) | Will not be paid |
| −N clawed back | The referred payment was refunded |
| (capped) | Limit for this referral reached |
| Clearing / Withdrawable / Being paid / Paid | Cash commission stages |
| Refund adjustment | A negative row netted from the next payout |

## FAQ

### Why is a row Under review?

Usually because the account was created on the same device as yours. Genuine referrals on SceneMixer are released.

### Why was a commission voided?

It came from your own account, did not pass review, or the payment was made while SceneMixer’s affiliate program was paused.

### Do I lose credits I already spent when something is clawed back?

SceneMixer never takes your balance below zero; what cannot be covered is owed and taken from later credit commissions.

### What does (capped) mean for the future?

That referral earns you nothing more in SceneMixer credits. New referrals start from zero.

Updated November 3, 2026


---

# Affiliate Payouts: When Cash Commission Becomes Withdrawable

URL: https://scenemixer.com/guides/affiliate-payouts-when-cash-becomes-withdrawable/
How a partner's cash commission moves from Clearing to Withdrawable to Paid: the monthly payout and its minimum, payout methods, refunds, converting cash to credits.

# Affiliate Payouts: Clearing, Withdrawable, Paid, and Converting to Credits

Quick answer: A cash commission starts as Clearing, becomes Withdrawable 30 days after the payment it came from, and is paid out in the monthly payout once your withdrawable balance reaches $50, to the payout method on your page. If you would rather spend it on the site, Convert to credits turns a withdrawable balance into credits at 100 credits per $1 — a one-way step.

Set your payout method on the cash commission card before your balance reaches the minimum, so the first payout is not held up.

## The three amounts on the card

- Clearing — commission from payments made in the last 30 days. It waits in case the payment is refunded or charged back.
- Withdrawable — cleared commission, ready for the next payout.
- Paid out — the total already sent to you.

Under them, one line tells you where you stand: $X to go before your first payout, You have reached the minimum. It goes out in the next monthly payout, or $X is being paid out.

## The monthly payout

SceneMixer pays out once a month to accounts whose withdrawable balance is at least $50. Everything withdrawable at that point is included. While a payout is being sent, its commissions show Being paid; when it has gone they show Paid, and the payout appears in the Payouts list on the page. Below the minimum, the balance simply carries over to the next month.

## Payout method

PayPal, Wise, Alipay or bank transfer, chosen when you apply and editable on the cash commission card under Payout method. Changing it does not affect a payout already in progress; the next payout goes to the new details. If it says Not set yet, set it before the payout date.

## Converting to credits

Convert to credits on SceneMixer turns your withdrawable balance into credits at 100 credits per $1 of commission. The confirmation says exactly how many credits you will get and that it cannot be undone. Converted commissions show Converted in your earnings, and the credits appear in your credit history. Only withdrawable money can be converted; clearing commission has to clear first.

## Refunds after you are paid

Commission follows what the referred person actually paid. If they are refunded or charge back while the commission is still clearing or withdrawable, it is reduced in place. If it was already paid out, converted or included in a payout, a Refund adjustment row with a negative amount appears instead, and it is netted off your next payout. Every earnings status.

| Status | Means | Next |
| Clearing | Payment was less than 30 days ago | Becomes withdrawable |
| Withdrawable | Cleared, waiting for the monthly payout | Paid when the balance is at least $50 |
| Being paid | In this month's payout | Paid |
| Paid | Sent to your payout method | — |
| Converted | Turned into credits | — |

## FAQ

### Why the wait before cash is withdrawable?

Refunds and chargebacks can arrive after a payment; SceneMixer’s clearing period covers them.

### Can I ask for a payout below the minimum?

No. Your SceneMixer balance carries over until it reaches the minimum; you can convert it to credits instead.

### Can I undo a conversion?

No. Converting to credits on SceneMixer cannot be undone.

### In what currency am I paid?

In the currency SceneMixer shows on your cash commission card.

Updated November 3, 2026


---

# Remake: After the Read, Generate the Storyboard (Why It Runs Longer)

URL: https://scenemixer.com/guides/after-the-read-generate-the-storyboard/
A remake reads your clip into a scene list with lines; the storyboard is then written like any script at the usual price. Why the result runs 1.5–2× the clip.

# Remake a Video: After the Read You Generate the Storyboard — and Why the Result Runs Longer Than the Clip

Quick answer: Reading a clip produces a scene-by-scene script: who is there, what happens, every line of dialogue, in order. It does not produce a shot list you can shoot straight away. After you pick the look in Step 1, open the episode in Step 3 and press Generate storyboard, priced like any episode; the storyboard is written from the scene list plus a shot-by-shot record of the original, and then segments are generated one by one at the tier you choose. The result usually runs 1.5–2× the clip’s length, because AI dialogue is paced slower than real actors and each shot gets the seconds it needs; the price shown before the read already estimates the video cost on that basis.

Upload a short clip from the Remake tab; when it says Read, open the project and watch Step 3 offer Generate storyboard for episode 1.

## What the read gives you

The read cuts the clip into shots, identifies the people, places and props, listens to every line and writes it all down as one episode: a scene list with prose and dialogue, plus a sheet for each character, place and prop. None of the clip’s frames are used as references for the new video; the cast is drawn from the descriptions, in the visual style you choose. Limits, read time and price of the read itself: Source Limits, Read Time and Price.

## Why the storyboard is a separate step

A clip’s cuts are a real editor’s cuts — one or two seconds each — and a video model cannot shoot that. SceneMixer’s storyboard step regroups the shots into segments a model can hold, gives spoken lines the seconds they take, writes the light, the camera and the movement for each shot, and runs the same consistency checks as any script. That is why it is generated, not copied, and why it costs the storyboard fee: Generate the Storyboard for an Episode.

## Why the result runs longer than the clip

Real actors talk fast and editors cut tight. AI dialogue is paced at reading speed, and each shot gets a minimum of a few seconds, so a 26-second clip with nine lines becomes about 45–60 seconds of video. This is the model’s pace, not a setting; you can trim seconds shot by shot in the editor afterwards, within the limits of the spoken lines. Shot Seconds and Timing.

## What it costs, in order

The read (by clip length, within the limit shown on the upload box), then the sheets in Step 2, then the storyboard for the episode, then each segment per second at the tier you pick, then the composite. The estimate SceneMixer shows before the read already scales the video part to the longer running time. Everything after the read is priced exactly like a script project: Where Your Credits Go in One Episode.

## Crowds and background people

A group in the clip — guards behind a counter, guests at a table — is not turned into a character with a sheet. It is written into the scene as background people, and the storyboard declares them where they appear, so the video shows a crowd rather than one face copied many times. Recasting the real characters: Recast a Remake.

| Step | What you get | Price |
| Read the clip | Scene list with lines, sheets to generate | By clip length |
| Step 1 | Pick the look, aspect, region | Free |
| Step 2 | Generate the sheets | Per sheet |
| Step 3: Generate storyboard | Segments and shots written from the scene list | Storyboard fee |
| Generate segments | Video, about 1.5–2× the clip's length | Per second, by tier |

## FAQ

### Why is there no storyboard right after the read?

SceneMixer’s read writes the story down; the storyboard is written from it in Step 3, like any script, so it can be paced and checked the same way.

### Can I make the result the same length as the clip?

You can trim shot seconds in SceneMixer’s editor, but spoken lines keep the seconds they need. Expect it to run longer.

### Does the remake use frames from my clip?

No. SceneMixer uses only the descriptions and the script; the picture is new.

### Is the storyboard fee extra?

It is the same fee any episode pays on SceneMixer; remake projects do not get the first project’s free first-episode storyboard.

### Why does a group of people not get a sheet?

Groups in SceneMixer are background people in the scene, not characters; one sheet would put one face on all of them.

Updated November 3, 2026


---

# After a Remake Read: Style, Region and Dialogue Language | SceneMixer

URL: https://scenemixer.com/guides/after-the-read-style-region-dialogue/
A remake opens at step 1 with the cast already read. Why you pick the three styles before any image is made, what Target Region changes, why dialogue language is fixed.

# After the Read: Style, Region, and Why Dialogue Language Is Fixed

Quick answer: When the read finishes, Open the project lands you at step 1 with the summary already filled — but no images yet. Choose the visual style, scene style and costume style (or let AI pick), the aspect ratio and, if you want the story relocated, the target region; then Next. The read cast, sets and props are restyled to your choices before step 2 generates their images. The dialogue language is not on this step: lines were transcribed and translated at read time, so it was chosen on the upload screen before you pressed Start remake (it defaults to your interface language).

The remake overview for the whole path; this page is the step you land on.

## Where you land

Step 1, Set Up Your Project, with the settings form on screen and the summary of what was read beneath it. There is no Parse here — the project was read from your video, so there is nothing to parse — but Next will not move on until the three style choices are made. That is deliberate: every image and every shot is drawn in the visual style, so it must be chosen before the first image exists.

## The three styles

Visual Style: how everything is drawn — live action, anime, 3D. Scene Style and Costume Style: what the sets look like and what people wear. When some are unset, a dialog offers Let AI choose from the read; accept or pick yourself. Press Next and the page says Applying your style choices to the cast, sets and props: the descriptions read from your clip are restyled so that step 2 draws them the way you asked. The presets · How the styles work together.

## Target Region

Optional. Pick a region and the cast’s ethnicity, the architecture and the props are moved to that market when the descriptions are restyled — names stay as read, because they are already written into the storyboard. Leave it on Follow script and the world stays as the clip showed it. Region.

## Why dialogue language is not on this step

In a drama, the dialogue language is applied when the script is read, and a script can be read again. A remake’s lines were heard and written down during the read, in the dialogue language you chose on the upload screen — that is the moment the choice takes effect, and there is no second read. So this step does not show the row: a picker that changed nothing would be worse than none. To remake a clip into another language, pick it on the upload screen before you start; lines heard in another language are translated into it.

## Aspect ratio

Choose it here like any project; it does not have to match the source. A landscape clip can be remade portrait. Aspect ratio.

## Then step 2 and onwards

Step 2 shows the cast, sets and props read from the clip, without images; generate them as in any project, or upload your own. Step 3 already has the storyboard, so the episode is ready to edit and shoot. From here it is a drama. Changing the cast and places.

| Setting | In a remake | When it applies |
| Visual / scene / costume style | Required before Next | When the read descriptions are restyled, then in every image and shot |
| Target Region | Optional | Same moment; names are kept |
| Aspect ratio | Your choice | Every video shot |
| Type | Set for you | — |
| Dialogue language | Not on this step | Chosen on the upload screen; fixed once the read starts |

## FAQ

### Why does Next refuse until I pick styles?

On SceneMixer, every image and shot is drawn in the visual style. It must exist before the first image is made.

### Can I change the language the remake speaks?

Not after the read. Pick the dialogue language on SceneMixer's upload screen before you start; the lines are written in it during the read.

### Will the region change the characters' names?

No. Names are already in the storyboard. Ethnicity, sets and props follow the region.

### Do I have to parse anything?

No. The project was read from your video. Next saves your choices and restyles the descriptions.

### Can I remake a landscape clip as portrait?

Yes. Pick the aspect ratio in SceneMixer; it need not match the source.

Updated November 3, 2026


---

# AI Create: After the Script Is Written, What Happens Next | SceneMixer

URL: https://scenemixer.com/guides/after-the-script-is-written/
What the writing panel shows, what Minimise does, where the finished script goes, what Cancel and Start Creating do, how to change a draft, where drafts are kept.

# After the Script Is Written

Quick answer: Press Start Writing and the panel shows Thinking, then AI Writing with a character count as the script grows. Minimise folds it into a corner capsule; you can leave and are told when it is done. When AI Writing Complete appears, the script is on screen with Start Creating ✦ N: that price is the read, and pressing it creates the project at step 1; Cancel just closes the window. To change the draft, write again with a sharper brief; every draft stays in Writing History, and any of them can be started later.

Upload and parse is what Start Creating does with the draft.

## While it writes

Thinking: the structure is being worked out. AI Writing…: the text grows with the character count. Minimise keeps it as a capsule at the bottom right; you can use the rest of the site, and the capsule opens the panel again. There is no stop button: a writing that has started runs to the end, and it is charged when the script is delivered.

## AI Writing Complete

The script is shown in full. Read it. The button at the bottom is Start Creating ✦ N: N is the read price for this script’s length, the same as uploading it. Press it and the project is created and opens at step 1 with the settings form; the script is the project’s text. Cancel closes the window instead: nothing is created, and the draft waits in history.

## Changing the draft

The script text is not edited in SceneMixer’s window. To change it: write again with a sharper brief (the usual fix), or copy the text, edit it in any editor, and paste it into Upload Script. Small changes are better made after the read: the summary, the cast descriptions and every storyboard are editable for free inside the project. A sharper brief.

## Writing History

The clock icon on the AI Create tab. Every draft is listed with its brief, length and date. Open one to read it and Start Creating from it at today’s read price. Drafts are not projects and do not count towards the project limit.

## What was charged

The writing, when the script was delivered: one AI script writing line in SceneMixer’s Credit History. The read, when you press Start Creating. Nothing for closing the result window, nothing for opening history. Credit History.

| Button | Does | Charged |
| Minimise | Folds the panel into a corner capsule | — |
| Cancel | Closes the result window; the draft stays in history | No |
| Start Creating ✦ N | Creates the project and reads the script | The read |
| Close window | Nothing; draft stays in history | No |
| Writing History | Lists every draft; start any of them | No |

## FAQ

### Can I leave while it writes?

Yes. Minimise it; SceneMixer’s capsule keeps it and you are told when it is done.

### What is the price on Start Creating?

On SceneMixer, the read is priced by the script's length, the same as uploading that text.

### Can I edit the script in the window?

No. Write again with a better brief, or paste an edited copy into Upload Script. After the read, everything in SceneMixer is editable.

### Where did my earlier drafts go?

Writing History on SceneMixer’s AI Create tab. Any of them can be started later.

### Can I stop a writing halfway?

No. Once started on SceneMixer, it runs to the end and is charged on delivery; minimise it and carry on with something else.

Updated November 3, 2026


---

# Make an AI Ad for a Car or a Dealership | SceneMixer

URL: https://scenemixer.com/guides/ai-ad-for-a-car/
How a car photo is handled: kept as shot, people removed, never redrawn. Which angles to photograph, the vibes that suit a car, and how the badge reaches the end card.

# An AI Ad for a Car

Quick answer: Photograph the car in its setting — three-quarter front, then a side, an interior, a detail — and the reading files it as a vehicle. A car is never cut out onto a white background; that would lose its proportions and its badge. Your photo is kept, the people in it are removed, and the ad puts a generated driver and a road around your actual car. Epic for a reveal, A day in the life for a family car, Show how it works for a feature. The name on the badge is read and printed on the end card.

Which photos work covers the general rules; this page is the car case.

## The photos

First: three-quarter front, the whole car in frame, in daylight, on a clean surface. Then, as extra images: a side profile, the interior from the driver’s door, one detail — the wheel, the light, the badge. A showroom photo with a salesperson in it is fine; the composer says People will be removed. A brochure render is not a photo of your car; use your car.

## What is kept

Cars, like shops and homes, are not redrawn. The shape, the colour, the badge, the wheels, the setting in your photo are what the shots are built from. Redraw is not offered for the main image (The main product image can only be redrawn from your original photo); if the photo is wrong, pick different images.

## What the ad shows

The car moving through a place that fits it: a coast road, a city at night, a driveway at dusk. A generated driver and passengers; a reveal, a drive, an arrival. The extra photos become the angles the shots cut to — the interior when someone gets in, the detail on the freeze. Generated video does not write legible text, so the badge is read for the end card rather than shown in close-up.

## Which vibe

Epic: scale, landscape, the reveal. A day in the life: school run, weekend, the boot full. Show how it works: one feature — the parking assist, the boot that opens itself. Minimal: the car, light, one line. Thirty seconds for a reveal and a drive; sixty for a story with the family in it.

## The one line

A seven-seat hybrid for families who camp, or the used-car lot on Main Street, open Sundays. Model names and dealership names go here so the script and the end card have them. SceneMixer leaves claims about range or safety ratings out of the script; say what can be seen.

## Dealership or single car

Selling one car: the car first. Selling the dealership: the lot or the showroom first, filed as a place, with a car as the second photo. The end card then carries the dealership’s name and Book a test drive. Ads for a place.

## The end card

Brand from the badge or your line, product name from the reading dialog (the model), tagline, and a call to action: Book a test drive, See it in the showroom. The end card.

| Photo | Role in the ad |
| Three-quarter front, whole car | The featured product; every shot is built from it |
| Side profile | A cut-to angle on the drive |
| Interior from the driver’s door | The moment someone gets in |
| One detail: wheel, light, badge | The freeze before the end card |

## FAQ

### Will the car be cut out onto a white background?

No. SceneMixer never redraws cars; the photo is kept and the ad is built around it.

### There is a salesperson in my showroom photo.

Fine. People are removed and the car and the room are kept.

### Will the badge be readable in the shots?

Not reliably; generated video does not write legible text. SceneMixer reads the name and prints it on the end card.

### Which vibe for a reveal?

Epic. For a family car, A day in the life.

### Can I advertise a used-car lot rather than one car?

Yes. Put the lot or showroom first; it is filed as a place, and a car can be the second photo.

Updated November 3, 2026


---

# Make an AI Ad for a Physical Product from One Photo | SceneMixer

URL: https://scenemixer.com/guides/ai-ad-for-a-physical-product/
Gadgets, cosmetics, drinks, tools, toys: how one photo becomes a product shot, how the ad keeps the same object in every hand, and which vibe fits.

# An AI Ad for a Physical Product

Quick answer: Photograph the product alone, filling the frame, in even light, and the reading files it as a product. It is cleaned into a product shot on a clean background and becomes the featured prop of the ad: the same object, held, opened, poured or switched on by a generated cast, in every shot. Show how it works demonstrates it; Epic, Twist or Funny dramatise it; Mouth-watering close-ups is for food and drink. Fifteen seconds for one idea; the end card carries the brand and Shop now.

The whole flow from one product photo is the canonical case; this page adds the details.

## The photo

The product alone on a plain surface, front three-quarter, large in frame, in even daylight. Packaging on if the packaging is what people recognise; out of the box if the object is. Extra images: the back, the product in use (people are removed; the object is read), a detail. A lifestyle photo with the product small in a busy scene reads badly as a first image.

## What the reading does

The photo is cleaned into a product shot: the object on a clean background with its shape, colour, material and labels kept. The name printed on the package is read for the end card. That shot is then a prop in the ad’s cast list, and every shot that shows the product shows that object. Why a prop stays the same.

## What the ad shows

Someone using it. A generated person in a real place takes the product out, opens it, pours it, switches it on, and something changes. The object stays your object throughout. Labels and text on it are what your photo shows; generated video does not write new legible text, so the shots do not invent a slogan on the box.

## Which vibe

Show how it works: the honest demo — use it, see the result — for gadgets, tools, cosmetics. Mouth-watering close-ups for anything edible or drinkable. Epic for a launch. Twist and Funny when the product has a story. Minimal for a design object. Fifteen seconds for one idea; thirty for problem and solution.

## The one line

A rechargeable hand warmer for people who cycle in winter. What it is and who it is for shape the person and the moment. Keep to what can be shown; efficacy claims, superlatives and guarantees are removed from ad scripts.

## The end card

Brand from the package or your line; product name from the reading; a tagline (spoken in the ad before the freeze); and a call to action — Shop now, Find it in stores. The card’s main image is the cleaned product shot. The end card.

## Expert mode for products

If you want a specific sequence — unbox, first use, the result — choose expert mode, read the script AI wrote, edit it, and shoot the shots yourself. The product shot is already prepared for you. One-click or expert.

| Product | First photo | Vibe | Length |
| A gadget or a tool | Out of the box, three-quarter | Show how it works | 15–30s |
| Cosmetics or skincare | The bottle or jar, front | Show how it works / Minimal | 15s |
| A drink or a snack | The pack, front | Mouth-watering close-ups | 15s |
| A toy | The toy, three-quarter | Funny / A day in the life | 30s |
| A design object | On a plain surface | Minimal | 15s |

## FAQ

### Will the product look like my product?

Yes. SceneMixer cleans the photo into a product shot, and that object appears in every shot.

### Should I photograph it in the box?

In the box if the packaging is what people recognise; out of the box if the object is.

### Will the label be readable in the shots?

The label is what your photo shows; the shots do not add new text. The name is delivered on the end card.

### Can I show it being used?

Yes. That is what the ad does: a generated person uses it in a real place. Show how it works is built for this.

### Can I claim it works better than others?

SceneMixer removes comparative and efficacy claims. Show what it does; let the viewer judge.

Updated November 3, 2026


---

# Make an AI Ad for a Restaurant or a Dish | SceneMixer

URL: https://scenemixer.com/guides/ai-ad-for-a-restaurant/
Which photos to take, whether to lead with the room or the food, which ad vibe fits, what happens to the customers in your photo, and what the end card should say.

# An AI Ad for a Restaurant, a Café or a Dish

Quick answer: Two kinds of restaurant ad: the room, or the food. Lead with a photo of the dining room or the shopfront and pick A day in the life, and the ad is people enjoying your place — your actual room, with the customers in your photo removed and a generated cast in their seats. Lead with a photo of the dish and pick Mouth-watering close-ups, and the ad is steam, pour and cut. Add the other kind as a second photo. Fifteen seconds is enough for one dish; thirty for the room and a dish. The end card carries the name from your sign.

Which photos work and the four pickers cover the general rules; this page is the restaurant case.

## Which photo first

The first photo is what the ad is about. The room or the shopfront first when the draw is the place: the reading files it as a store / property, keeps your actual interior or sign, removes the people, and sets the ad inside it. The dish first when the draw is the food: it is cleaned up into a product shot and the ad is built around it. Add the other as the second photo and both appear.

## Taking the photos

Room: stand at the door and shoot the whole space in daylight; include the sign if the name matters. Dish: fill the frame, top-down or three-quarter, on the table you actually serve it on. Customers and staff in the room photo are fine — the composer says People will be removed. A menu photo is not a good first image; it reads as text.

## Which vibe

Mouth-watering close-ups for the food: heat, pour, the first cut, the bite. A day in the life for the place: a couple at a window table, a family, the evening rush. Warm for a family restaurant; Minimal for a fine-dining plate; Funny for a street-food stall with a personality.

## Length and format

Fifteen seconds: one dish, one idea. Thirty: the room, then the food. Sixty: a small story — arrival, the meal, the goodbye. Portrait for feeds and stories; landscape for a website or a screen in the window.

## The one line

Type what it is and who it is for: our family-run Sichuan restaurant, for people who like it spicy. It tells the reading this is a place, not a product, and gives the script its angle. Skip health claims and superlatives; they are removed anyway.

## The end card

The name on your sign is read and printed as the brand; the product name is what you confirmed in the dialog. Add a tagline and a call to action — Book a table, Order on the app. The name is delivered here, not on the shots. The end card.

## What to expect

The dishes look like your dishes; the room looks like your room; the diners are a generated cast that fits the place. Lines are spoken in the dialogue language shown in the composer. If a shot is wrong, open it in the editor, change the text, regenerate that shot, and rebuild the final cut. The progress page.

| You are selling | First photo | Vibe | Length |
| A signature dish | The dish, filling the frame | Mouth-watering close-ups | 15s |
| The place | The room or the shopfront | A day in the life / Warm | 30s |
| Both | The room, then the dish | A day in the life | 30–60s |
| A café’s drink | The drink, poured | Mouth-watering close-ups / Minimal | 15s |

## FAQ

### My restaurant photo has customers in it.

Fine. People are removed and the room is kept; the ad's own cast sits in it.

### Will the ad show my actual dining room?

Yes, when the room is your first photo. SceneMixer draws it from your photo, without the people.

### Will the dish look like my dish?

Yes. SceneMixer cleans a dish photo into a product shot, and the ad is built around it.

### Does the restaurant's name appear in the shots?

Not reliably; generated video does not write legible text. The name is delivered on the end card.

### Can I say the food is the best in town?

SceneMixer removes superlatives and health claims from ad scripts. Say what it is and who it is for.

Updated November 3, 2026


---

# Make an AI Ad for a Store, a Salon, a Gym or a Property | SceneMixer

URL: https://scenemixer.com/guides/ai-ad-for-a-store-or-property/
Shops, salons, gyms, hotels, homes for sale: how a place photo is read, why it is kept rather than redrawn, what happens to the customers in it, which vibe shows it off.

# An AI Ad for a Store or a Property

Quick answer: Add a photo of the place — the shopfront with its sign, the room, the lobby, the kitchen of the house — and the reading files it as a store / property. A place is never cut out or redrawn: the sign, the depth and the scale would go. Your photo is kept, the people in it are removed, and the ad is set inside it with a generated cast: shoppers, a stylist and a client, a family walking through the house. Lead with A day in the life, thirty seconds, and let the end card carry the name from the sign and Visit us.

Which photos work covers the general rules; this page is the place case.

## What counts as a place

Anything the customer walks into: a shop, a café (see restaurants), a salon, a gym, a clinic, a hotel, an office, a house or a flat for sale or rent. The reading files it as store / property. If the photo is ambiguous — a product on a shelf inside your shop — the one line settles it: write our shop and it is a place, not a product.

## The photos

First: the frontage with the sign, straight on, in daylight; or, for a property, the best room. Then up to three more: the interior from the door, a second room, a detail. Customers, staff, the family who lives there: all fine — the composer says People will be removed, and they are, while the room stays exactly as it is.

## Why the place is kept

Products are cleaned onto a white background so the ad can put them anywhere. A place cannot be treated that way: cut out, a shopfront loses its sign, a room loses its depth and scale. So the photo is kept and used as the first scene the ad happens in; the ad’s people are generated into it. That is why the redraw button is not offered for the main image.

## What the ad shows

Your place, with life in it: the door opening, someone browsing, a haircut in the chair, a couple standing in the living room deciding. Extra photos become the rooms the shots move through. The second-photo rule: for places, every photo you add becomes a set, so a shop, its interior and its fitting room can all appear.

## Which vibe

A day in the life: people using the place. Warm: a neighbourhood shop, a family salon. Minimal: a design store, a high-end property. Show how it works: a gym class, a service with steps. Epic: a hotel with a view. Thirty seconds for a place; sixty for a property tour with more rooms.

## The one line

A family barbershop on the corner of Elm Street, walk-ins welcome; a two-bedroom flat by the park, ready to move in. The name and the offer go here so the end card has them. Keep it to what is visibly true; superlatives are removed.

## The end card

The sign is read and printed as the brand. If the brand and the product name are the same — a shop named after itself — one line is printed. Add a call to action: Visit us, Book an appointment, Arrange a viewing. The name reaches the viewer here, not on the shots. The end card.

| Place | First photo | Vibe | Length |
| A shop | The frontage with the sign | A day in the life / Warm | 30s |
| A salon or clinic | The room from the door | Show how it works / Warm | 30s |
| A gym | The floor in use | Show how it works / Epic | 30s |
| A hotel | The lobby or the view | Epic / Minimal | 30–60s |
| A home for sale | The best room | A day in the life / Minimal | 60s |

## FAQ

### Will the ad show my actual shop?

Yes. SceneMixer keeps a place photo and uses it as the set; only the people are removed.

### There are customers in my photo.

Fine. People are removed and the room stays; the ad's own cast is generated into it.

### Why can't I redraw the shop photo?

Cutting a place out would lose its sign, depth and scale. Places are kept as shot.

### Can I add several rooms?

Yes. For places on SceneMixer, every photo you add becomes a set the shots can move through, up to four.

### Will the shop's name be visible in the ad?

On the end card, printed from the sign. Generated video does not write legible text, so the shots are not where the name is delivered.

Updated November 3, 2026


---

# Make an AI Ad for an App or a Website from a Screenshot | SceneMixer

URL: https://scenemixer.com/guides/ai-ad-for-an-app-or-website/
How to make an ad from a screenshot: which screen to use, why it is never redrawn, why it must be one clean screen with no faces, and what the ad can show.

# An AI Ad for an App or a Website

Quick answer: Add a screenshot of the screen that matters — one clean screen, roughly phone- or computer-shaped, with no real faces — and the reading files it as an app or a website. The screenshot is used exactly as uploaded, never redrawn, so the interface stays readable. The ad is the moment of use: a generated person in a real situation reaching for the app and getting the result. Pick Show how it works, fifteen or thirty seconds, and let the end card carry the name and Download the app.

Which photos work has the screenshot rules; this page is the app case.

## Which screenshot

The screen that shows the promise: the result page, the booking confirmed, the map with the route. Not the sign-in page, not the settings. One screen only: a full-page capture is far taller than any video frame and is refused (This screenshot is too tall or too wide — crop it). Crop to the phone or the browser window and upload that.

## No faces

A screenshot with a real face — a profile photo, a video call — is refused: This screenshot has a real face in it. Please use one that shows only the interface. The reason is in the hint: editing a face out of a screenshot smears the interface with it, so a clean one is required. Use a screen without avatars, or a demo account.

## Never redrawn

Products and places are cleaned into product shots; screenshots are not. Yours is kept pixel for pixel and used as the featured product, so the buttons, labels and layout stay exactly as they are. That is also why the redraw button is not offered for it (App and website screenshots are used as they are). If the screenshot is wrong, pick different images.

## What the ad shows

The moment of use. A generated person in a real situation — on a train, in a kitchen, at a desk — reaches for the app and something changes: the taxi arrives, the meal is ordered, the graph goes up. The screenshot appears as the thing on their screen. The ad does not fake a walkthrough of the interface; generated video cannot draw menus and text reliably, so the script stays with what the app does for the person, and the screen itself is your real screenshot.

## Which vibe and length

Show how it works: problem, tap, result. It is built for this. Minimal for a design-led product; Funny for a consumer app with a personality; Twist when the app solves something unexpected. Fifteen seconds for one promise, thirty for problem and solution. Portrait for app-store previews and feeds.

## The one line

A grocery app that delivers in twenty minutes, for busy parents. Who it is for shapes the person in the ad; what it does shapes the moment. Keep claims to what is visibly true; superlatives are removed.

## The end card

Brand and product name from the reading and your line; the call to action is the download or the sign-up: Get the app, Try it free. The main image on the card is your screenshot. The end card.

| Do | Do not |
| One clean screen, phone- or browser-shaped | A full-page capture |
| The result screen, the promise | The sign-in page |
| A demo account without avatars | A screen with real faces |
| Show how it works, 15–30s, portrait | A 60-second walkthrough of menus |

## FAQ

### Will the ad redraw my interface?

No. SceneMixer uses screenshots exactly as uploaded, never redrawn, so the interface stays readable.

### Why was my screenshot refused?

SceneMixer refuses a screenshot if it has a real face in it, or if it is a full-page capture. Crop to one clean screen.

### Can the ad show someone tapping through several screens?

It shows the moment of use and the result, with your screenshot as the screen. It does not fake a multi-screen walkthrough.

### Which vibe for an app?

Show how it works: problem, tap, result.

### Can I add more than one screenshot?

Yes, up to four images on SceneMixer. The first is the featured screen; others become supporting details.

Updated November 3, 2026


---

# Make an AI Ad for Clothing, Shoes or Accessories | SceneMixer

URL: https://scenemixer.com/guides/ai-ad-for-clothing/
How to photograph a garment so it is read right, what happens when a model is in the photo, how the ad dresses generated people in it, and which vibe sells clothes.

# An AI Ad for Clothing

Quick answer: Photograph the garment flat or on a mannequin, front on, in even light, and the reading files it as apparel. It is cleaned into a product shot — the cut, the colour, the fabric — and the ad dresses a generated cast in it: someone walking out of the door, turning in a mirror, sitting at a café table. A photo with a real model in it is fine for a shop, but for clothing the piece reads better alone. Pick A day in the life or Minimal, portrait, thirty seconds, and let the end card carry the brand and Shop now.

Which photos work covers the general rules; this page is the clothing case.

## The photo

The piece alone: flat-lay on a plain surface, or on a mannequin or hanger, front on, in even light so the true colour shows. Fill the frame. As extra images: the back, a fabric close-up, the piece worn (a real model is fine here — people are removed and the garment is what is read). Shoes: three-quarter, the pair. Accessories: on a plain surface, large in frame.

## What the reading does

The garment is cleaned into a product shot on a clean background: the cut, the colour, the fabric, the details kept. That shot is the featured product every shot of the ad dresses its people in. If the colour came out wrong, it is usually the photo’s light; reshoot in daylight and pick different images.

## Who wears it

The ad’s cast is generated to fit the piece and the audience in your line: a linen shirt for men who hate ironing gets a man in a hurry; a school-run parka gets a parent at the gate. You do not supply a model. If the piece is for a specific body or age, say so in the line.

## What the ad shows

The piece in motion: a walk, a turn, a sit-down, a coat going on. Fabric moving is what sells clothes on video, so the vibes that give the cast something to do work best. Close-ups of a seam or a print come from your extra photos.

## Which vibe

A day in the life: the piece worn through a morning. Minimal: the garment, light, one line — for a brand with a look. Twist: it looks formal, then it is not. Funny: a hoodie with a personality. Portrait for feeds; thirty seconds for one piece, sixty for a small collection with several photos.

## The one line

A wool overcoat for city winters, for people who walk to work. Material, occasion, wearer. Sizes and prices belong on the end card or the landing page, not in the shots; SceneMixer removes claims like “best quality”.

## The end card

The brand from the label in your photo or your line; the product name from the reading; a tagline; and Shop now or See the collection. The main image on the card is the cleaned product shot. The end card.

| Piece | First photo | Extra photos | Vibe |
| A coat or a dress | On a mannequin, front on | Back, fabric close-up | A day in the life |
| A T-shirt or a hoodie | Flat-lay | The print up close | Funny / Minimal |
| Shoes | The pair, three-quarter | Sole, detail | Minimal / Show how it works |
| A bag or jewellery | On a plain surface, large | Worn, detail | Minimal / A day in the life |

## FAQ

### Do I need a model in my photo?

No. SceneMixer generates the cast. Photograph the piece alone so the cut and colour are read right.

### There is a model in my photo already.

Fine. People are removed and the garment is what is read; the ad dresses its own cast in it.

### Will the colour be right?

If the photo is lit evenly, yes. A yellow-lit photo reads as a yellow-tinted garment; reshoot in daylight.

### Can I show a whole collection?

Add up to four pieces to SceneMixer as separate photos; the first is featured and the others appear. Sixty seconds gives them room.

### Can the ad show sizes and prices?

Put them on the end card or your landing page. Generated shots do not write legible text.

Updated November 3, 2026


---

# An AI Ad for Jewelry: Sparkle, Scale and the Worn Shot

URL: https://scenemixer.com/guides/ai-ad-for-jewelry/
Make a jewelry ad with AI from a photo: the piece redrawn as a studio shot, worn shots and close-ups from extra photos, Minimal or Warm tone, no price in the picture.

# An AI ad for jewelry

Quick answer: Photograph the piece alone on a plain surface, large in the frame, in soft even light, and the reading files it as jewelry: it is cleaned into a product shot with the stones, the metal and the clasp kept, and the ad puts it on a generated wearer at the right scale: a ring on a hand, a necklace at a collarbone, earrings in a turn of the head. Add a worn photo and a macro of the setting as extra images. Pick Minimal for a brand with a look, Warm for a gift, Epic for a hero piece; thirty seconds in portrait. Carat, metal and price belong on the end card and the landing page, not in the shots. Which photos work covers the general rules; this page is the jewelry case.

One line for the box next to the upload: A recycled-gold signet ring with a hand-cut onyx, made in our workshop; for people who want one ring they never take off.

## The photo

- The piece alone, on a plain matte surface (linen, stone, paper), large in frame, soft light from a window or a tent. Hard flash makes a stone look like a white dot.
- Extra photos: the piece worn (a real hand or neck is fine; people are removed and the piece is what is read), a macro of the setting or clasp, the box.
- Scale: if the piece is unusually large or small, say so in the line; a photo alone does not tell a model how big a pendant is.
Jewelry is redrawn as a clean studio shot on white, the same in every scene, so the gold is one gold throughout. If the colour came out wrong, it is usually the light: reshoot in daylight. An ad for a physical product.

## Who wears it

You do not supply a model. The cast is generated to fit the piece and the buyer in your line: a signet ring for men who never wore one gets a man adjusting his cuff; a graduation pendant gets a mother fastening a clasp. Real people in your photos are removed, so a brand ambassador's face does not survive the read: this is a feature, since it is a face you may not have permission to put in an ad. Real faces.

## What the ad shows

Jewelry on video is light moving over a surface. The vibes that give the wearer something to do work best: a hand reaching for a cup, a turn to look back, a clasp fastened in a mirror. Close-ups of the stone come from your macro. Avoid asking for a hundred stones in one shot; one piece, one hand, one light.

| Tone | Jewelry use |
| Minimal | The piece, one light, silence, one line; for a brand with a look. |
| Warm | A gift opened, a hand held; anniversaries, graduations. |
| Epic | A hero piece revealed; a launch. |
| A day in the life | Everyday pieces worn through a morning; stacking rings, small hoops. |
| Twist | It looks like a drop of water, then it is the stone. |

## The one line, and what gets removed

Material, maker, wearer: recycled 18k gold, hand-set in our Lisbon workshop, for people who wear one ring. Claims SceneMixer cannot verify (flawless, best quality, investment grade) are removed from the script. Carat weights, sizes and prices belong on the end card or the landing page, not in the shots, because text in the picture is not drawn reliably. The end card.

## The end card

The brand from the box or your line; the product name from the reading; a tagline; and Shop now or Book a fitting. The main image on the card is the cleaned product shot, so the last thing seen is the piece itself. For a collection, make the 60-second cut with several photos, one piece per segment. Length, format and quality.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Piece | First photo | Extra photos | Tone |
| A ring | On a plain surface, large in frame | On a hand; macro of the stone | Minimal / Warm |
| A necklace or pendant | Laid out, clasp visible | Worn at a collarbone; the box | Warm / A day in the life |
| Earrings | The pair, together | Worn, in profile | Minimal |
| A watch or bracelet | Three-quarter, on a surface | On a wrist; the clasp | Epic / Minimal |

## FAQ

### Do I need a model for a jewelry ad?

No. The wearer is generated to fit the piece and the buyer in your line. A worn photo helps the read; the person in it is removed.

### Will the stone look right?

SceneMixer redraws the piece as a studio shot from your photo; soft, even light in the photo is what makes a stone read as a stone.

### Can I show the price or carat weight?

On the end card and in the spoken line, not in the picture; text is not drawn reliably.

### Which tone suits jewelry?

Minimal for brand pieces, Warm for gifts, Epic for a hero launch, A day in the life for everyday pieces.

### How long should it be?

30 seconds for one piece; 60 for a small collection with several photos, one piece per segment.

### Can I say it is real gold?

Say the material plainly in the line; superlatives like best or flawless are removed from the script.

### Portrait or landscape?

Portrait for feeds and stories, landscape for a website or an in-store screen. Changing it later is a new ad.

### What does it cost?

One figure is quoted on SceneMixer before it starts and it is a ceiling; length and quality set it.

Updated November 3, 2026


---

# AI Perfume Ad: A Fragrance Spot From One Bottle Photo

URL: https://scenemixer.com/guides/ai-ad-for-perfume/
Make a perfume ad with AI from one bottle photo: sell a scent nobody can see with a mist, a turn of the head and a place, then edit the end card for each gift date.

# An AI perfume ad

Quick answer: A scent cannot be filmed, so a perfume ad films what the scent does: the mist in the light, a turn of the head, a doorway, a memory of a place. Photograph the bottle alone, front on, label readable, in soft side light, and the reading files it as a product and redraws it as a clean studio shot with its shape, colour and label kept. The ad then puts that bottle in the hands of a generated wearer in a place that matches the notes you wrote in the one line. Pick Minimal for a niche brand, Twist or Epic for a launch, Warm for a gift; fifteen or thirty seconds, portrait. Size, concentration and price go on the end card or the landing page, not in the picture.

One line for the box next to the upload: A 50 ml eau de parfum in a square glass bottle with a brass cap; top notes of bergamot, a heart of fig leaf, a base of cedar; for people who want a fragrance that smells like a morning terrace, not a perfume counter.

## The photo: glass is the hard part

- The bottle alone, on a plain matte surface (linen, stone, paper), large in the frame, cap on. Glass shows everything around it, so keep windows, phones and your own reflection out of it.
- Soft light from the side, not a hard flash: flash turns a faceted bottle into a white blur and a gold cap into a mirror.
- The label readable. The brand name is copied letter for letter from what is printed, so a tiny or curved label is better backed up by typing the brand and the product name in the one line, which wins over the photo.
- Up to three extra photos of the same bottle: the box, the atomiser close up, the bottle next to a cut fig or a sprig of cedar. Extras become details and sets the shots move through.
The bottle is redrawn as a studio shot on a clean background and then used as the same object in every scene, so the gold of the cap is one gold throughout. If the colour came out wrong it is usually the light; reshoot in daylight and pick different images. Which photos work · An ad for a physical product.

## Selling a smell on a screen

The ad cannot make a viewer smell anything and will not claim to. What it can do is put the bottle next to the things people associate with a scent: the spray in a shaft of light, a wrist, a turn in a doorway, the place the notes point to. So write the notes as places and moments in the one line, and the script has something to film:

| Note family | A scene to write | Tone that suits it |
| Citrus and green | A morning terrace, a bowl of lemons, linen drying | A day in the life, Minimal |
| Floral | A garden doorstep after rain, a bouquet on a table | Warm |
| Woody and amber | A library at dusk, a fire, a coat on a hook | Minimal, Epic |
| Gourmand (vanilla, tonka) | A kitchen at night, a warm window, a shared cup | Warm |
| Fresh and aquatic | A shoreline at dawn, wet stone, wind | Epic, Minimal |
| Musky and clean | A bare bright room, white linen, an open window | Minimal, Twist |
These are scenes for the picture, not claims about the product. The ad's script is written from your line and the photo, and sets the product in the place; the cast is generated to fit the buyer you describe. Writing the one-line brief.

## Which tone for which kind of bottle

| Tone | Perfume use |
| Minimal | The bottle, one light, near silence, one line; for niche and design-led brands. |
| Epic | A launch: scale, a skyline or a coast, the reveal of the bottle at the end. |
| Twist | It looks like one thing (rain on a window), then it is the mist; ends on the bottle. |
| Warm | A gift opened, a hand on a shoulder; for gift sets and the weeks before a gift date. |
| A day in the life | An everyday scent worn through a morning: a spray at the mirror, a bus, a desk. |
| Show how it works | Rarely: the spray and the first walk-out; only for travel sprays or refills. |
Length and quality move the price; tone and format do not. Portrait for feeds and stories, landscape for a website or a counter screen. Fifteen seconds is one beat: the bottle and one idea. Length, format and quality.

## What is removed, and what goes on the end card

Wording that cannot be backed up is removed from the script: the best, irresistible, number one, and promises about how people will react to the wearer. That is a filter on wording, not advice; the claims you make are yours. Because video cannot draw readable lettering, the ad is made without on-screen text. Brand, product name, a tagline and the call to action are delivered on the end card: Shop now, Find your scent, Gift this season. Volume, concentration and price belong on the card or the landing page. The end card.

## One bottle, four gift dates

Perfume is bought for dates. Shoot the ad once and change only the end card: Edit end card, then Rebuild final cut re-joins the same shots without reshooting. Dates in the United States:

| Date | Occasion | End card tagline and call to action |
| Friday, December 25, 2026 | Christmas | Gift this season / Shop the gift set |
| Sunday, February 14, 2027 | Valentine's Day | For the one who stays / Order by the 10th |
| Sunday, May 9, 2027 | Mother's Day | A scent she will keep / Shop gifts for mum |
| Sunday, June 20, 2027 | Father's Day | Woody, warm, his / Shop now |
Make the 30-second main ad early and a 15-second teaser from a second photo (the box) for the week before. A Christmas ad · A Black Friday ad · Rebuild the final cut.

## Shoes, bags, eyewear: the rest of the counter

Fashion accessories use the same recipe as the bottle: the item alone, the cleaned studio shot, a generated person who wears it in a real place. The clothing page covers garments; here is the rest.

| Item | First photo | Extra photos | What the ad does | Tone |
| Shoes | The pair, three-quarter view, on a plain surface | The sole; a detail of the stitching | Walks out of a door onto wet pavement, a sit-down, laces tied | A day in the life, Minimal |
| A handbag | Front on, large in frame, handles up | The inside; the clasp | A bag taken from a shoulder, opened at a café table | Minimal, Warm |
| Sunglasses | The pair open, front on | Folded; the hinge | A head turn in bright sun, a hand lifting them | Minimal, Epic |
| A watch or scarf | Three-quarter, on a surface | Worn; the clasp or the weave | A cuff pulled back, a scarf tied at a doorway | Minimal, Warm |
A “fashion video” here is one product in an ad of 15 to 60 seconds, or a drama in which outfits are set per character and per scene. An ad for clothing · An ad for jewelry · Outfit consistency in a series.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Part of the ad | What to give it | Where the scent lives |
| First photo | The bottle alone, label readable, soft side light | In the bottle shot |
| The line | Notes written as places, the buyer, and the brand name | In the script's setting |
| Tone | Minimal, Twist, Epic or Warm | In the mood |
| End card | Brand, name, tagline, call to action | In the one place text is readable |

## FAQ

### How do I make a perfume ad with AI?

Upload one to four photos of the bottle, write one line with the notes as places and the buyer, pick a tone, a length and portrait or landscape, and SceneMixer writes, shoots and finishes the spot with an end card.

### Can an AI ad show what a perfume smells like?

No. It shows the bottle, the mist and the places and moments the notes suggest. Describe the scent in the line, not as a promise.

### What photo works best for a perfume bottle?

The bottle alone on a plain matte surface, large in frame, label readable, in soft side light without reflections of windows or people.

### Will the bottle look the same in every shot?

Yes. It is redrawn once as a studio shot and used as the same object throughout; the label and the colours of your photo are kept.

### Can I put the price or the size on screen?

Not in the picture; text there is not drawn reliably. Use the end card and the landing page.

### Can I reuse the ad for Valentine's Day and Mother's Day?

Yes. Edit the end card and rebuild the final cut; the same shots are re-joined with a new tagline and call to action.

### Can I make an ad for shoes or sunglasses the same way?

Yes. The item alone is cleaned into a studio shot and a generated person wears it. The clothing and jewelry pages cover the details.

### Does it use a real model or my brand ambassador?

No. People in your photos are removed and the cast is generated; SceneMixer offers no face swap and no voice cloning.

### What does it cost?

One figure is quoted on SceneMixer before it starts and it is a ceiling; length and quality set it, tone and format do not. A 15-second ad is the cheapest length.

Updated November 3, 2026


---

# An AI Ad for Skincare and Beauty: Texture, Routine, No False Claims

URL: https://scenemixer.com/guides/ai-ad-for-skincare-and-beauty/
Make a skincare or beauty ad with AI from a product photo: texture and routine on a generated face, Show how it works or A day in the life, claims kept honest.

# An AI ad for skincare and beauty

Quick answer: Photograph the bottle, jar or tube front on, label readable, in even light, and the reading files it as beauty: the product is cleaned into a studio shot and the ad shows it in use on a generated face or hand: a pump, a drop on a fingertip, a routine at a bathroom mirror, a morning in daylight. Pick Show how it works for a routine, A day in the life for a mood, Minimal for a brand with a look. Say what the product is and who it is for; before-and-after promises and medical words are removed from the script, because the ad can show a texture and a routine but cannot prove a result. Which photos work covers the general rules; this page is the skincare and beauty case.

One line for the box next to the upload: A fragrance-free night cream with squalane, in a glass jar; for people with dry skin who want one step before bed.

## The photo

- The product alone, front on, label readable, in even light. The brand name is transcribed from the label exactly as printed, so make it legible.
- Extra photos: the texture (a swatch on paper or the back of a hand), the open jar, the box. A photo with a real face in it is fine for a brand shot, but the face is removed; the product is what is read.
- Sets: up to four photos of one product. A routine of three products is three ads, or one 60-second ad with one product per segment and a line that names them in order.
An ad for a physical product.

## Whose face

Beauty is sold on faces, and the faces here are generated to fit the buyer in your line: dry skin in winter gets a woman at a frosted window, a men's moisturiser gets a man shaving at six. You do not supply a model, and a real face in your photo does not survive the read. If the product is for a specific skin tone, age or hair type, say so in the line and the cast follows. Describing a face.

## What the ad shows

Texture and routine, not results. A pump pressed, a drop spreading on a fingertip, a swatch pulled across the back of a hand, the jar set down on a shelf, a face in morning light. Close-ups of the texture come from your swatch photo. Avoid asking for a split-screen before and after: it cannot be shown honestly and the claim would be removed anyway.

| Tone | Skincare and beauty use |
| Show how it works | Pump, apply, the routine in order; serums, cleansers, tools. |
| A day in the life | The product inside a morning or a night; moisturisers, body care. |
| Minimal | The jar, the light, one line; premium brands. |
| Warm | A gift set, a shared bathroom shelf. |
| Mouth-watering close-ups | Textures, drips and swirls shot like food; balms, butters, glosses. |

## Claims that are removed, and what to say instead

| Do not write | Write instead |
| Clinically proven, dermatologist recommended | What is in it and who it is for: with 2% niacinamide, for oily skin |
| Removes wrinkles in 7 days | What it does when used: sinks in without a film; for under make-up |
| Cures acne, treats eczema | Nothing medical. If the product is a medicine, this is not the tool. |
| Best, number one, flawless | The texture, the scent, the routine |

The script writer removes superlatives and unverifiable claims; medical claims are a content-rule matter and a shot built on one is blocked. Content rules · If the ad is blocked.

## The end card

The brand from the label; the product name from the reading; a tagline; Shop now or Find your shade. The main image is the cleaned product shot. Ingredient lists, sizes and prices go on the landing page, not in the picture. The end card.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Product | First photo | Extra photos | Tone |
| A serum or cream | The bottle or jar, front on, label readable | A swatch; the open jar | Show how it works |
| A lipstick or gloss | The tube, cap off | The swatch; the shade range | Mouth-watering close-ups / Minimal |
| A tool (brush, device) | Three-quarter, on a surface | In use on a hand | Show how it works |
| A gift set | The box open | Each product; the box closed | Warm |

## FAQ

### Can I make a skincare ad with AI?

Yes. Upload the product photo, write one line about what it is and who it is for, pick a tone, and the ad shows the texture and the routine on a generated face or hand.

### Can it show a before and after?

No. SceneMixer removes result claims from the script and blocks medical claims; the ad shows texture and routine instead.

### Do I need a model?

No. The face is generated to fit the buyer in your line; a real face in your photo is removed.

### Can I say dermatologist tested?

Unverifiable claims are removed. Say what is in the product and who it is for.

### Which tone suits beauty?

Show how it works for routines and tools, A day in the life for moisturisers, Minimal for premium, close-ups for textures.

### Can I advertise a medicine or a treatment?

No. Medical products and claims are outside the content rules.

### Will the brand name be right?

SceneMixer transcribes it from the label exactly as printed; make the label legible in the first photo.

### How long should it be?

15 seconds for one product and one idea, 30 for a routine, 60 for a set with one product per segment.

Updated November 3, 2026


---

# AI Cinematic Realism: The Settings That Make It Look Filmed

URL: https://scenemixer.com/guides/ai-cinematic-realism/
Realistic and cinematic are different dials. Which SceneMixer settings turn each: the four live-action looks, Film type, cuts, long takes and three settings sheets.

# AI Cinematic Realism: Two Dials, and Where Each One Lives

Quick answer: “Realistic” and “cinematic” are two dials. Realistic means the picture looks photographed, and it is decided by the Visual Style row in step 1 and by how concretely you describe people and places. Cinematic means it looks directed, and it is decided by the Type, the shape, how the shots are cut and how long each take runs. Of the 23 finished samples on this site, 20 chose a live-action look, 16 of them Digital HD and 4 Shot on Film. This guide puts each setting on the dial it turns, and ends with three settings sheets you can copy.

## Two dials, not one

| Dial | What it asks of the picture | Where you set it | What moves it |
| Realistic | Looks photographed: real skin and cloth, believable rooms | Step 1, Visual Style; step 2, the descriptions | The look you pick; how concrete the descriptions are; how many people are in a shot |
| Cinematic | Looks directed: paced, framed, cut with intent | Step 1, Type and shape; step 3, the storyboard | Film or Short Drama; 16:9 or 9:16; the shot sizes and seconds; 15 or 30 second segments; the quality you finish on |
The two can be set apart. A clean, modern, true-to-life picture cut like a thirty-second feed clip is realistic and not very cinematic. A grainy, dreamy picture with one long conversation take is cinematic, and less “real”. Decide which one you are short of before you change anything.

## The realism dial: four live-action looks

SceneMixer’s Visual Style row groups its cards by medium. The first group is Live-action, with four looks. The row is fixed once you continue past step 1, because it shapes every sheet and every shot.

| Look | The card says | Used by | Reach for it when |
| Digital HD | Crisp and true-to-life, like today’s streaming dramas | 16 of the 23 samples | A modern story that should look like something shot yesterday |
| Shot on Film | Film grain, warm colour and soft blacks | 4 of the 23 | A period, coastal or family story that wants warmth and texture |
| Soft Glow | Hazy highlights, low contrast, pale and dreamy | None of the samples | Romance or memory, where softness is the point |
| Black & White Film | Rich blacks, whites and grays, like a classic film | None of the samples | Noir, or any story that should read as a different era |
Two neighbours sit in the 3D group and are computer graphics, not live action: Game Engine Realism (sharp materials, no grain), which one sample uses, and Realistic 3D (near-human 3D characters, like film CG). If you want to look photographed, stay in the live-action group.

A note on names. The sample pages still print the older names, Realistic HD for Digital HD and Filmic Live-Action for Shot on Film; they were renamed on October 17, 2026. Visual style presets, one by one has all of them. The four samples that chose Shot on Film have episode titles about a blade, a pilot at sea, an olive table and a bird that was never meant to sail.

## The realism dial: words in step 2

The look sets the medium. Whether a face reads as a person is decided by the description in step 2. The description is the sheet: SceneMixer’s reading writes it, you can edit it, and every shot copies it. The habits that move realism are covered in How to make AI video look more realistic: specific, slightly imperfect faces, physical causes for actions, two or three people to a shot. Words such as “8K” and “masterpiece” are not on that list because they give the model nothing to place in the frame. Why those words do not help.

One piece of realism is deliberately not in the sheets. Character, place and prop sheets are plain: figures on white, rooms empty, objects alone. Whatever mood a sheet carries, every shot copies, so the film look is written per shot in the storyboard, not painted into the reference. Why sheets look plain.

## The cinematic dial: Type, shape and takes

| Setting | Cinematic choice | What it changes |
| Type in step 1 | Film | Slow-paced and atmospheric with no episode cap, and a script treated as one piece divided into acts instead of a run of hooks. Short Drama episodes stop at 240 seconds; Film does not |
| Shape in step 1 | 16:9 Landscape | The recommended shape for film. The 9:16 portrait shape is for feeds. 22 of the 23 samples are 16:9 |
| Segment length in step 3 | Long take 30s | One generation holds a longer stretch with the camera cutting inside it, so a conversation is not interrupted by a generation boundary. Only on tiers that support 30 seconds |
| Shot seconds in step 3 | Give movement the time it needs | Seconds are written per shot and add up to the segment. A shot given too few seconds reads as a flicker. Shot seconds and timing |
| Quality per segment | Draft low, finish high | A cheap tier for drafts, a better one for the keepers, then Enhance a finished clip a step at a time, up to 2K. Enhance a clip |
How the shots are cut. A SceneMixer storyboard is a run of shots, each with a size and an angle, and the writing step is built so that two shots of the same people back to back do not repeat the same framing: when it cuts back to the same people, the size changes by at least a step or the angle changes. That is the grammar of a scene cut by someone, and you can still edit any shot’s text for free. How to write one shot.

Where a take ends. A segment is one generation of up to 15 seconds, or 30 on the long-take tiers, and one continuous shot cannot be longer than one segment. A long unbroken take is the one cinematic thing this method cannot give. Everything else is joined. 15 or 30 second segments.

## Three settings sheets to copy

These are starting points, built from preset names that exist in SceneMixer’s step 1. Change anything after the story is read by editing the descriptions in step 2; the scene and costume styles are written into those descriptions at the read.

| Setting | Corporate thriller | Coastal family drama | Night-street noir |
| Type | Film | Film | Film |
| Shape | 16:9 Landscape | 16:9 Landscape | 16:9 Landscape |
| Visual Style | Digital HD | Shot on Film | Black & White Film |
| Scene Style | Corporate Cold | Mediterranean Coast | Retro Hong Kong |
| Costume Style | Power Tailoring | Soft Everyday | Retro Hong Kong |
| Segment length | Standard 15s | Long take 30s for the dinner scene | Standard 15s |
| Draft on | The cheapest tier | The cheapest tier | The cheapest tier |
| Finish on | A better tier for the keepers, then Enhance | A better tier for the keepers | A better tier for the keepers |
The pairs follow the rule on the scene and costume presets page: a scene style sets what the sets are made of, a costume style what people wear, and they should come from the same world.

## A one-evening test

- Pick the look and the Type, then read a one-page script. The first 15 setup images on a SceneMixer account are free, so the cast and the places cost nothing to look at.
- Open the storyboard of episode 1, which is free on the first drama, and read the shot sizes. If two shots in a row feel the same, change one.
- Use the free preview: one shot of about 5 seconds with your own cast, on the cheapest tier.
- Generate one segment as a draft. Judge the face and the framing, not the sharpness; sharpness is what the finishing tier is for.
- Only then render the rest. Draft at the lowest quality, finish at the highest.

## FAQ

### What is the difference between realistic and cinematic?

In SceneMixer, realistic means the picture looks photographed, which the Visual Style row and the descriptions decide. Cinematic means it looks directed, which the Type, the shape, the cuts and the takes decide.

### Which visual style looks most real?

SceneMixer’s Digital HD is the crisp, true-to-life live-action look and the one most of the finished samples use. Shot on Film adds grain, warm colour and soft blacks for a cinema feel.

### Is Realistic 3D the same as live action?

No. In SceneMixer, Realistic 3D and Game Engine Realism are computer graphics. They sit in the 3D group, and the live-action group has four looks.

### Does choosing Film type make it more cinematic?

SceneMixer changes the pacing and the cut: slower, atmospheric, no episode cap, a script treated as acts. It is one of several dials, not a filter.

### Can I change the look after I start?

No. SceneMixer fixes the Visual Style once you continue past step 1, because it shapes every sheet and shot. To change it, start a new project with the same script.

### Do words like 8K or hyperrealistic help?

No. They give the model nothing it can place in the frame. Concrete descriptions of faces, places and actions do.

### Can I get a long unbroken take?

A single SceneMixer take stops at one segment, 15 seconds or 30 on the long-take tiers. Longer scenes are segments joined.

### Why are my sheets so plain?

On purpose. Every SceneMixer shot copies its sheets, mood included, so the film look is written into the shots instead.

Updated November 2, 2026


---

# AI Clip Generator: The Smallest Clips You Can Make Here

URL: https://scenemixer.com/guides/ai-clip-generator/
The smallest clips SceneMixer makes, from a free five second preview to a 15 second ad, what each costs, and what it will not do: cut clips from a long video.

# AI Clip Generator: How Small a Clip Can You Make?

Quick answer: SceneMixer makes clips by shooting them, not by cutting them out of something longer. The smallest things you can get are a free preview of about 5 seconds, a segment that is billed from 4 seconds up to 15 (or 30 on the long-take tiers), and a 15 second ad built from your photos. Anything longer is those pieces joined. It does not pull highlights from a long video, and there is no box where a prompt returns a clip: a clip starts from a short script. This guide lays the sizes out as a ladder, says what each rung costs, and ends with what to do when a clip is a few seconds too long.

## Which kind of clip do you want?

“Clip” means three different things, and only two of them are things this site does.

| You want | Can you do it here? | Where |
| A new short clip from an idea | Yes, by writing a very short script | A one-scene script, then one segment. The ladder below |
| A short branded spot from product photos | Yes | Ad Video, 15, 30 or 60 seconds. An ad from one photo |
| A few seconds taken out of a clip you generated | Yes | Edit this clip, trim or remove. Trim a clip |
| Highlights or short clips cut out of a long video | No | The remake path reads a video and shoots it again with your cast. It does not export pieces of the original |
| A single photo that moves | No | AI image to video explains what pictures are used for |
The last two rows are the honest part. A remake reads your clip shot by shot, then makes a new picture; the source is never handed to the video model. Remake a video covers it.

## The ladder, from smallest to largest

| Size | Length | What it is | Cost |
| Free preview | About 5 seconds | One shot of your own episode, with your own cast, on the cheapest tier. Once per account. It can be shared | Free |
| Shortest segment | 4 seconds | The least a segment can be. A shorter total is rounded up to 4 | Four seconds at the tier’s per-second price |
| Standard segment | Up to 15 seconds | Three to five shots, one generation, one clip | Seconds times the tier’s price, on the card before you press |
| Long take | Up to 30 seconds | One generation with the camera cutting inside it. Only on tiers that support it | Same per-second price |
| Ad spot | 15, 30 or 60 seconds | One, two or four segments plus an end card, from up to four photos | One all-in price on the button. It is a cap, not a bill |
| Episode | Many segments | The segments joined into one file | Joining is 1 credit a second |
SceneMixer prices video by the second. The cheapest tier starts at 10 credits a second, about $0.06 at the lowest credit price. The price on a segment card follows the tier you pick for that segment, so a draft on the cheapest tier and the keeper on a better one are two separate choices. Draft at the lowest quality, finish at the highest explains the pattern.

## Make one clip from an idea

There is no one-box path from a prompt to a clip, because a shot needs a cast and a place to be drawn from. The shortest honest path is a script of one scene.

- Write one scene. In SceneMixer’s AI Create choose the shortest length, Mini, which the menu describes as one scene, a test. Or paste three or four paragraphs yourself. Choosing the script length.
- Read. The reading finds the people and the place. It is priced by length at 4 credits per thousand characters on the default, and English counts 3,333 characters as 1,000, so a short scene is a small read.
- Look at the pictures. Step 2 draws a sheet for each person and place. The first 15 on an account are free, and one scene rarely needs more than a handful.
- Open the storyboard. The first drama’s episode 1 storyboard is free. A scene usually comes back as several segments.
- Generate one segment. Nothing is shot until you press the button on a card, so shoot the first segment and leave the rest. You pay only for the seconds you generate.
- Delete the extras if you do not want the timeline of unused cards. Deleting a segment is free; an episode always keeps at least one. Inserting and deleting segments.
The preview is a quicker way to see whether the cast works: once the storyboard is open it makes one shot of about 5 seconds for nothing. The free preview.

## What a clip costs, step by step

| Step | Charge | Free allowance |
| Read the scene | 4 credits per 1,000 characters, English counted at 3,333 per 1,000 | None. A short scene is a small read |
| Sheets for the cast and place | Per image, at the image model’s price | The first 15 on an account |
| Storyboard | By the episode’s length in characters | The first 1 episode of the first project |
| The clip itself | Seconds times the tier’s price per second | A free preview of about 5 seconds, once per account |
| Trimming it afterwards | 5 credits per save, as a new version | None |
For the full price list see the pricing page; for what is free on a new account see What you get for free. A new SceneMixer account also starts with 50 credits.

## When the clip is a few seconds too long

In SceneMixer, select the segment, open Edit this clip and use Trim. In Keep mode the two handles frame what stays; in Remove mode they frame what goes, which cuts a moment out of the middle. The cut is made at the exact frame, picture and sound together, and nothing is generated again, so the faces and voices are the ones you already had. It costs 5 credits and is saved as a new version; the clip you started from is kept. Remove a moment from the middle.

To make a clip shorter before it is generated, change the seconds of its shots: they add up to the segment, and the segment is what you pay for. Change the segment duration.

## What it will not do

- Cut clips out of a long video. There is no highlights or clipping tool. The remake reads a video and shoots it again with your own cast.
- Return a clip from a text box. A SceneMixer clip comes from a script, a cast and a storyboard, so the shortest path has a few steps in it.
- Animate one photo. Pictures are references for who and where, not frames to move.
- Make a single shot longer than the tier’s cap. A segment stops at 15 seconds, or 30 on the tiers that allow long takes.
If what you want is a finished, longer thing, an episode is these clips joined. Where credits go in one episode shows the whole sum.

## FAQ

### What is the shortest clip I can generate?

SceneMixer bills a segment for at least 4 seconds. The free preview is shorter, about five seconds, but it is one shot made once per account for you to judge the cast.

### Can I make a clip from just a text prompt?

Not in one box. A SceneMixer clip starts from a short script, which is read into a cast and a place, then shot as a segment. A one-scene script is the shortest honest path.

### Can I cut clips from a long video?

No. SceneMixer has no clipping or highlights tool. The remake path reads a video and makes a new picture; it does not export pieces of the original.

### How long can one clip be?

A SceneMixer segment is up to 15 seconds, or 30 on the tiers that support long takes. A longer video is segments joined into an episode.

### Is there a free clip?

SceneMixer gives a free preview of about five seconds, once per account, made from your own script and cast. The first setup images and the first storyboard are also free.

### How do I make a 15 second ad?

Use SceneMixer’s Ad Video tab with up to four photos of one subject and choose 15 seconds. The price on the button is a cap.

### How do I shorten a clip I have already generated?

Open Edit this clip in SceneMixer and trim it, or remove a moment from the middle. Each save is a small charge and creates a new version.

### Can I make the clip vertical?

Yes. SceneMixer asks for the format when you start the project or the ad, landscape or portrait, and every shot follows it.

Updated November 2, 2026


---

# AI Create: Why Some Script Lengths Are Missing From the Menu

URL: https://scenemixer.com/guides/ai-create-why-some-lengths-are-missing-from-the-menu/
The length menu in AI Create hides targets whose finished script would not fit the parse limit. How the limit is counted per language and what happens on overflow.

# AI Create: Why Some Lengths Are Missing From the Menu, and What Happens If the Script Comes Out Too Long

Quick answer: AI Create writes to a target length and the result is parsed like any pasted script, so it has to fit the parse limit. The menu therefore only offers targets that will fit for your writing language: a Chinese script is counted in characters, an English one in words converted to characters, and alphabetic languages count lighter, so the same menu shows different options. If a draft still runs over the limit, the task fails, no project is created, and the writing charge is refunded. Pick the next length down, or write in parts and parse them as volumes.

Switch the interface language and open AI Create: the length menu changes with it because the count unit changed.

## What the length menu shows

Targets from a short episode up to the largest that fits: the cap on a single parse (30000 characters by default; the two-pass read for long scripts does not raise it for AI Create) divided by the expected size of a finished script at each target. Targets that would overflow are simply not listed. Choosing a length: Choosing Script Length and Cost.

## Why it differs by language

Chinese counts one character as one. SceneMixer counts English and other alphabetic languages at roughly three characters as one for the limit and the price, because an English sentence is longer in characters than the same sentence in Chinese; Japanese and Korean count like Chinese. The target itself is given to the writer in that language’s natural unit (characters or words). Length Caps and How Characters Count.

## When the draft runs over

Writers do not hit a target exactly. A draft that comes out above the cap is refused at the end of writing: the task fails with a message, no project is created, and the writing credits are refunded. Choose a smaller target, or ask for the story in two parts and parse each as its own volume.

## Writing in your language

AI Create in SceneMixer writes in your interface language and is priced by that language’s unit. The finished script is parsed into the project’s working text; the dialogue language is chosen in Step 1 as usual. Write the Script in Your Language.

## Ads

Ad scripts are short by design and never approach the limit; their length is set by the ad’s duration, not by this menu.

| Language | Counted as | Effect on the menu |
| Chinese, Japanese, Korean | One character = one | Fewer of the longest targets |
| English, other alphabetic | About three characters = one | More targets fit |
| Any | Draft over the cap | Task fails, refund, no project |

## FAQ

### The menu used to have a longer option.

Your interface language changed, or the cap did; SceneMixer’s menu follows both.

### Was I charged when the draft ran over?

No. SceneMixer refunds the writing charge when the draft is refused.

### Can I write a full novel?

In parts: write and parse each volume in SceneMixer as a project or an episode range.

### Does the target count dialogue?

Yes; SceneMixer counts the whole script text.

### Why is English 'cheaper' per character?

Because it takes more characters to say the same thing; SceneMixer’s factor keeps the price about equal per story.

Updated November 3, 2026


---

# AI Drama: Choose the Size of Your Season Before You Start

URL: https://scenemixer.com/guides/ai-drama/
Four sizes of AI drama, pilot, mini-season, season and long run, with the script length, reading, first steps and ceilings that go with each, and the 50-episode limit.

# AI Drama: Choose the Size of Your Season First

Quick answer: An AI drama is a script that becomes episodes, with the same cast and places in every one. The first decision is not the story, it is the size, because size decides how much script you need, how it is read and how much to make before you judge it. There are four useful sizes: a pilot of one episode, a mini-season of a handful, a season of about ten and a long run. One project holds up to 50 episodes; past that, part two is its own project and the cast comes with it. This guide gives each size its script length, its first steps and its ceilings.

## Four sizes

| Size | Episodes | Script | What you commit to first |
| Pilot | 1 | AI Create Brief (one or two short episodes), or your own pages | Everything: read, sheets, one storyboard, a few segments. The whole path for the least money |
| Mini-season | 3 to 6 | AI Create Short (a handful of episodes) | Episodes 1 and 2 only. Read the others as scripts, make them later |
| Season | About 10 | AI Create Medium (a short season), or a longer script of your own | Episodes 1 to 3. The cast has to hold up before you pay for seven more |
| Long run | Up to 50 in one project | AI Create Long (a full series), or a novel read in two passes | Episodes 1 to 3, then add the rest in batches as the story is written |
SceneMixer’s AI Create menu names its lengths after what they give, and the read decides the split: a longer script makes more episodes, not longer ones. Choosing the script length has the prices.

## What does not change with size

Whatever the size, SceneMixer makes each episode the same way and keeps it small. A Short Drama episode stops at 240 seconds, a TV Series episode at 300 seconds and a Film has no cap. An episode is up to 60 segments, each of 15 seconds or 30 on the long-take tiers. The ten English samples run from about one and three-quarter minutes to three and a half, with four to nine characters in one to three places, and that is a comfortable size for one person to finish. Short Drama, TV Series or Film sets the cap and the cut.

What scales is the number of times you repeat it. A 60 second episode on the default tier costs 660 credits for video and joining, about $4.11 at the lowest credit price, so ten of them cost ten times that before any retakes. The cast, the places and the voices are made once and reused. How much a micro drama costs has the worked sum.

## The first episode is the test

Whatever size you plan, make the first episode before you commit to the second. In a drama the first episode tells you three things a script cannot: whether the cast looks like the same people from shot to shot, whether the voices suit them, and whether the pace works at the length you chose. The first project’s episode 1 storyboard is free, the first 15 setup images on an account are free, and the free preview shows one shot with your cast before you pay for any video. What you get for free.

A sensible order is: read the whole script, check the summary, make the cast sheets, storyboard episode 1, preview, shoot episode 1 as drafts on the cheapest tier, fix single shots, then join. If episode 1 holds, repeat for episodes 2 and 3 before touching the rest. Checklist before you generate videos lists what to check.

## Growing a drama after the first read

A season does not have to exist as a script on day one. The Add an episode card at the bottom of SceneMixer’s step 3 takes more pages, pasted or uploaded, and only the new pages are read and charged. Let AI continue writes one to three episodes from where the story stands. AI story continuer says what it reads. Anything that first appears in the new pages is added to step 2 with the price shown.

A project holds up to 50 episodes, the same ceiling the first reading uses. When you get there, the next part is its own project: write it, start a new project with the same visual, scene and costume styles, and in step 2 use Pick existing on each returning character. The whole entry comes across: name, images, description, voice and outfit states, with no charge. Reuse a character, scene or voice across projects.

## What fixes at the start, what can change

| Decision | When it is fixed | How to change it later |
| Type: Short Drama, TV Series or Film | When the script is read | New project with the same script |
| Visual style | When you continue past step 1 | New project |
| Dialogue language | When the script is read | New project, or a second-language version (guide) |
| Aspect ratio | Step 1 | Set it before the shots; it is the shape of every one |
| Quality tier | Per segment | Any time. A new generation on a new tier replaces nothing: it adds a version |
| Episode order and count | The reading | Add episodes at the end; delete or insert segments inside an episode |
The practical rule: decide the type, the look and the language before you press the button that reads the script. Everything else is cheap to change.

## Where to begin

If you have a story but no script, start with a pilot from one line and read it. If you have a script of several episodes, read it whole and make the first one. If you have only a premise, the one-line brief is worth ten minutes: Writing the one-line brief. The rest of the series follows one step at a time in How to make an AI short drama, and the generator itself is described at AI drama generator.

## FAQ

### What is an AI drama?

A serialised story whose cast, places and footage are generated from a script. Each episode is a run of short clips joined into a video, with the same characters throughout.

### How long is an AI drama episode?

As long as the type allows: SceneMixer lets Short Drama run up to 240 seconds and TV Series 300, with no cap for a Film. The samples on this site run between about one and three-quarter minutes and three and a half.

### How many episodes can I make?

A SceneMixer project holds up to 50 episodes. For more, start a new project for part two and bring the cast with Pick existing.

### How many episodes should I make first?

One, then two or three before you commit to the rest. The first tells you whether the cast and the pace work.

### Do I need a finished script?

No. You can give SceneMixer’s AI Create a line and read the draft, or start with a pilot and add episodes later.

### Does the cast stay the same across episodes?

Yes. SceneMixer makes the cast once and every episode uses the same pictures, descriptions and voices.

### Can I change an episode’s type later?

No. SceneMixer fixes the type when the script is read, because the episode breaks were cut for it. Start a new project with the same script to use another type.

### What does an AI drama cost?

SceneMixer charges by the stage: a read by script length, pictures, a storyboard per episode and video by the second. The price is shown on each button, and the first episode has free allowances.

Updated November 2, 2026


---

# AI-Edit a Character Sheet: Outfit, Hair, Expression | SceneMixer

URL: https://scenemixer.com/guides/ai-edit-a-sheet-outfit-hair-expression/
AI Edit Image changes a sheet for the whole series; AI Fine-tune in the editor changes it for one episode. What to type, what each keeps, and what it costs.

# AI-Editing a Sheet: Outfit, Hair, Expression

Quick answer: AI Edit Image on a step-2 card keeps the same character and applies a short change you describe — navy blue shirt, short curly hair, smiling — then Apply Edit. The new image replaces the sheet, so every later shot follows it. AI Fine-tune in the episode editor does the same from a reference of the same type, but the result is applied to that episode only, with Keep or Undo. Both are priced like an image. On the current plan settings both are membership features.

Open a finished episode and look at the left panel: an episode-only image would show here, marked as this episode's.

## AI Edit Image, in step 2

Open the card, press AI Edit Image. The hint under the box says what it does: it uses the current image as reference — AI keeps the same character and applies your edits. Type the change in the box (Describe what to change…) and press Apply Edit. A moment later the card shows the edited image in place of the old one.

This changes the character for the whole project: every shot generated from now on uses the edited sheet. Shots already generated keep the old look until you regenerate them.

## What to type

One change, in plain words, about something visible: an item of clothing and its colour, a hairstyle, an expression, glasses on or off, a scar. “Navy blue shirt”; “short curly hair”; “smiling”; “add a leather jacket”. Two or three small changes at once are fine; a redesign is not — for that, rewrite the description and regenerate. Rewriting the description.

The description does not change when you AI-edit. If the edit is meant to last, update the Clothing or Appearance text as well, so a future regeneration does not undo it.

## A change for one episode: AI Fine-tune

In SceneMixer's episode editor, open a character or scene from the left panel and choose AI Fine-tune. Select a reference of the same type and describe the change; Generate. You can close the panel — it prompts you to confirm when it is done. The result is applied to this episode: Keep it, or Undo to restore the original. The rest of the series keeps the step-2 sheet.

Use it for a costume that belongs to one episode, a bruise that heals by the next, a room decorated for one occasion. For a look that recurs across episodes, a state is the better tool. States.

## Cost and who can use it

Each edit or fine-tune is priced like an image; the price is on the button. On the current plan settings both are membership features, and the button says so when they are not available. Paid plans.

## When the edit does something else

An edit that also moved the face or changed the age asked for too much at once. Undo (in the editor) or regenerate from the description (in step 2), then ask for one thing. An edit that had no visible effect usually described a mood rather than a change; name the garment or the hair.

| Tool | Where | Scope | Result |
| AI Edit Image | Step 2, the card | Whole project | Replaces the sheet |
| AI Fine-tune | Editor, left panel | This episode only | Keep or Undo |
| Rewrite description + Regenerate | Step 2, the card | Whole project | New sheet from the text |
| State | Step 2, Add a state | The segments you assign | A second image alongside the main one |

## FAQ

### What is the difference between AI Edit Image and AI Fine-tune?

Edit Image in SceneMixer changes the character's sheet for the whole project. Fine-tune produces an episode-only image you can keep or undo.

### Does AI Edit Image change the description?

No. Update the Clothing or Appearance text yourself if the change should survive a future regeneration.

### Can I edit a scene image the same way?

Yes. The tools work on scene and prop cards too.

### Does it cost credits?

Each edit on SceneMixer is priced like an image; the price is on the button.

### The edit changed more than I asked.

Undo or regenerate, then ask for one visible change at a time.

Updated November 3, 2026


---

# AI Edit Messages Explained: Why an Edit Was Dropped, What to Do

URL: https://scenemixer.com/guides/ai-edit-messages-explained/
Used a name the project doesn't have, came back in another language, changed too much in one go: every AI edit message, why it happens, and the next step.

# AI Edit Messages: Why a Revision Was Dropped, and What to Do Next

Quick answer: Every AI edit message that says a revision was dropped also says you weren’t charged — failed attempts cost nothing. Most come from one rule: AI edit only uses names that already exist in the project, and one revision can’t change too many lines of dialogue at once. The messages about shots and timing near the top of this page are leftovers from the older rule — adding, removing and retiming shots is something AI edit does now. When the fix is to edit by hand, the message carries Open the editor →.

When a message asks you to say it another way, name the shot and the result: 'in shot 2, she answers without looking at him'.

## If you see “can’t add shots” or “can’t change timing”

These two are older messages; they can still show on an editor page opened before the update. AI edit adds, removes and retimes shots now: say it in the same box, and the review shows shots X → Y and the new total seconds. If you see one of them, reload the editor and ask again — or make the change by hand, which is free and instant.

- AI edit can’t add shots… / can’t remove shots… / can’t change shot timing… — the older rule; reload the page and ask again.
- The AI changed the shot timing, so the result was dropped… — you did not ask for a timing change, but the AI made one, so the result was dropped. Your request was fine; run it again, or say the change you want explicitly.

The box can also warn under it as you type: AI edit keeps the number of shots and their timing as they are; for changes like that, edit by hand. If you see that note, adding, removing or retiming shots is not available at the moment: make the change by hand, or try again a little later. AI edit or edit by hand.

## A name or a line is not allowed

- The AI can only use characters, scenes and items already in this project… To add someone new, add them to the asset library first.
- The AI wrote in a character, scene or item this project doesn’t have, so the result was dropped… — use names that already exist, or add the new one first. Add a character, scene or prop.
- The AI changed too much of the dialogue… Try changing one or two lines at a time. — one revision can’t add or drop too many lines of dialogue at once.
- The AI altered the fixed header lines, so the result was dropped… — the lines at the top of the segment must stay in place; say it another way.

## The AI could not do it well

- The AI couldn’t tell what to change… Say which shot and what to change, e.g. “make her line in shot 2 shorter”.
- The AI can’t make that change… Try saying it another way, or edit by hand.
- The AI didn’t change anything… Try being more specific.
- The revised text came back in a different language, so it was dropped… Please try again.
- The AI’s reply came back incomplete… Clicking “Revise” again usually works.
- The AI didn’t respond and you weren’t charged. Please try again.
- That edit didn’t work and you weren’t charged. Try saying it another way. — the catch-all.

## Before it runs

- Describe the change in up to 500 characters. — the box is empty or too long.
- This segment has no storyboard text yet. — older wording: a blank or hand-written segment now opens AI edit in Write mode (Let AI write this segment), and the AI writes the segment for you. If you see this message, reload the editor. Insert or delete a segment shows the whole flow.
- This segment’s text just changed. Please try again. — the text was saved elsewhere (another tab) after you opened AI edit.
- Too many edits in a short time. Please try again later. — the hourly limit.
- AI edit is temporarily unavailable. — switched off for the moment; hand editing still works.
- The buy-credits window — the balance does not cover one revision.

## Warnings in the review (they do not block saving)

- The dialogue in shot N takes about X s to say, but the shot is Y s long, so it may not fit. Use “Revise again” to shorten it.
- The new line is spoken by NAME, who wasn’t in this segment before. If that’s not who you meant, use “Revise again” and say who should speak.

| Message starts with | Why | Next |
| AI edit can't add / remove shots | Older page — structure changes are supported now | Reload and ask again |
| This segment has no storyboard text yet | Blank or hand-written segment | AI edit opens in Write mode |
| The AI changed the shot timing | The AI slipped | Try again |
| The AI can only use characters… | Unknown name | Add it to the library first |
| The AI changed too much of the dialogue | Too many lines added or dropped at once | Change fewer lines per run |
| The AI couldn't tell what to change | Vague request | Name the shot and the result |
| The AI's reply came back incomplete | A glitch | Press Revise again |

## FAQ

### Was I charged for any of these?

No. Only a revision that reaches the review on SceneMixer is charged.

### Why does the same request sometimes work on a second try?

SceneMixer’s AI writes a fresh revision each time; a slip (timing, incomplete reply) often does not repeat.

### Where do I add a new character?

In the episode's asset library, or in SceneMixer's Step 2 for the whole project.

Updated November 3, 2026


---

# AI Edit or Edit by Hand: Which to Use on an AI Storyboard

URL: https://scenemixer.com/guides/ai-edit-or-edit-by-hand/
Editing by hand is free, instant and changes anything; AI edit rewrites the wording and can also add, remove or retime shots. How to combine the two.

# AI Edit or Edit by Hand: Which to Use for a Storyboard Change

Quick answer: Use Edit (by hand) for anything you can fix faster than you can describe it: one word, a character chip from the @ menu, a number of seconds typed in directly. It is instant, free and exact. Use AI edit (1 credit per revision) when the change is easy to say but tedious to write: a colder tone across the segment, a character taken out of three shots, the weather changed everywhere it is mentioned — or the shape of the segment: more or fewer shots, different seconds. The best of both: run AI edit, then Edit by hand from its review to adjust a word before saving.

If you catch yourself rewriting the same idea in five places, that is an AI edit. If you know the exact words, type them.

## What editing by hand is best at

- Exact wording, down to a single character, placed exactly where you want it.
- Inserting a character, location, prop or voice as a chip with the @ menu. The @ menu.
- Speed: the change is saved the moment you press Save — nothing runs in the background, nothing to review first.

Editing by hand costs nothing. Adding and removing shots and changing seconds used to be hand-only work; AI edit does them too now, and its review shows shots X → Y and the new total seconds. What you can change in the storyboard.

## What AI edit is good at

- A change that touches several shots: a mood, a relationship, a character removed.
- Rewording in the voice of the rest of the text, in the language it is in.
- Camera and staging described in words, when you know the effect but not the phrasing.
- Adding a voice-over line in the right place and format.
- Reshaping the segment: one shot more, one fewer, or a shot made longer or shorter — say it in the same box, and the review shows the old and new counts and seconds.

It checks its own work: the number of shots, their numbers and seconds, the header lines, the names it uses and the language must all match, or the revision is dropped and you are not charged. Eight quick requests with examples.

## What AI edit will not do

AI edit on SceneMixer will not bring in a character or location that is not in the project as a proper chip, and one revision cannot add or drop too many lines of dialogue at once. A request that needs either is stopped before it runs, at no charge, and its message may offer Open the editor → so you can do that part by hand. (Adding or removing shots and changing seconds are fine to ask for — the review shows the new count and timing before anything is saved.)

## Combining the two

- Run AI edit for the broad change.
- In Review changes, press Edit by hand: the normal editor opens with the AI’s version already in it.
- Adjust what you like — a word, a number of seconds, a chip — and save.

Both kinds of save go into the segment’s text history on SceneMixer, labelled Before an AI edit or Before a manual edit, so either can be undone. Text history.

| Change | By hand | AI edit |
| Fix one word | Best | Works, but costs credits |
| Change the mood of the whole segment | Tedious | Best |
| Remove a character from several shots | Tedious | Best (not too many dialogue lines at once) |
| Add or remove a shot | Yes | Yes — ask in the same box |
| Change a shot's seconds | Yes — instantly | Yes — the review shows the new timing |
| Insert a character chip with @ | Yes | Only characters already in the project |
| Price | Free | 1 per revision shown |

## FAQ

### Is editing by hand really free?

Yes. Only generating costs on SceneMixer; saving text does not.

### Does either one regenerate the video?

No. Both change the text. Regenerate the segment to see the change on video.

### Can I undo an AI edit I saved?

Yes. Restore the previous text from History.

### Why did AI edit send me to the editor?

That message means the change is easier to make by hand this time — or that the page still follows the older rule, which kept shots and seconds fixed. If what you asked for was more or fewer shots or different seconds, reload the editor and ask again: AI edit does that now.

Updated November 3, 2026


---

# AI Edit: Rewrite an AI Storyboard Segment With One Sentence

URL: https://scenemixer.com/guides/ai-edit-the-storyboard-text/
Tell AI edit what to change in a segment's storyboard text: reword it, add or remove shots, change seconds — see every change in review, save nothing until you do.

# AI Edit: Rewrite a Storyboard Segment by Saying What to Change

Quick answer: On a segment’s storyboard card, next to Edit, press AI edit. Write what should change — “make her line in shot 2 colder and shorter”, or “add a shot after shot 3 where she reads the letter” — and press Revise. In a few seconds you get Review changes: every changed line marked, a count of what changed, and — when you asked for more or fewer shots or different seconds — the old and new shot count and seconds side by side. Save it, Revise again, Edit by hand or Discard. Each revision that comes back costs 1 credit; one that fails or changes nothing costs nothing. Saving changes the text only — regenerate the segment to see it on video.

Start with one of the eight quick requests above the box: it types the beginning of a sentence for you to finish.

## Step by step

- In the episode editor, find the segment’s storyboard card and press AI edit in the row of buttons under it. The card turns into Let AI revise this segment, with the line Say what to change: wording, action, camera, sound, or the shots and their timing.
- Write what to change, in your own words, up to 500 characters. Name the shot if it matters: “in shot 3, she slams the door instead of closing it”. The eight quick requests above the box type the start of a sentence for you to finish; free-form text works the same way.
- Press Revise. The button on SceneMixer shows the price (1 credit) and the line under it says No charge if the edit fails. It usually takes a few seconds, and can take close to a minute when the site is busy — do not assume it failed just because it is slow.
- Read Review changes. Removed text is struck through, new text is highlighted, and a summary line counts the changed lines. When you only asked for wording, it also reports the shots and total seconds as they are; when you asked for more or fewer shots or different seconds, it shows the change itself — Shots: 4 → 5 · 18s → 22s — plus a line telling you what the segment becomes and, if its regeneration price changes, the new price.
- Press Save to keep it. Nothing is saved before that. If the length changed, saving updates the segment’s duration along with the text.
- Regenerate the segment’s video to see the new text on screen.

On a phone, the same steps happen in a sheet that rises from the bottom of the Script tab. When the episode has more than one segment, the panel also offers Apply to the whole episode…: the same sentence runs on each segment you select, and segments that need no change are not charged. To put one requirement across the whole episode in a single run, see Revise the Whole Episode: One Request, Every Segment.

## What you can ask for

Rewording the segment in SceneMixer — lines, actions, moods, the time and weather, how a shot is framed in words, who is in a shot, a voice-over, even the location. Three requests that work well:

- “Make her line in shot 2 colder and shorter.”
- “Change the weather to light rain in every shot that mentions it.”
- “Take the man in the red coat out of shot 3 and 4.”

Changing the shots and their timing — add a shot, remove one, merge two, give a shot different seconds, or let the segment grow or shrink as a whole. The review shows the old and new shot count and seconds before you save, and saving writes the new duration into the segment. Three requests that work well:

- “Add a shot after shot 3 where she reads the letter.”
- “Remove the flashback shot.”
- “Give the last shot six seconds instead of three.”

Two boundaries still apply. The total has to stay inside your episode’s segment length — if the change would make the segment too long or too short, the result is dropped and you are not charged; ask for fewer or shorter shots instead. And it only uses characters, locations and props that already exist in your project, written in the language the text is already in. AI edit or edit by hand.

## What it cannot change

Some things are not part of the storyboard text, so asking has no effect — SceneMixer’s review names them and points you to the right place instead: a character’s face, hair, build and outfit come from the character image (update it, or add a state for a different outfit); a location’s fixed furnishings come from the scene image; voices are chosen on the character; background music has its own switch by the player; the spoken language is a project setting. When the whole request is about one of these, it comes back as a message and you are not charged. A visual glitch that appeared during generation is not in the text either — regenerating the segment usually clears it. Every AI edit message.

## After the review

- Save — the new text replaces the segment’s text. The text before it is kept in History. Text history.
- Revise again — back to your request, still in the box; the next try starts from the saved text again, not from the rejected revision.
- Edit by hand — opens the normal editor with the AI’s version in it, so you can adjust a word before saving.
- Discard — nothing changes. SceneMixer does not return the credit for this revision: the revision was made.

## Two warnings you may see

If a changed line looks too long for its shot, the review says the dialogue in shot N takes about X seconds to say but the shot is Y seconds long, and suggests revising again to shorten it. If a new line is given to a character who was not in this segment before, it names them so you can check that is who you meant. Neither blocks saving. Every AI edit message.

## Which text it edits

The one on your screen. If your storyboard is shown translated into your interface language, AI edit revises that translation and SceneMixer carries the change into the original, the same way a hand edit to the translation does. Translated storyboards.

## When it is not available

While the episode’s storyboard is being generated, while a previous edit to the translated text is still being carried into the original, and whenever too many edits happen in a short time — then AI edit pauses for a while and you’ll see a message asking you to try again later. Manual editing and history are not affected. If you see This segment has no storyboard text yet on a blank segment, refresh the page and open AI edit again.

| Button | What happens | Credits |
| Revise | AI rewrites the text and shows the changes | 1 if a revision comes back; 0 if it fails |
| Save | The text is replaced; the old one goes to History | 0 |
| Revise again | Back to your request, same text as a base | 1 for the next revision |
| Edit by hand | Opens the editor with the AI version | 0 |
| Discard | Nothing changes | The revision already shown stays charged |

## FAQ

### Does saving regenerate the video?

No. Saving in SceneMixer changes the storyboard text only. Press regenerate on the segment when you want a clip from the new text.

### Can it add a shot or change seconds?

Yes. Say so in the same box — “add a shot after shot 3 where she reads the letter” — and the review in SceneMixer shows the shot count and seconds before and after you save. Saving also updates the segment’s length.

### Can it bring in a new character?

SceneMixer uses only characters already in the project. Add a new one to the asset library first.

### Is my original text lost?

No. SceneMixer keeps the original version and the texts before your last five changes in History.

### Does it work on a phone?

Yes: SceneMixer’s AI edit button on the Script tab opens the same flow in a sheet.

### What does it cost?

The price shows on the Revise button before you run it, and the review shows what was charged. A revision that fails or changes nothing costs nothing on SceneMixer; saving and editing by hand are free.

### Why did it say I was editing too fast?

Too many edits in a short time. Wait a little and try again. Manual editing and history on SceneMixer keep working meanwhile.

### The segment has no storyboard text yet — what do I do?

Open SceneMixer’s AI edit: the panel becomes Let AI write this segment — say what happens in it and press Write. If you instead see This segment has no storyboard text yet, refresh the page and try again.

Updated November 3, 2026


---

# AI-Generated Video: Spotting It, Labelling It, What the File Carries

URL: https://scenemixer.com/guides/ai-generated-video/
What marks an AI-generated video: badge, metadata tag, platform declaration. What SceneMixer adds by plan, what survives a download, and how far to trust your eyes.

# AI-Generated Video: How It Is Marked and What Travels With the File

Quick answer: An AI-generated video can carry four different kinds of mark, and they are not the same thing: a visible badge in the picture, a tag in the file’s metadata, a declaration you tick when you upload, and the caption you write. SceneMixer adds the first two at delivery: a small AI badge burned into the bottom-right corner and an AI-generated tag in the file’s metadata, on every video from a free account and, on a paid plan, off by default. A shared link always carries the badge. The platform declaration and the caption are yours. This page says where each mark lives, what survives a download and a re-upload, what the platforms asked for when we last read them, and how far you can trust your own eyes to tell.

Try it: download a finished clip, play it, and look at the bottom-right corner. Then open the same clip through its share link and compare. The badge you see on the share page is a separate copy that is always marked.

## Four marks that all get called a label

| Mark | Who adds it | Where it lives | What happens to it |
| Visible badge | SceneMixer, at delivery | Burned into the picture, bottom-right corner | Part of the picture, so it survives download, re-upload and re-encoding |
| Metadata tag | SceneMixer, in the same pass | A text tag in the file’s metadata saying the video is AI-generated | Stays in the file we deliver; another tool or platform may drop it when it re-exports the video |
| Platform declaration | You, when you upload | The upload form of the platform you post to | The platform decides what to show viewers |
| Your caption or opening frames | You | The description, or a notice on screen | Whatever you wrote |

There is also a layer SceneMixer does not control: whatever the model services themselves may apply at the moment of generation. It is separate from the switch described below. What you get to keep has the terms side.

## What SceneMixer puts on the file, by plan

| You are | Badge and metadata tag | A shared link | A generated image |
| On a free account | Always on | Badge always | No badge |
| On a paid plan | Off by default; you can switch them on | Badge always | No badge |

The switch is a pill beside the model picker in the editor, reading AI Badge or No Badge, and the same control is in Settings. It is read when a segment is generated: clips you already have do not change, and a joined episode is stitched from the clips as they are, so generate all of an episode under one setting. Regenerate a clip to apply a new setting; the earlier version stays in its history. The panel warns, before you remove the badge, that you then assume the compliance risk of how the video is labelled wherever you publish it. The AI badge switch.

Sharing works differently on purpose. A shared video travels on without you, reposted and embedded, so the share page, its preview picture and the file your phone’s share sheet sends all carry a badged copy, made once per clip in about a minute. If you want an unbadged file for your own channel as a member, use Download, not Share. Why shared videos carry the badge.

## What the platforms asked when we last read them

Rules in this area move, so the dated version is the useful one. Our answers page read the platforms and the law on October 14, 2026. YouTube asks you to disclose, when you upload, realistic content that was altered or generated with AI, and shows its own label; photorealistic uploads can be labelled in the player and stylised or animated ones in the description. TikTok asks creators to label AI-generated content that shows realistic scenes or people, and excuses things like anime styles and generic narration voices. The EU AI Act’s transparency article applies from August 2, 2026, and binds providers of generative systems and those who publish content that resembles real people, places or events.

What these have in common is that the trigger is realism, and the declaration is on the uploader. A badge in the corner is not the same as ticking the platform’s box. This is general information, not legal advice; check the platform’s page on the day you publish. Do AI-generated videos need a label or a watermark? has the sources, and the 2026 compliance checklist covers likeness and copyright.

## Telling by eye: how far to trust it

There is no reliable test by looking. A good cue is a label, so look for those first: a visible badge, a platform label, a disclosure in the description. Next look for provenance: the file’s metadata, which a media-info tool can show, and the account that posted it. A SceneMixer file carries an AI-generated tag there, unless a paid plan turned it off.

The picture itself gives habits, not proof, and the habits fade as models improve. In our own pipeline the recurring ones are these: readable text inside the frame does not hold, because video models do not render it reliably; faces start to blend when four or more named people share one frame; a character’s outfit or injury can drop out across a cut when the story did not write it again; and generated dialogue is paced at reading speed, with little overlap. Treat each as a reason to look for a label, not as a verdict.

## Before you upload: a four-line routine

- Keep your own copy. Download the full-quality file; deleting on the site is one-way.
- Check the corner. Decide the badge setting before you generate an episode, so the joined file is consistent.
- Tick the platform declaration if the cut looks like filmed reality, and put a plain notice in the caption or the first frames if the badge is off.
- Do not upload what you have no right to use. No real person’s face or voice without documented consent.
Every sample project on this site is headed “AI-generated short drama” and says it was produced end to end with SceneMixer, which is the same habit applied to our own pages.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Frozen Wilderness; Costume Style Outdoor Gear; Visual Style Realistic HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two survivors with a description each; the cabin and the ridge as locations; the radio as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Where a mark can live | Put there by | Survives a re-upload? |
| Corner badge | SceneMixer | Yes: it is part of the picture |
| Metadata tag | SceneMixer | Not always: tools and platforms can drop it when they re-export |
| Platform declaration | You | Belongs to the post, not the file |
| Caption or opening notice | You | Belongs to the post or the picture you made |

## FAQ

### How can you tell if a video is AI-generated?

Not reliably by eye. Look for a visible badge, a platform label, a disclosure from the creator, and the file's metadata. Visual habits such as unreadable text or blended faces are hints, not proof.

### Does SceneMixer put a watermark on AI-generated video?

SceneMixer adds a small AI badge in the bottom-right corner and an AI-generated tag in the metadata. Free accounts always get them; on a paid plan they are off by default and can be switched on.

### Can I remove the badge?

SceneMixer members can switch it off for clips generated afterwards. Shared links always carry it, and the panel notes that you take on the labelling responsibility where you publish.

### Does the badge survive downloading and re-uploading?

Yes, when it was on: it is burned into the picture, so it stays through download, re-upload and re-encoding.

### Does the metadata tag survive?

It is in the file we deliver. Another tool or platform can drop metadata when it re-exports a video, so do not rely on it alone.

### Do I still need to tick the platform's AI box?

Yes, if the platform asks for it. A badge in the corner and a declaration on upload are separate things.

### Do AI-generated images carry the badge?

No. SceneMixer adds the badge to video, not to generated images.

### Do stylised or anime videos need a label?

The trigger on the platforms we read was realism, and anime-style content was treated more lightly, but rules change; check the platform on the day you publish.

Updated November 2, 2026


---

# AI Halloween Video: Ads, Greetings and Spooky Shorts for 2027

URL: https://scenemixer.com/guides/ai-halloween-video/
Halloween 2026 was Saturday, October 31; the next is Sunday, October 31, 2027. Make a Halloween ad, a greeting, a kid-friendly cartoon or a no-gore scare, and plan ahead.

# AI Halloween video

Quick answer: Halloween 2026 was Saturday, October 31; the next one is Sunday, October 31, 2027, so this is the page to plan from, not to rush from. Four kinds of Halloween video come out of SceneMixer: a shop or venue ad from a photo (a costume shop, a pumpkin patch, a haunted attraction, a cafe's autumn menu), a 15-second greeting with your name on the end card, a kid-friendly cartoon with a friendly ghost, and a no-gore scare for the week people look for one. Write the autumn into the one line and keep the Halloween words for the end card, and the same shots can serve until Thanksgiving.

One line for the box next to the Ad Video upload: Our family pumpkin patch and corn maze, open weekends in October; for families with small kids; last entry at dusk. Tone: Warm; 30 seconds; portrait.

## Where the date is

| Date | Day | What it means for a video |
| October 31, 2026 | Saturday (past) | Past. A Halloween ad posted now has a year to wait; a post-Halloween offer or a thank-you is the live option. |
| November 26, 2026 | Thursday | Thanksgiving in the United States: edit the end card of an autumn ad. |
| December 25, 2026 | Friday | Christmas: the Christmas ad guide |
| October 31, 2027 | Sunday | The next Halloween: work back from here (calendar below). |
Make the ad in the first week of October, not the last, so a shot that missed can be regenerated alone. Edit one shot and regenerate.

## Four Halloween videos, and which tool

| Video | Made as | Start from | Length |
| A shop, patch or attraction ad | An ad project | One to four photos of the place or the product | 15, 30 or 60 seconds |
| A Halloween greeting | An ad project, Warm or Funny, end card Happy Halloween from… | One photo of your shop or product | 15 seconds |
| A kid-friendly cartoon | A Short Drama or Film with a soft look | A one-line story | 60 to 120 seconds |
| A scare for the week | A Film or Short Drama | A one-line horror premise | 60 to 120 seconds |
The first two use the ad flow (photos and one line); the last two are drama projects. AI ad video generator · AI kids story video generator · AI horror movie generator.

## The ad: one line per business

| Business | First photo | One line | Tone |
| A costume shop | The shopfront with its sign | Our costume shop on Main Street, rentals and make-up, open until 9 in October; for last-minute party-goers. | Funny, A day in the life |
| A pumpkin patch | The patch from the gate | Our family pumpkin patch and corn maze, open weekends; for families with small kids. | Warm |
| A haunted attraction | The front of the building at dusk | Our walk-through haunted barn, ages 12 and up, open Friday and Saturday nights. | Twist, Epic |
| A cafe or bakery | The seasonal item, close | Our pumpkin-spice loaf, baked on Fridays; for people who want it still warm. | Mouth-watering close-ups |
| An online shop | The product alone, large | A set of six glass lanterns with flickering lights, for porch and table; arrives before the 31st. | Minimal, Show how it works |
A place photo is kept as it is, the people in it are removed, and the ad is set inside it with a generated cast; a product is redrawn as a clean studio shot. The end card carries the name and the button (Visit us, Book your night, Shop now), because readable text cannot be drawn into the picture. An ad for a shop or a venue · An ad for a restaurant · Which photos work.

## A greeting that lasts past the 31st

Write autumn into the one line (an October evening on our street, lanterns on the steps) and keep Halloween for the end card: Happy Halloween from… and a tagline. After the 31st, Edit end card and Rebuild final cut re-join the same shots with a new card (Thank you for a spooky season, then Thanksgiving hours); no reshoot. The end card · Rebuild the final cut.

## Kid-friendly, and scary without gore

For children: choose a soft look (Family 3D Animation or Watercolor Storybook), a gentle set, and a ghost that wants a friend; a friendly ghost, a lost cat and a lantern make a 60-second story with a happy ending. Bedtime stories · Educational cartoons.

For a scare: something ordinary that is slightly wrong, a sound before a sight, a cut at the worst moment. Gore and extreme violence are blocked, and a blocked attempt is normally not billed. The horror short guide has the whole recipe, including a three-evening schedule. How to make an AI horror short · A haunted house drama · Content rules.

## Counting back from October 31, 2027

| Date | Day | Do |
| Friday, October 1, 2027 | Friday | Make the main 30-second ad. Read the end card and the one line. |
| Monday, October 4, 2027 | Monday | Fix any shot that missed; rebuild the cut; export. |
| Friday, October 15, 2027 | Friday | Post a 15-second teaser made from a second photo. |
| Friday, October 22, 2027 | Friday | Post the main ad: the last full weekend before the date. |
| Sunday, October 31, 2027 | Sunday | Post the greeting; change the end card to Happy Halloween. |
| Monday, November 1, 2027 | Monday | Edit the end card to a thank-you or the next offer. |
A drama or a horror short needs more time: write in September, make the cast and the storyboard in the first week of October. A New Year video · A love story video.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Picker | Options | Changes the price? |
| Vibe | Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life | No |
| Length | 15, 30 or 60 seconds (one, two or four segments) | Yes |
| Format | Portrait 9:16 or Landscape 16:9 | No |
| Quality | 480P, 720P or 1080P | Yes |

## FAQ

### When is Halloween?

October 31. In 2026 it was Saturday; in 2027 it is Sunday.

### Can I make a Halloween video with AI?

Yes: a shop or venue ad from a photo, a 15-second greeting with your name on the end card, a kid-friendly cartoon, or a no-gore scare.

### Is it too late for Halloween this year?

Halloween 2026 has passed. Make the ad now and post it in October 2027, or use the same shots with a new end card for Thanksgiving.

### Can it make a scary video?

Yes, without gore. SceneMixer blocks gore and extreme violence. Use dread, sound and a cut at the worst moment.

### Can I make a kid-friendly Halloween cartoon?

Yes. Choose a soft look such as Family 3D Animation or Watercolor Storybook and write a gentle story with a happy ending.

### Can I put my business name on a Halloween video?

On the end card, yes: brand, product name, a tagline and a button. Text is not drawn into the picture.

### Can I reuse a Halloween ad after the 31st?

Yes. Edit the end card and rebuild the final cut; write autumn rather than Halloween into the line if you want the shots to last.

### How long does an ad take?

A 15-second one-click ad on SceneMixer is typically ready in about ten minutes; longer ads take longer because there are more segments.

### What does a Halloween ad cost?

One figure is quoted on SceneMixer before it starts and it is a ceiling; length and quality set it, tone and format do not.

Updated November 3, 2026


---

# AI Image to Video: Four Methods, and Which Ones Work Here

URL: https://scenemixer.com/guides/ai-image-to-video/
Image to video AI is four methods: animate a still, pin first and last frames, use pictures as references, or turn photos into an ad. What each needs and what works here.

# AI Image to Video: Four Methods, and Which You Get

Quick answer: “Image to video” covers four different things: bringing one still picture to life, pinning the first and last frame of a clip, using pictures as references for who and where, and turning several photos into an ad. They need different inputs and give different results. SceneMixer does two of them. Pictures as references: a photo can become the face, the room or the object that every shot follows. And photos as an ad: up to four photos become a 15, 30 or 60 second spot. It does not animate a single still on its own and does not take a first or last frame as pictures. This guide sorts out which method you need and what to prepare for it.

## Four things people mean

| Method | What the picture does | You give | Here? |
| Animate a still | The picture is frame one; a sentence says how it moves | One image and a line about the motion | No. There is no one-button animate-this-photo tool |
| Pin first and last frame | Two pictures fix where a clip starts and where it ends | Two images | No. A segment is made from the cast pictures and its own storyboard text, not from frame pictures you supply |
| Pictures as references | The pictures say who and where; the model draws a new shot of them | One picture per character, location or prop | Yes. It is how every shot is made, and your own photo can replace any of them |
| Photos into an ad | Up to four photos are read, a product is redrawn, a spot is written around it | One to four photos and one line | Yes. 15, 30 or 60 seconds |

The difference decides what you see. With a reference, the picture says what a person or place looks like, not how the shot begins. The model composes a new picture of them from the shot text, close or wide, standing or sitting, so the result is your subject in a new shot, not your photo moving. If you need the exact frame to be your photo, that is the first method, and this tool is not it. Image to video AI covers the two routes that are offered in more detail.

## What you have, and what to do with it

| You have | Do this | Expect |
| A product photo, or up to four of the same thing | Ad Video tab; the first photo is the subject | A made-up spot around your product, redrawn as a clean studio shot, ending on a card with your brand |
| A shop, restaurant, car or house | Ad Video tab | The place is kept as it is, people are removed, and the first scene is built from your own photo |
| A room or street a story should happen in | Upload & Replace on the location card; no people, 16:9 | Every shot at that location is built on it |
| An object that matters to the plot | The prop card, the object alone on a plain background | It is the reference whenever the prop is in frame |
| A character you drew or made | Upload & Replace, then Use this image as is | It becomes that character’s sheet untouched |
| A photo of a real person | The character card, with care | It may be flagged if it reads as a real person’s face; celebrities and brand mascots are refused |
| An app or website screenshot | Ad Video tab; one screen, no faces | Never redrawn, so the interface stays readable |
| One still you want to see move | Not offered | Closest: put the subject on a card and write a one-shot scene around it |

Putting your own picture on a card is a membership feature on the current plan settings; photos for an ad are a separate door and are not part of that. Use your own photo · Which photos work for an ad.

## What happens to a picture when you pick it

On SceneMixer, your browser reads WebP, HEIC, PNG or JPEG, flattens any transparency onto white and re-encodes it, so what arrives is always JPEG or PNG. The only refusals are files your own browser cannot open, such as SVG, camera RAW and HEIC outside Safari, and you find out the moment you pick the file. An ad takes up to four photos of the same subject, up to ten megabytes each and sixteen in all, and reads them once, for free, before anything is charged. The first time you upload anything, a window asks you to confirm that you have the right to use it and that anyone whose face or voice is in it consented. What you can feed in.

On a character card you choose how the picture is used: Optimize into a reference sheet redraws it in the project’s look on a clean background and is priced like an image, while Use this image as is is not an image generation. Either way, new videos for that character look at it from then on; clips made earlier are untouched. Your own reference images: what changes.

## The same picture, three different shots

Because the shot text decides the framing and the action, one photo on a card gives you as many different shots as you write. With a character called Ana on a card:

| Shot text | What it asks for |
| [Shot 1] 3s: Close-up on @Ana at a window. She lets the curtain fall. | Her face large, one small action |
| [Shot 2] 5s: Wide shot, a hallway. @Ana walks away from the camera and stops at a door. | The whole figure in a room, moving away |
| [Shot 3] 4s: Over-the-shoulder, @Ana facing a stranger. @Ana: You are early. | A second person in frame and a spoken line |
A segment of shots like these is one generation, charged per attempt, and SceneMixer keeps every take. Editing the text is free, so write the shot until it says what you want, then generate. How to write one shot.

## Across clips: what carries over from a picture

Each SceneMixer segment starts from the cast pictures and its own storyboard text. It does not start from the last frame of the segment before, so the same photo gives the same person in every segment, but a segment does not open on the exact picture the last one ended on. What has to carry across a cut, a bandage, a coat taken off, is kept by writing it as an event in the shot where it happens, or by a state picture for something that lasts. Why the next segment opens like the last one ended · What the system keeps track of.

## When the picture you start from is a video

A finished clip can itself be the starting point. SceneMixer’s AI visuals makes the clip again from itself with one change from a fixed menu of eight, and everything else about how it was shot carries over. That is video to video, not image to video, and it is priced and limited differently. AI visuals: what it can change.

## FAQ

### Can I turn one photo into a video here?

As an ad, with up to four photos of one subject. A photo can also become a character, location or prop that shots are made around. SceneMixer has no tool that animates a single still on its own.

### Can I choose the first and last frame?

Not as pictures. In SceneMixer you can describe how a segment opens in its first shot, but a segment is made from the cast pictures and its own storyboard text.

### Will the video look exactly like my photo?

SceneMixer redraws a product as a clean studio shot, keeps a place, and a character sheet made from a photo keeps the face, hair and clothing redrawn in the project’s look.

### What files can I upload?

Your browser reads WebP, HEIC, PNG or JPEG and sends JPEG or PNG. SceneMixer refuses SVG and camera RAW when you pick them.

### How many photos can an ad use?

A SceneMixer ad takes up to four photos of the same subject, ten megabytes each and sixteen in total.

### Does the same photo give the same person in every shot?

Yes. SceneMixer makes every shot that shows the character from the same entry. Outfit changes are explicit states you choose.

### Can I use a photo of myself?

A photo that reads as a real person’s face may be flagged, and uploading to a card is a membership feature. The guide on using your own photo explains the details.

### What does it cost?

SceneMixer reads ad photos for free. Video is 10 credits a second on the default tier, and the price shows on the button before you press it.

Updated November 2, 2026


---

# AI Image to Video Creator: What to Do After the First Cut

URL: https://scenemixer.com/guides/ai-image-to-video-creator/
After a photo ad is made: fix one shot, rebuild the final cut, edit the end card, make variants from the same photos, and what to do when a shot is blocked by review.

# AI Image to Video Creator: Finishing the Ad After the First Cut

Quick answer: The first cut of a photo ad is rarely the last. With an ad made from up to four photos, the work after the first cut is a short loop: look at the shots, fix the one that is wrong without touching the rest, edit the end card, rebuild the final cut and, if you want to test, make variants from the same photos. Most of that is cheap and none of it reshoots the whole ad. A different vibe, length or format is a new ad. This guide is that loop, with what each step costs and what to do when a shot is blocked by content review.

## The loop after the first cut

| Step | Where | What it does | Cost |
| 1. Watch | The finished ad on the progress page | Look at each shot, the end card and the sound | Free |
| 2. Fix a shot | Fix this shot in the editor | Edit that shot’s text, regenerate that 15-second segment | The segment’s seconds at the tier price. The other segments are not touched |
| 3. Fix the end card | Edit end card, then Assemble | Change brand, product name, tagline or call to action | A re-join, priced by the finished length |
| 4. Rebuild | Rebuild final cut | Re-join the same shots with the current end card; nothing is reshot | The same re-join. If it fails, the previous cut stays |
| 5. Vary | Make another | Back to the composer with new settings | A new ad at its own price |
| 6. Publish | Download or share | The finished file | Free; a free account carries the small AI badge |
SceneMixer’s progress page keeps this list as buttons, and a stage that fails offers Try again, which continues from that stage and keeps what is finished. You can leave the page while it runs; the home card for the ad shows its stage. The ad progress page.

## Fixing one shot without redoing the ad

SceneMixer makes an ad in 15-second segments, each shot on its own, so a bad shot is a local problem. In a one-click ad the script is fixed but the shots are not: open the ad in the editor, change that shot’s text and regenerate that segment, then Rebuild final cut. Pick the repair by what is wrong.

| What is wrong | Repair | Notes |
| The picture of one shot | Edit the shot text, regenerate that segment | Earlier takes are kept, and switching back is free |
| A spoken line | Redub on the clip | Lip sync is 1 credit a second of clip. In an ad scored the new way, the ad’s music returns by itself when you join again |
| A stumble or a dead moment | Trim | 5 credits, and a trimmed segment is cheaper to join |
| A colour, the time of day, the weather, background people | AI visuals | 2 times the per-second price of the tier; see the guide on the eight changes |
| The product looks wrong | In step 2 of an expert-mode ad, Redraw from original on the product card | The product is only ever redrawn from your own photo, never from text |
| A brand name or tagline | Edit end card | The one place readable text is guaranteed |
The script of a one-click ad cannot be edited afterwards. If the wording of the script matters, that is the reason to choose expert mode in the first place, where the script can be edited or rewritten by AI until it has been broken down into cast and shots. One-click or expert mode · Edit this clip.

## Variants from the same photos

The photos can be used again for another ad, and the reading is free each time. What a variant changes depends on which setting you touch.

| Change | Counts as | Changes the price? |
| Ad vibe | A new ad | No |
| Format: portrait or landscape | A new ad | No |
| Length: 15, 30 or 60 seconds | A new ad | Yes |
| Quality | A new ad | Yes |
| The one line | A new ad | No |
| End card text | The same ad, rebuilt | A re-join only |
A sensible test set is therefore the same photos in a few vibes and both formats at one length and quality, which keeps the prices equal and compares only the creative. Each ad is its own project with its own all-in price, so six variants is six prices, and the account has limits: there is a cap on projects (“You’ve reached your project limit — delete one to continue”) and a rate limit on starting ads in a short time. Start with the cheapest quality and the shortest length, and pay for the step up only on the variant that wins. Vibe, length, format and quality · The generator’s settings.

## When a shot is blocked by content review

Sometimes SceneMixer’s check at the video stage refuses a segment. The ad stops there, the refused segments are refunded within the daily limit, and the progress page lists them with the reason. Retry is refused until you have changed the refused text, because the same text would be refused again and every block counts toward an account-level pause on video.

| Reason | Likely cause in an ad | Change |
| A person or face | A named or recognisable person in the script | Describe a generic person |
| A lookalike | A celebrity, a character or a famous logo | Rename it or describe it generically |
| A claim | A cure, a guarantee, a medical effect | Show it instead of claiming it |
| Violence | A weapon used for a joke | Cut it |
| Clothing | Outfit wording the review flags | Describe the outfit plainly |
Edit the refused segments in the shots view, then retry. The segments that were delivered are charged and kept, and only the edited ones are regenerated. The one-price cap still applies to the whole ad. An ad blocked by content review.

## What you cannot change on a finished ad

- The mode. A one-click ad cannot become an expert-mode ad or the reverse. Each is a fresh ad.
- The length of the ad and the number of segments. SceneMixer fixes segment length at 15 seconds, and the number of segments comes from the ad length.
- The vibe and format. They are part of how the ad was made.
- The music. It is written for the ad and cannot be picked. To use your own, download the final cut and replace the track in an editor. Where the music comes from.
- Whether the ad can be cancelled. Cancelling works before shooting starts; credits spent on steps that have started are not refunded, and a cancelled ad cannot be resumed.
For how the ad is first made, from the photos to the first cut, see photo to video AI and AI ad video generator.

## FAQ

### Can I change one shot of a finished photo ad?

Yes. Open the ad in the SceneMixer editor, change that shot’s text and regenerate that segment, then rebuild the final cut. The other shots are not reshot.

### How do I change the text on the end card?

Use Edit end card in SceneMixer, change the brand, product name, tagline or call to action, and Assemble. The final cut is rebuilt without reshooting.

### Can I change the vibe or length after the ad is made?

No. A different vibe, length or format is a new ad. SceneMixer reads your photos for free each time.

### What if a shot is blocked by content review?

The ad pauses, SceneMixer refunds the refused segments within the daily limit, and retry is refused until you change the refused text. Only the edited segments are made again.

### What does it cost to fix a shot?

Regenerating a segment costs its seconds at the SceneMixer tier price. A trim is 5 credits, lip sync is 1 credit a second, and a picture change is 2 times the per-second price.

### Can I edit the script of a one-click ad?

No. SceneMixer lets you edit the shots but not the script. Expert mode lets you edit or rewrite the script until it has been broken down.

### Can I cancel an ad?

Before shooting starts. Credits spent on steps that have started are not refunded, and a cancelled ad cannot be resumed.

### Does a variant need new photos?

No. You can add the same photos again. Each SceneMixer variant is its own ad with its own price.

Updated November 2, 2026


---

# AI Image to Video Generator: Every Setting and What It Changes

URL: https://scenemixer.com/guides/ai-image-to-video-generator/
Every control in the photo-to-ad generator: photos, one line, vibe, length, format, quality, end card. What each changes, what it costs, and three settings sheets.

# AI Image to Video Generator: The Settings That Matter

Quick answer: The generator takes one to four photos and makes a 15, 30 or 60 second ad. What you can control is short and specific: which photo leads, one line about what it is and who it is for, the ad vibe, the length, the format, the quality, the lines on the end card, and whether the AI or you runs the steps. Only length and quality change the price. Music has no control, because it is written for the ad. This guide goes control by control, says which can still be changed after the ad is made, and then fills in the settings for three real jobs: a restaurant, an app and a skincare bottle.

## Every control, and what it does

| Control | Options | What it changes | Price? | After the ad is made |
| Photos | 1 to 4, the first is the hero | The subject, the picture on the end card, the places the shots are set in | No | A different set is a new ad |
| One line | Free text, optional | The direction of the story. It wins over the photo when they disagree | No | New ad |
| Product or place name | Shown for correction after the reading | The name used on the end card | No | Edit end card |
| Ad vibe | Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life | How the story is told, the look of the ad and the colours of the end card | No | New ad |
| Length | 15, 30 or 60 seconds | How many 15-second segments there are: 1, 2 or 4 | Yes | New ad |
| Format | Landscape or Portrait | The shape of every shot and of the end card | No | New ad |
| Quality | 480P, 720P, 1080P | Sharpness of every shot | Yes, shown in the menu | Fixed for the ad |
| Mode | One-click generate or Expert mode | Who writes and runs the steps | A cap, or per step | Cannot be switched |
| End card lines | Brand, product name, tagline, call to action; any can be blank | The only readable text in the ad | A re-join, if changed later | Edit end card, then Assemble |
Notice what is on the right-hand side of the table. Almost everything is fixed when you press Make my ad, and a change of vibe, length or format means a new ad. That is why the reading and the settings are worth a minute: they are free to change until the button is pressed. Vibe, length, format and quality covers the four pickers on their own.

## What the one line really does

The line under the pickers in SceneMixer is optional, and it is the only control that is plain language. It is read together with the photo and wins when the two disagree: write “our family restaurant” under a picture of a dish and the reading treats the photo as a place, not a product. It also seeds the story. If you leave it blank, the chooser tells you the direction the AI picked instead, in a line that starts “No brief written”, so you can see what it will build before you spend anything.

A good line answers two questions in one breath: what it is, and who it is for. “A fragrance-free night serum for people with dry skin” gives the script a customer and a reason. “Great product” gives it nothing, and superlatives and absolute claims are cleaned out of the line, the tagline and the end card, so a line that leans on “the best” will not survive to the ad.

## Length is segments, and segments are the unit of everything

SceneMixer makes an ad in 15-second segments joined in a final cut with the end card at the end. So 15 seconds is one segment, 30 is two and 60 is four. Each segment is shot on its own, which has three consequences you can use.

- Price follows segments. More segments means more shots to pay for, which is why length is a price control and format is not.
- A bad shot is a small problem. Open the finished ad in the editor, change that shot and rebuild the final cut; the other segments are not shot again.
- Length decides how much story fits. Fifteen seconds is the product and one idea. Sixty seconds has room for a cast and a small turn, and uses the places and people the reading found.
The tagline is spoken out loud before the final freeze, so the end card repeats what the viewer just heard. The end card.

## Format and quality are the shape and the sharpness

Landscape suits sites, screens and pre-roll; portrait suits feeds and stories. The format is the shape of every shot and of the end card, and your product and place photos are used at their own shape either way. It never changes the price.

In SceneMixer, quality is the only picker that spends money on pixels. 480P is for drafts and small screens, 720P is what most people ship, and 1080P is for big screens. The menu shows the all-in price beside each, so you see the step up before you take it. If the quality you picked is unavailable when shooting starts, the ad says so and offers to continue with one that is. The same logic runs through draft at low resolution, finish at high.

## Three settings sheets for three real jobs

These are suggested settings, not rules. The behaviour in the first two rows of each column is how SceneMixer’s reading treats that kind of photo; the rest is a sensible starting point you can change.

| | A neighbourhood restaurant | A budgeting app | A night serum |
| Photos | Shopfront first, then the signature dish, then the dining room | Two screenshots, one screen each, no faces | The bottle on a plain surface first, then a texture close-up and one in use |
| What the reading does | Category Store / property. Customers are removed, the place is kept | Category App. Used exactly as it is, never redrawn | Category Product. Redrawn as a clean studio shot |
| The one line | Our family restaurant, for the lunch crowd | What it does and who uses it | What it is and for whom |
| Ad vibe | Mouth-watering close-ups, or Warm | Show how it works | Minimal |
| Length | 15 seconds to start | 30 seconds: two beats, problem then screen | 15 seconds |
| Format | Portrait, for feeds | Landscape for a site, portrait for stories | Portrait |
| Quality | 720P | 720P | 720P |
| Check on the end card | Brand read from the sign, call to action such as “Book a table” | Product name and a call to action such as “Download the app” | Brand, product name, a short tagline |
For the app column, a screenshot must be roughly a phone or computer screen shape: a full-page capture is far taller than any video frame and is refused until you crop it. Which photos work for an ad.

## Controls that are not there, and why

| You might look for | What happens instead |
| A music picker | None. The music is written for the ad, composed as one piece and mixed under the dialogue, then carried across the end card. To use your own, download the final cut and replace the track in an editor. Where the music comes from |
| On-screen text for the name and price | Video models do not write legible text, so ads are shot without it. The end card is where the name, tagline and call to action appear as readable words |
| A first-frame or last-frame picture | Not offered. Photos are references for the subject and places, and every shot is composed from the script |
| Your own footage | Not accepted by the ad route. Remake is the door for an uploaded video, and it reads the clip rather than reusing it. The spoken lines of an ad are written with the script and performed by the generated cast |
| A switch from one-click to expert mode | Not possible. Each is a fresh ad |
For what to do once the first cut is in, see AI image to video creator: finishing the ad.

## FAQ

### What can I control in an AI image to video generator?

In SceneMixer: the photos and their order, one line about what it is, the ad vibe, length, format, quality, the end card lines and whether the AI or you runs the steps.

### Which controls change the price?

Length and quality set the SceneMixer price. The vibe, the format and the one line do not. The quality menu shows the all-in price beside each option.

### How many photos can I use?

Up to four of the same subject. The first is the featured product and the picture on the end card.

### Can I change the vibe after the ad is made?

No. A different vibe, length or format is a new ad. The end card lines and any single shot can be changed without starting over.

### Can I choose the music?

No. SceneMixer writes music for each ad and mixes it in for you. To use your own, download the final cut and replace the audio in an editor.

### What language is the ad in?

The chooser shows the dialogue and end card language before you press either button. Check it there; some languages note that voices may occasionally mispronounce a word.

### Does the video start from my photo?

No. SceneMixer treats photos as references. A product is redrawn as a clean studio shot, a place is kept and a screenshot is used as it is, and each shot is composed from the script.

### How long does it take?

A one-click SceneMixer ad is typically ready in about ten minutes; longer ads take longer because they have more segments to shoot.

Updated November 2, 2026


---

# AI Micro-Drama in 2026: Market Numbers, Sources and What They Mean

URL: https://scenemixer.com/guides/ai-micro-drama-market-data-2026/
Downloads, revenue and release counts for micro-dramas in 2026, how much is made with AI, where each number comes from, and what it means if you make one.

# AI micro-drama in 2026: the numbers

Quick answer: Three numbers frame it. Short-drama apps passed 850 million downloads worldwide in the first quarter of 2026, up 140% on a year earlier, and took in about US$750 million in in-app purchases (Sensor Tower). Outside China, AI-made titles went from about 3% of vertical-drama releases in the first quarter to about two-thirds in the third (Owl & Co, as reported on 20 October). In China the regulator counted 430,000 micro-dramas released between January and August, more than 90% of them made with AI. The three sources count different things in different places, so read them side by side, not as one curve. Numbers checked on 26 October 2026.

## The numbers and where they come from

| Measure | Figure | Source and period |
| Short-drama app downloads, worldwide | More than 850 million in Q1 2026, up 140% on Q1 2025 | Sensor Tower, “State of Short Drama Apps 2026” (June 2026) |
| In-app purchase revenue, worldwide | About US$750 million in Q1 2026, up 20% on a year earlier | Sensor Tower, “State of Short Drama Apps 2026” (June 2026) |
| Top-grossing apps | DramaBox and ReelShort, each close to US$140 million in Q1 2026 | Sensor Tower, “State of Short Drama Apps 2026” (June 2026) |
| Where the downloads are | Southeast Asia 32%, Latin America 23%, India 22% — together more than three-quarters | Sensor Tower, “State of Short Drama Apps 2026” (June 2026) |
| AI share of vertical-drama releases outside China | About 3% in Q1, 28% in Q2, about two-thirds in Q3 2026, across 20 top platforms | Owl & Co, “What We Watched on Vertical”, as reported by The Hollywood Reporter on 20 October 2026 |
| Micro-dramas released in China | 430,000 from January to August 2026, 13 times the whole of 2025; AI more than 90% | National Radio and Television Administration, State Council Information Office press briefing, reported by Xinhua on 18 September 2026 |
| Micro-drama users in China | More than 800 million | National Radio and Television Administration, State Council Information Office press briefing, reported by Xinhua on 18 September 2026 |
| Size of China’s micro-drama market | More than 100 billion yuan in 2025 | National Radio and Television Administration, State Council Information Office press briefing, reported by Xinhua on 18 September 2026 |
The Owl & Co figures reach us through press coverage; we have not read the full report. The Chinese figures are the regulator’s own, as stated at the briefing.

## How to read them

- Downloads are not revenue. India is 22% of downloads; Sensor Tower names the United States as the revenue leader. Where people install an app and where they pay for it are different maps.
- The revenue figure is in-app purchases only. The report does not say whether it is gross or net, which stores it covers, or whether China is included, so it is not comparable to a box-office number. Downloads are estimates for the App Store and Google Play and leave out pre-installs, re-downloads and third-party Android stores.
- An AI share of releases is a share of titles, not of viewing. Two-thirds of new releases made with AI says how most new titles are produced. It does not say how many people watch them or for how long.
- Nobody defines “made with AI”. The pages we read do not say whether a fully generated series and a live-action one with AI-assisted editing count the same. Treat the percentages as indications, not as a measurement.

## China, counted separately

China is reported on its own: the regulator’s figures cover every micro-drama released there, and the global app figures do not say whether China is included. At the State Council Information Office briefing the regulator gave these numbers:

- 430,000 micro-dramas released from January to August 2026, 13 times the whole of 2025, with AI-made titles above 90%.
- More than 800 million users; a market above 100 billion yuan in 2025.
- 68,000 non-compliant micro-dramas taken down and more than 1,200 accounts dealt with, over 90% of the content handled coming from AI-made titles.
- The micro-drama development and management measures (微短剧发展管理办法) took effect on 1 September 2026; a micro-drama made with AI has to carry a prominent notice on each episode, and platforms must uphold portrait and voice authorisation.
- The stated line is that AI is a tool and people remain the authors. The regulator also said live-action dramas are the main force for high-quality work and that it will support a live-action section on platforms.
Another Chinese industry body’s quarterly counts, reported second-hand, are 128,000 titles released in the first quarter (122,000 of them AI-made) and 239,000 in the second (153,000 AI-made). The two quarters add up to 367,000 against the regulator’s 430,000 for eight months, and the second-quarter AI share works out at about 64% by our arithmetic, so the two sets do not obviously fit together. Neither publishes how it counts. Use the regulator’s figures for the January–August total and do not splice the two. Disclosure rules in other markets are in the compliance checklist.

## What it means if you make one

This is our reading, not something the sources say.

- AI-made is now the normal way to make a micro-drama, so it is not the pitch. What a viewer notices is the story and whether the people look the same from one shot to the next. Keeping characters consistent.
- Volume is enormous. A title that has to share a feed with hundreds of thousands of others needs its reason to keep watching in the first seconds. Hooks and cliffhangers.
- The audience is not where the money is. Most downloads are in Southeast Asia, Latin America and India; the United States leads revenue. Choose the dialogue language and region with that in mind: AI short drama in 15 languages.
- Know what a series costs before you start. What a micro-drama costs to make · How to make an AI short drama.

## What these numbers do not tell you

- No source here gives watch time or the share of viewing that goes to AI-made titles.
- The latest Sensor Tower quarter we found is Q1 2026; nothing for the second or third quarter yet.
- The Owl & Co figures come from a press summary, and the global and Chinese counts use different rules.
The figures above are as of 26 October 2026.

## FAQ

### How big is the micro-drama market in 2026?

In the first quarter of 2026 short-drama apps passed 850 million downloads worldwide, up 140% on a year earlier, and took in about US$750 million in in-app purchases, up 20% (Sensor Tower, June 2026). China is counted separately: the regulator put its 2025 market above 100 billion yuan.

### How much micro-drama is made with AI?

Outside China, about two-thirds of vertical-drama releases in the third quarter of 2026, up from about 3% in the first and 28% in the second (Owl & Co, as reported on 20 October 2026). In China, the regulator said AI-made titles were above 90% of the 430,000 released from January to August.

### Which countries download the most short-drama apps?

Southeast Asia took 32% of first-quarter 2026 downloads, Latin America 23% and India 22%. Sensor Tower names the United States as the revenue leader.

### How many micro-dramas are released in China?

The regulator said 430,000 from January to August 2026, 13 times the whole of 2025.

### Are these numbers comparable with each other?

No. They come from different bodies, measure downloads, revenue or releases, cover different places and periods, and none publishes what counts as made with AI.

### Does the AI share mean most viewers are watching AI-made series?

No. It is a share of titles released, not of viewing time.

### Do AI-made micro-dramas have to be labelled?

In China, since 1 September 2026 a micro-drama made with AI has to carry a prominent notice on each episode, according to the regulator. Other markets have their own rules; the compliance checklist lists them.

Updated October 26, 2026


---

# AI Movie Idea Generator: 48 Movie Ideas, Titles and Plot Twists

URL: https://scenemixer.com/guides/ai-movie-idea-generator/
Movie ideas you can use today: 48 one-line concepts by genre, a three-column idea mixer, title formulas, 16 plot twists and how to turn one idea into an AI Create brief.

# AI movie idea generator: ideas, titles and plot twists

Quick answer: This page is a working idea generator: 48 original one-line movie ideas across eight genres, a three-column mixer for making your own, ten title formulas and sixteen plot twists. Pick one, change the person, the place and the ending until it is yours, then paste it into SceneMixer's AI Create as a one-line brief: it writes the script, and reading the script gives you the cast and the storyboard. The Inspire Me button on that tab also fills the box with an example to edit.

Pick an idea below, add a lead, an obstacle, an ending and a tone, and paste it into AI Create; the price of the writing is shown on the button before you start.

## Drama and romance ideas

Drama

- A retired piano tuner is hired to tune a piano in a house he swore never to enter again; the owner is the daughter he lost touch with.
- Two sisters have one weekend to empty their late father's flat, and a locked room makes them decide what to keep.
- A bus driver on her last route knows every regular's story, and one of them is planning to get off for the last time.
- A boy teaches his grandmother to use a phone, and she uses it to find the sister she was told had died.
- A baker who has not spoken since a funeral must give the speech at the town's award dinner, because the winner is her rival.
- A night cleaner at a courthouse finds the missing page of a verdict.
Romance

- Two translators fall for each other's work before they meet; the book they share ends mid-sentence.
- Rival food-truck owners must share one parking space for a festival weekend.
- A lighthouse keeper and the supply-boat pilot have only ever met in storms.
- A wedding planner who is allergic to weddings is hired by her ex.
- Two strangers swap phones by accident on a train and spend one night returning each other's lives.
- A widower and a bookseller write to each other in the margins of the same copy of a novel.

## Thriller and horror ideas

Thriller

- A hotel night manager notices that the same guest checks in every night under a new name and always asks for room 9.
- A translator at a peace summit hears a sentence that was never said in the room.
- A courier delivers a parcel to the wrong flat, and its owner is already waiting with the right address in hand.
- A mountain rescue volunteer is sent to find a hiker whose trail log is signed by her late brother.
- A bank clerk finds that the safe-deposit box she guards was rented in her name ten years ago.
- A juror realises the defendant is the only person in the courtroom lying about being a stranger.
Horror

- A house sitter keeps finding the dog waiting at the door of a room that has no door.
- Every night at 3:07 the neighbourhood's smart speakers say the same name; tonight it is yours.
- A family finds the previous tenants' growth chart on the doorframe, and there is a new mark each morning.
- A radio host takes calls on the last night shift, and one caller describes the studio from the inside.
- A school trip to a lighthouse comes back with one more child than it left with.
- A village sets an extra place at the table on harvest night, until a guest asks why.
Make an AI horror short · A psychological thriller · A murder mystery.

## Sci-fi and fantasy ideas

Sci-fi

- A repair technician finds that the station AI has been rerouting the supply ship to postpone the one day it is switched off.
- The last human-run lighthouse on a flooded coast is kept by a maintenance robot that has started a diary.
- A courier delivers parcels to a city that exists only on Tuesdays.
- The one student on a generation ship asks where the destination is on the map, and nobody can show her.
- A town discovers that its clocks run backward inside the old library.
- A translator hired to speak to the first alien envoy slowly realises the envoy is translating her.
Fantasy

- A dragon's accountant discovers that the hoard is being quietly paid out as someone's pension.
- An apprentice mapmaker is told to draw a country that does not exist yet, and each map he draws becomes true.
- A village witch who sells only small spells, a good sleep or a found key, is asked to cast one big one.
- A knight who cannot swim must cross a kingdom of lakes to deliver one letter.
- The moon sends a bill to a small town for fourteen centuries of tides.
- A young cartographer of dreams must map one that rearranges itself whenever she looks.
A sci-fi short drama · A dragon fantasy drama · A first-contact drama.

## Comedy and family ideas

Comedy

- A town's only mime must give the emergency announcement when the loudspeaker breaks.
- A ghost who is showing a house to buyers tries to scare them away, then tries to sell it to the right family.
- A pet sitter accidentally takes the wrong dog to a royal dog show.
- An office goes on a team-building retreat that is running a murder-mystery weekend for another company at the same time.
- A tour guide with a made-up history of the town faces the one visitor who is a historian.
- Two neighbours' feud over a hedge escalates into a town-wide election.
Family and adventure

- Three siblings follow a treasure map drawn on their grandmother's tablecloth.
- A girl and a stray robot dog cross a flooded city to return a library book that is forty years overdue.
- A lost kite carries a message across a mountain range, passed from hand to hand.
- A lighthouse cat becomes the unofficial harbour master.
- A boy who is afraid of water must row to the island school on the first day.
- A caravan of retired circus animals refuses to give up the show.
A sitcom or comedy skit · An animated short film · A dinosaur adventure.

## The three-column idea mixer

Pick one from each column, or roll a ten-sided die three times. Ten by ten by ten is a thousand ideas.

| # | Who | Must | Against |
| 1 | a retired lighthouse keeper | deliver one last letter | a flood that closes the road |
| 2 | a night nurse | return something borrowed | a locked door |
| 3 | a teenage locksmith | keep a promise to a stranger | a blackout across the city |
| 4 | a failed magician | win one election | a rival who knows the truth |
| 5 | a museum guard | find a missing page | a witness who changed their story |
| 6 | a ferry captain | hide a secret for one night | a storm that cancels the last boat |
| 7 | a translator | get home before dawn | a language they cannot speak |
| 8 | a goat farmer | stop a wedding | a lie from long ago |
| 9 | a piano tuner | fix one clock | a clock that runs fast |
| 10 | a bus driver | be believed | a promise made to someone else |
Two rolls: a teenage locksmith must return something borrowed against a blackout across the city; a translator must be believed against a witness who changed their story. Both already have a person, a want and an obstacle, which is all a one-line brief needs.

## Title formulas

| Formula | Examples |
| The [job] of [place] | The Clockmaker of Low Street · The Lamplighter of Harrow Bay |
| Last [thing] to [place] | Last Ferry to Kessel · Last Train to Marrow Hill |
| [Number] [plural] | Nine Quiet Rooms · Forty Borrowed Hours |
| Don't [verb] the [noun] | Don't Wake the Harbour · Don't Answer the Lift |
| The [adjective] [noun] | The Borrowed Winter · The Patient Thief |
| [Person]'s [thing] | Ines's Last Prescription · Otto's Ropes |
| When the [thing] [verb] | When the Tide Forgets · When the Bells Stop |
| A [thing] for [person] | A Lantern for Pim · A Key for Mr. Lull |
| Beneath the [place] | Beneath the Salt Hotel · Beneath the Third Bridge |
| One word | Tether · Salt · Kessel · Hollow |
The promo cover paints your title into the poster, and a title of 12 characters or fewer works best, so keep a short version in your pocket. Search a title before you commit to it. Make a promo cover poster · AI movie poster generator.

## Sixteen plot twists

| Twist | What the audience realises | Plant it in act one as |
| The helper is the problem | The person helping has been steering the hero | A kindness that costs the hero something small |
| The narrator is not neutral | The story was told by someone with a reason | A detail the narrator gets wrong |
| The victim arranged it | The crime was staged by the one it happened to | A plan that is a little too neat |
| The villain is protecting something | The enemy blocks the hero to keep someone safe | The villain's one soft gesture |
| The missing person left on purpose | No one took them; they walked away | A packed bag nobody notices |
| The prize is the burden | Winning means inheriting the cost | A warning treated as a joke |
| Same place, different years | Two timelines are one place | An object that appears in both |
| The rule was a lie | The rule that kept everyone safe also kept them stuck | A rule no one has ever tested |
| The message was for someone else | The hero acted on a letter meant for a stranger | A delivery with a half-legible name |
| The enemy wants the same thing | The rivals' goals are identical | Both hands reach for the same object |
| The rescue is the trap | Getting out was what the villain wanted | An exit that was too easy |
| The answer was in the first scene | The clue sat in plain sight | An object shown early without emphasis |
| Nobody was there | One person was never real, or was already gone | A character no one speaks to directly |
| The hero made the villain | The hero's choice created the enemy | A small cruelty the hero forgets |
| The ending is the beginning | The last image is the cause of the first | The opening shot, repeated |
| The order was wrong | Events happened in a different order than shown | A change of season between two scenes |

## Turn an idea into an AI Create line

A good brief has five things: the genre, the lead and what they want, the conflict, the ending and the tone. Add them to an idea and paste it into AI Create; pick the length in SceneMixer (Mini for a single scene, Short for a handful of episodes) and press Start Writing. Writing the one-line brief · Choosing the script length.

| Idea | The brief |
| A retired piano tuner is hired to tune a piano in a house he swore never to enter again. | A quiet drama: a retired piano tuner is hired to tune a piano in the house he swore never to enter again, and the owner is the daughter he lost touch with; she wants him gone, he wants one chance to say sorry; he finishes the tuning and plays one note; warm, slow, bittersweet. |
| A hotel night manager notices that the same guest checks in every night under a new name. | A thriller: a hotel night manager notices that one guest checks in every night under a new name and always asks for room 9; her boss tells her to leave it alone; she opens room 9 and finds her own handwriting on the wall; tense, quiet, a final reveal. |
| A ghost showing a house to buyers tries to scare them away. | A comedy: a ghost who is showing a house to buyers tries to scare them away, then tries to sell it to the right family; the estate agent can hear him and nobody believes her; the house is sold to the family that needed it; gentle, absurd, a warm ending. |
Then read the script, approve the cast and read the storyboard, which is free on your first drama. A movie from a prompt · AI film script generator · AI story maker · Logline examples.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, English Manor; Costume Style Soft Everyday, Smart Casual; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the piano tuner, his daughter and the house; the piano is a prop that appears in every scene. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Five things in a brief | Example (the piano tuner) |
| Genre | A quiet drama |
| The lead and the want | A retired piano tuner who wants one chance to say sorry |
| The conflict | The owner is the daughter he lost touch with, and she wants him gone |
| The ending | He finishes the tuning and plays one note |
| The tone | Warm, slow, bittersweet |

## FAQ

### Is there an AI movie idea generator?

SceneMixer's AI Create turns a one-line idea into a script, a cast and a storyboard, and its Inspire Me button fills the box with an example to edit. This page gives you 48 ideas, a mixer, titles and twists to start from.

### How do I turn an idea into a movie?

Add a lead, a want, a conflict, an ending and a tone, paste it into AI Create, read the draft script, approve the cast and read the free storyboard before you render anything.

### Can I use these ideas?

They are starting points. Change the person, the place and the ending until the idea is your own.

### What makes a good movie idea?

A specific person who wants one thing, something in the way, and an ending you can picture.

### Can it generate movie titles?

The title formulas here give you a start. The promo cover on SceneMixer paints your title into three poster options.

### What is a plot twist generator?

A list of structures that change what the audience believes. The sixteen above each say what to plant in act one.

### Do I need to write a script first?

No. One line is enough for AI Create to draft the script. You can edit it, or paste your own.

### Is it free to try?

AI Create shows the price of the writing on the button before you start; new SceneMixer accounts start with 50 credits, and the first storyboard on your first drama is free.

Updated November 3, 2026


---

# AI Movies: How Long They Run, How They Look and How They Are Built

URL: https://scenemixer.com/guides/ai-movies/
What an AI movie can be today: real runtimes of the site’s sample films, the ceilings on a segment, an episode and a script, the quality tiers and what one costs.

# AI Movies: Length, Quality and How One Is Built

Quick answer: An AI movie today is a short film or a pilot, built from clips of up to 15 or 30 seconds joined into one piece, with the same cast in every shot and the lines spoken in the picture. The sample films on SceneMixer run from 1 minute 46 seconds to 3 minutes 30. One episode can hold up to 60 segments, which is 15 minutes at 15 seconds a segment and 30 minutes at 30 seconds, a Film has no episode cap, and a project can run to many parts. Quality is a tier you choose per segment, from 480P up. This guide gives the real numbers, what limits them and what they cost.

## What the sample films actually are

Every sample on the site was made end to end from a script by the same tools you would use, so their sizes are the honest answer to “how long can an AI movie be”. All 27 are a single episode, between 1 minute 46 seconds and 3 minutes 30, with four to eleven characters and one to four sets. Ten of them in English:

| Film | Look | Shape | Cast | Sets | Runtime |
| The Veil at Dawn | Realistic HD | 16:9 | 6 | 2 | 1:46 |
| Starfall Command | Realistic HD | 16:9 | 7 | 2 | 2:14 |
| The Eagle’s Vow | Realistic HD | 16:9 | 9 | 2 | 2:16 |
| The Last Envelope | Realistic HD | 16:9 | 6 | 2 | 2:20 |
| Rocket Pup | Feature 3D Animation | 16:9 | 7 | 3 | 2:20 |
| The Glass Floor | Realistic HD | 16:9 | 5 | 2 | 2:22 |
| Crimson Vow | Realistic HD | 16:9 | 4 | 1 | 2:23 |
| The Violin on Mercer St | Realistic HD | 9:16 | 5 | 1 | 2:28 |
| Neon Debt | Game Engine Realism | 16:9 | 7 | 2 | 2:32 |
| Rust City Knockout | Painterly Anime | 16:9 | 6 | 1 | 3:30 |
Read the table the way a producer would: this is what one pass at a short film looks like here. Roughly two to three and a half minutes, a handful of people, one or two places. It is not a limit. It is a size that holds together and that one person can finish.

## The ceilings, from one clip up to a project

| Unit | Ceiling | What sets it |
| A segment (one generation) | 15 seconds, or 30 on the long-take length | The Standard 15s or Long take 30s choice for the episode, and the tier, since older model families stop at 15 |
| An episode | 60 segments: 15 minutes at 15 seconds, 30 minutes at 30 | The segment limit. Short Drama episodes are also capped at 240 seconds and TV Series at 300; a Film has no cap |
| A script read | Under 10,000 equivalent characters in one pass, up to 100,000 in two (300,000 English characters) | The reading, which is refused rather than truncated above the cap |
| A project | Many episodes, sharing one cast and one set of locations | The length of the script you bring |

SceneMixer caps one continuous take at one segment, so a long shot in the sense of a single unbroken minute is the one thing this method cannot give. Everything else is joined. Each episode is composited on its own; to watch parts as one film, download each and join the files in any video editor. Composite and share an episode · Long AI video generator.

## How an AI movie is cut: Film, not Short Drama

In SceneMixer’s Step 1, the Type row decides how the script is divided and how each scene is written for the camera. For a movie you want Film: slow-paced and atmospheric, no cap on episode length, and the script treated as one piece divided into acts rather than as a run of episodes that each end on a hook. The type is fixed when the script is read, so choose before you continue. 16:9 Landscape is the recommended shape for film, and for the look you can pick Shot on Film for a cinema grade, Digital HD for a clean filmed look, Black & White Film for monochrome, or an animation group. Short Drama, TV Series or Film · How to make an AI short film.

## Quality: what you decide per segment

- Resolution is the SceneMixer tier you pick for each segment: 480P by default, 720P, 1080P, and a 2K tier on some models. Draft low, then regenerate the segments you keep on the tier you will deliver on.
- Sound is made in the same pass as the picture: dialogue spoken in the video, one voice per character, in any of 15 languages, and music scored for each episode.
- The cast is drawn once and used in every shot, so a face in the last minute is the face in the first.
- The file you download is the full-quality one; the player on the site streams a lighter copy.
What it still does badly is worth knowing before you plan a feature. Legible words in the picture are not drawn reliably. A single take longer than a segment is not possible. Wide shots of crowds and brawls tend to blur, so they are cut as close shots. And real faces from photos, famous characters and brand mascots are refused. What comes out · Writing fights, crowds and group scenes.

## What one costs

SceneMixer charges 10 credits a second for video on the default tier, and joining the segments is 1 credit a second. A 60-second episode comes to about 660 credits including the join, before retakes, so a 140-second sample film is about two and a third times that. Reading the script is 4 credits per 1,000 characters, English counted lighter, and editing text is free. At the other end, fifteen minutes of finished video, ten episodes of 90 seconds from a 30,000-character novel on the lowest tier, comes to about 10,026 credits, about $62 at the lowest credit price, nearly all of it video. A regenerated take is charged again, which is why the cheap steps come first. Where credits go in one episode · AI video cost calculator.

## How to plan one that gets finished

- Count the places. The samples use one to four. Every place is drawn once and reused, so fewer is steadier.
- Write what the camera can show. Actions and spoken lines; a thought becomes a look held or a voice-over.
- Make one sequence first. Parts of a film can be storyboarded and generated in any order, so test whether the look holds on one before you commit to the rest.
- Fix single shots, not scenes. In SceneMixer, edit the text of the shot and regenerate that segment.
- Join each part, then the parts. Composite per episode and join the files in your own editor if you want one file.
For a second part that should stand on its own, start a new project with the same look and use Pick existing to bring each returning character across. Making the next episode.

## FAQ

### How long can an AI movie be?

A SceneMixer segment is up to 15 or 30 seconds, an episode holds up to 60 segments, a Film has no episode cap, and a project can run to many parts. The sample films on the site run from 1:46 to 3:30.

### What quality are AI movies?

You choose a SceneMixer tier per segment, from 480P up to 1080P and a 2K tier on some models. Dialogue and music are made with the picture.

### Can an AI movie keep the same characters?

Yes. SceneMixer draws each character once and makes every shot from that picture, so faces stay the same across the film.

### How are AI movies divided?

As a Film, SceneMixer treats the script as one piece divided into acts, without the short-drama length cap or the end-of-episode hook. Each part is composited on its own.

### Can I make a feature-length movie?

A script up to 300,000 English characters can be read in one project, but every finished second is a paid generation, so start with a short, a pilot or a single sequence and judge the look before committing to more.

### Can an AI movie have real actors?

No. SceneMixer’s cast is generated or drawn by you. Photos that read as a real person’s face and famous characters are refused.

### Do AI movies have dialogue?

Yes. SceneMixer speaks lines in the video itself, one voice per character, in any of 15 languages.

### What does an AI movie cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join. A 60-second episode is about 660 credits before retakes.

Updated November 3, 2026


---

# AI Powered Creator Video Tool: Set Up a Vertical Series

URL: https://scenemixer.com/guides/ai-powered-creator-video-tool/
What an AI powered video tool must settle for a creator who posts vertical: the frame, the episode cap, segment length, a reusable cast and the finishing kit.

# AI Powered Creator Video Tool: Set Up a Short-Form Series

Quick answer: For a creator who posts short vertical video, an AI powered tool is only as useful as the number of format decisions it settles before you spend. In SceneMixer there are six: the frame, the episode type and its length cap, the dialogue language, the look, the segment length (15 or 30 seconds, never under 4) and the tier you render on. Four of them lock when the script is read, so this page lists them in the order they lock. It then covers how one cast carries across a series and what the finishing kit gives you for a post: cover, music switch, captions, badge and file.

## Six settings, in the order they lock

Short-form creators lose the most time to a decision they cannot undo. Here is what is fixed when, so you can make each choice once, on purpose.

| Setting | Where | Choices | Can it change later? |
| Frame | Step 1, Aspect Ratio | 16:9 Landscape (the default, marked Recommended) or 9:16 Portrait | No. It is fixed once you continue past step 1. For the other shape, start a new project with the same script. |
| Type | Step 1, Type | Short Drama (episodes up to 240 seconds), TV Series (up to 300 seconds), Film (no cap) | Not after the script is read. The price does not depend on it. |
| Dialogue language | Step 1 | One of 15; the lines are written in it from the start | Not after the read. Edit the lines, or start again with the same script. |
| Look | Step 1, Visual Style (plus Scene Style and Costume Style) | 21 visual presets; members can write their own | No. It shapes every sheet and every clip. |
| Segment length | Step 3, per episode | Standard 15s or Long take 30s | Per episode, before you generate its storyboard. Re-parsing re-cuts the episode and discards its clips. |
| Tier | The model button in the editor | From the cheapest tier up to a 2K tier | Yes, for every segment. Each segment keeps the tier it was made on. |

Everything in the first four rows is settled before the first picture is drawn; SceneMixer charges the read of your script first. 9:16 changes the framing the model is asked for, not just the crop: a vertical drama gets close framing that fills a phone, a landscape one gets wider staging. Setup images do not follow the frame: character sheets are always portrait and location and prop images always widescreen. Vertical or landscape and Short Drama, TV Series or Film cover the first two in full.

## A drop is a row of segments, not a number of minutes

A SceneMixer segment is one generation, and a finished post is several of them joined end to end. At the default Standard 15s a segment runs up to 15 seconds. Long take 30s allows up to 30 seconds, but only on the tiers whose cap is 30 seconds; a 30-second segment on a 15-second tier is refused before anything is charged. The floor is 4 seconds: the editor keeps a segment's total between 4 seconds and the episode's segment length. A segment length is a ceiling, not a target, so a 60-second post is at least four 15-second segments or two 30-second ones, and usually a few more, because the story decides where a segment ends.

The price is per second either way. A 60-second post on the default tier comes to 660 credits with the joining step included, about $4.11 at the lowest credit price. What the setting changes is the size of a retake: redoing a long take costs twice a standard segment, which is why 15 seconds is the default and the safer habit when a series is on a schedule. 15-second or 30-second segments.

The type caps how long an episode may run, not how long it has to be. Short Drama splits your script so that every episode opens on a hook and ends on a reversal or an open question; the 240-second cap is the longest an episode may run. Where a post should land inside it is a pacing choice you make in the script. Writing hooks and cliffhangers into your script.

## One show, many posts: how the cast carries over

A series is one SceneMixer project. The script is read once; the characters, locations and props are drawn once and shared by every episode, which is why a face in episode nine is the face from episode one. The order that saves money is to finish episode 1 end to end, watch it on your phone, fix any cast problem while it is cheap, and only then storyboard the rest. A problem in the cast found in episode 1 costs one regeneration; found in episode 8 it costs seven episodes of them. Making a multi-episode series.

When the story keeps going, Add an episode at the bottom of step 3 takes the continuation (pasted, uploaded, or written by AI), reads only the new pages and reuses the cast you already have. A second season or a spin-off can instead be a new project that brings characters, scenes and props over from the asset library with Pick existing: the name, description, images and voice move together, and it costs nothing on any plan. Reusing a character, scene or voice.

How many posts you can have in flight is set by how many generations run at once: 3 on a free account, 5 on Pro, 8 on Studio and 10 on Enterprise, with more above that. Extra generations queue and start by themselves. One segment takes the same time on every plan. The weekly arithmetic of that is in AI video creator: how one person plans a week.

## The finishing kit for a vertical post

Short-form posts need more than the video. This is what the tool hands you for each piece, and what it leaves to you.

| Piece | What you get | Good to know |
| Cover | A poster with the title painted in, in portrait 9:16 and landscape 16:9; 3:4 (a short-drama platform cover) and 1:1 can be made from the one you pick | The first set of three is free on your first drama; each extra shape is priced as one image. Keep the title to 12 characters or fewer and check its spelling on each option; the image model draws it. |
| Music | Every episode is scored automatically once its first shot is generated and mixed under the voices | A Music switch beside the player turns it off. Turn it off before you download to get the version without music and lay your own licensed track under it. There is no track picker. |
| Captions | On accounts that have bought credits or a plan at least once, a Subtitles switch burns your script's lines into the download, and a whole episode can also export an SRT file | The text is your script's, so invented names are spelled as you wrote them. On other accounts, add captions in the platform's own editor. |
| Badge | A small SceneMixer AI mark in the bottom-right corner of the picture, plus a line of AI-generated-content metadata in the file | Free accounts always carry it. Members get clips without it by default, and the switch affects clips generated afterwards. A share link always shows it. |
| File | Download gives the full-quality original; the site's player uses a lighter copy | An episode is joined from its segments at 1 credit a second. Nothing is posted for you: you upload the file. |

Covers, music and subtitles each have their own guide: promo cover posters, the built-in score and subtitles.

## Check it on a phone before you post

Your audience will watch on a phone, so check on one. The same SceneMixer account and projects are there when you sign in on a phone. The episode editor stacks into three views, preview, timeline and storyboard text; the model picker and badge switch sit in a bottom sheet. Download saves the original-quality file, and Share… hands it to the phone's own share menu. Play the downloaded cut full screen once before you upload it: that is the picture your viewers get. Using the site on a phone and posting to TikTok, Shorts and Reels go through the upload side.

## What it does not do for a creator

SceneMixer does not post. The share menu lists social sites for sending a link, but no platform receives an upload from here. It is not a timeline editor. The joined episode is made of hard cuts between segments; fades, wipes and speed ramps belong in an editor you download the file into. It does not take requests for music. The score is chosen for the episode, with no track picker and no volume control, so a trending sound or a licensed track goes on afterwards, in the platform's editor or your own. If a post needs any of those, plan for one more tool at the end.

## FAQ

### Can I make vertical videos for TikTok, Reels and Shorts with it?

Yes. Choose 9:16 Portrait in SceneMixer’s step 1 and every video in the project is made in that frame. The setting is fixed once you continue, so choose it before you read the script.

### Can I turn a landscape project into a vertical one later?

No. SceneMixer fixes the frame once you continue past step 1. Start a new project with the same script and pick 9:16 Portrait; the cast can come across from the asset library for free.

### How short can a clip be?

A SceneMixer segment can be as short as 4 seconds. A spoken shot needs enough seconds for its line, and the editor does not check that for you.

### Does a vertical video cost more than a landscape one?

No. SceneMixer prices video by the second, not by the shape of the frame.

### How long can one episode be?

SceneMixer caps Short Drama episodes at 240 seconds and TV Series episodes at 300; Film has no cap. An episode holds up to 60 segments. The cap is a ceiling, not a target.

### Does it post to the platforms for me?

No. You download the file and upload it yourself. A SceneMixer link is fine for sending a preview, but feeds favour videos uploaded natively.

### Can I keep the same characters in my next series?

Yes. In step 2 of a new SceneMixer project, open the card, press Edit, then Pick existing. The name, description, images and voice move over and it costs nothing.

### Will my free videos carry a mark?

Yes, a small SceneMixer AI badge in the bottom-right corner of the picture, plus a line of AI-generated-content metadata in the file. Members' new clips are badge-free by default and can switch it back on; a shared link always shows it.

Updated November 2, 2026


---

# AI Product Video Generator: A Finished Video From One Photo

URL: https://scenemixer.com/guides/ai-product-video-generator/
Upload one to four product photos, get a finished video with sound and an end card. Which product videos it makes (promo, demo, explainer, lifestyle) and which not.

# AI product video generator

Quick answer: An AI product video generator takes the photo and does the rest: it reads what the product is, redraws it as a clean studio shot, writes a short script with two to four spoken lines, builds a cast and a place, shoots 15, 30 or 60 seconds with the sound made alongside the picture, and ends on a card with your brand, product name, tagline and button. Of the usual kinds of product video it makes four well: lifestyle, demo, promo and a one-promise app or site explainer. It does not make a turntable spin, a customer testimonial or an edit of footage you shot. You are quoted one price before it starts, and charged no more than that.

One line for the box next to the upload: A collapsible silicone lunch box that fits in a coat pocket; for commuters who hate carrying a bag. Tone: Show how it works; 15 seconds; portrait.

## Five kinds of product video, and which this makes

| Kind | What the viewer gets | Made here? | How |
| Lifestyle | The product in someone's day | Yes | A day in the life |
| Demo | Open it, use it, see the result | Yes | Show how it works |
| Promo or commercial | A story or a gag that ends on the product | Yes | Epic, Funny, Twist or Warm |
| Explainer (app or site) | One promise: the moment of use, a real screenshot on the person's screen | Yes, one promise per ad | A clean screenshot, Show how it works |
| Food and drink close-ups | Steam, pour, cut | Yes | Mouth-watering close-ups |
| Turntable or 360 spin | The object rotating on white | No | The product is a studio shot placed in generated scenes |
| Customer testimonial | A real buyer's review | No | The cast is generated; there is no real customer |
| Edited footage you shot | Your own clips cut together | No | Uploads are references, not source clips |
SceneMixer’s eight tones are narrative recipes, not filters: each remakes the same product into a different kind of scene. The four pickers · An ad for an app or website.

## What happens between the photo and the video

- It reads the photo for free. Category, product name, any brand name legible on the label, the selling points worth filming, a suggested tone. If it cannot tell what the subject is, you learn that here, at no cost.
- It redraws the product as a clean studio shot (shape, colour, material and labels kept) and uses that one object in every scene.
- It writes the script: two to four spoken lines, a cast, locations, a storyboard of 15-second segments.
- It shoots every segment with dialogue, ambience and music generated together with the picture.
- It finishes on the end card with brand, product name, tagline and call to action.
The one figure on the button is a ceiling, not an estimate: each stage freezes and charges its actual price, so the total lands at or below the number you saw. What you actually pay · The whole flow, step by step.

## A source photo that works

The product alone, large in the frame, in even light, on a plain surface. Up to four photos of the same product: the first is the hero the end card is built around; the others are angles and details. A lifestyle photo with the product small in a busy scene reads badly as a first image. A screenshot on SceneMixer is used exactly as uploaded and never redrawn, and must be roughly phone- or computer-shaped with no real face in it. Which photos work.

## Promo, commercial, explainer: the same machine, three lines

| You want | One line to write | Tone, length |
| A promo for a launch | A rechargeable hand warmer for people who cycle in winter; launching this month. | Epic, 15 seconds |
| A commercial with a turn | A reusable coffee cup that seals; for people who always leave it on the bus. | Twist, 30 seconds |
| An explainer for an app | A grocery app that delivers in twenty minutes; for busy parents. (a screenshot of the result screen) | Show how it works, 30 seconds |
When the line and the photo disagree, SceneMixer lets the line win, so use it to name the product, the buyer and where it ships from. Claims such as best seller are removed from the script. Writing the one-line brief.

## Where the finished video goes

| Placement | Format | Length | Guide |
| Your product page or a website banner | Landscape 16:9 | 15 or 30 seconds | — |
| Instagram, TikTok, Stories | Portrait 9:16 | 15 or 30 seconds | TikTok Shop |
| A marketplace listing | Check the marketplace's page | Short | Amazon · Etsy |
| A store screen or pre-roll | Landscape 16:9 | 30 or 60 seconds | — |
The same four photos can make several versions with different tones and lengths; each is a separate ad with its own quote, and tone does not change the price. On SceneMixer, free accounts carry a small AI label in the corner of each video; a paid membership turns it off by default. The AI label switch · AI ad video generator.

## What it will not do

- No readable lettering in the picture. The ad is made without on-screen text; the name, tagline and button are on the end card.
- No real person. People in your photos are removed; the cast is generated; no face swap, no voice cloning.
- No absolute claims. Superlatives and regulated phrasing are stripped from the script, from the tagline and from the end card. It is a wording filter, and the claims you make are yours.
- No longer than 60 seconds, built from four 15-second segments.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Picker | Options | Changes the price? |
| Vibe | Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life | No |
| Length | 15, 30 or 60 seconds (one, two or four segments) | Yes |
| Format | Portrait 9:16 or Landscape 16:9 | No |
| Quality | 480P, 720P or 1080P | Yes |

## FAQ

### How do I make a product video from an image with AI?

Upload one to four photos of the product on SceneMixer’s Ad Video tab, write one line, pick a tone, length and format, and press Make my ad. The script, cast, shots and end card are made for you.

### Can it make a promo video or a commercial?

Yes. Epic, Funny, Twist and Warm tones on SceneMixer turn the same product into different kinds of promo or commercial, at 15, 30 or 60 seconds.

### Can it make an explainer video?

For an app or a website, yes: one promise per ad, with your real screenshot on the person's screen. SceneMixer does not draw a step-by-step interface walkthrough, because video cannot render menus and text reliably.

### Does it make a 360-degree product spin?

No. The product in SceneMixer is a studio shot placed in generated scenes; there is no turntable.

### Will the product look the same in every shot?

Yes. SceneMixer redraws it once from your photo and uses it as the same object in every scene.

### Can I use footage I filmed?

No. Everything on SceneMixer is generated from the storyboard; uploads are references, not source clips.

### How long can the video be?

15, 30 or 60 seconds on SceneMixer, made from one, two or four 15-second segments, plus the end card.

### How much does it cost?

One figure is quoted on SceneMixer before it starts and it is a ceiling. Length and quality set it; tone, format and the line do not.

### Is it free to try?

Reading the photo on SceneMixer is free, new accounts start with 50 credits, and the first 15 setup images are free. Ads are never the free first project.

Updated November 3, 2026


---

# AI Real Estate Video: A Listing Tour From Four Photos

URL: https://scenemixer.com/guides/ai-real-estate-video/
Make a real estate video with AI from listing photos: rooms kept as shot, people removed, four 15-second segments for a 60-second tour, an end card for the viewing.

# AI real estate video

Quick answer: Add up to four photos of the property and the ad is set inside them: the rooms are kept as you shot them, the people in the photos are removed, and a generated cast moves through the place. A segment can only be one place, so a 60-second tour of four photos reads naturally as four rooms, one per 15-second segment, ending on a card with your agency name and Arrange a viewing. Treat it as a mood film that sits beside the real photos and the floor plan, not as a measured walkthrough; price, size and address go in the listing and on the end card, not in the picture.

One line for the box next to the upload: A two-bedroom flat by the park with a south-facing living room and a new kitchen, ready to move in; for first-time buyers. Tone: A day in the life; 60 seconds; landscape for the listing page.

## What the ad does with a property photo

Products are cleaned onto white so the ad can put them anywhere; a place cannot be treated that way, because cut out, a room loses its depth and scale and a shopfront loses its sign. So a property photo is kept as it is and becomes the first scene the ad happens in. With up to four photos, every extra becomes another set the shots can move through. The people in your photos (the owners, the agent) are removed before anything is made, and the ad's own cast is generated into the rooms. An ad for a store or a property covers the general case; this page is about listings.

## Four photos, four rooms, four segments

An ad on SceneMixer is made in 15-second segments, one, two or four, and a segment is one place and one timeline. That is why a four-photo, 60-second ad can be a tour: each segment has room for one set.

| Photo | What to shoot | What the segment is for |
| 1 (the hero) | The best room, or the front of the building, straight on, in daylight | The opening; the end card is also built around this image |
| 2 | The main living space from the doorway, curtains open | Life in the room: a coffee, a window, a conversation |
| 3 | The kitchen or the feature room (a terrace, a view, a garden) | The reason someone buys: a meal, a morning |
| 4 | A bedroom, or the street and the park outside | The close: rest, the neighbourhood |
The first photo is the featured subject and fills the frame; the others become the sets the shots move through, and the script decides how each is used. If you need the rooms in a fixed order, choose Take the wheel: you pay for the cleanup and the script, then land in the editor with the ad already set up and can rewrite the script and edit any shot. Hand it to AI or take the wheel · One place per segment.

## Photographing a room for it

- Daylight, curtains open, one clear view of the room from the doorway; tidy surfaces, because the room is kept as it is.
- Straight on and level, not tilted: the room is kept as shot, so a level, undistorted photo gives the ad a cleaner set to start from.
- People are fine in the photo: the composer says People will be removed and they are; the room stays.
- No floor plans or screenshots as photos. Show the plan in the listing, next to the video.
- SceneMixer takes JPG or PNG, under 10 MB each, of a property you have the rights to use.
Which photos work for an ad.

## Which tone, which length, which button

| Listing | Tone | Length and format | End card button |
| A flat or house for sale | A day in the life, Minimal | 60 seconds, landscape for the listing page | Arrange a viewing |
| A rental | A day in the life | 30 seconds, portrait for stories | Book a viewing |
| A holiday let | Warm, Epic for a view | 30 or 60 seconds | Book your stay |
| A just-listed teaser | Minimal | 15 seconds, portrait | See the listing |
| An open house | Warm | 15 or 30 seconds | Open house Sunday |
| A new development | Epic, Minimal | 60 seconds | Register your interest |
Length and quality move the price; tone and format do not. Ads on SceneMixer are never the free first project. Length, format and quality.

## One tour, several end cards

The end card is the one place text is readable: your brand, a product name (here the property: Two-bed flat, Maple Park), a tagline and a button. After the tour is made, Edit end card and Rebuild final cut re-join the same shots with a new card without reshooting. The same shots can open with Just listed, then Open house Saturday, then Offers invited. Keep the price and the address in your listing and caption, because text in the picture is not drawn reliably. The end card · Rebuild the final cut.

## What it is and is not

- The ad from SceneMixer is a staged film of your photos. The rooms are generated from them, so small details can differ from the real room. Put it next to the real photos and the floor plan so a viewer can check.
- It is not virtual staging. It does not furnish or rearrange a room in your photo; the generated people live in the room as it is.
- It does not use your face or voice. The cast is generated; there is no face swap and no voice cloning, so the agent appears on the end card by name, not on screen.
- It can speak in the buyer's language. The spoken lines and end card follow the language you choose; a second language is a second ad from the same photos.
- It will not say the best. Superlatives and unverifiable claims are removed from the script. Making AI video look realistic.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| For the listing | Do | Do not |
| Photos | Up to four of one property, daylight, hero first | Floor plans, screenshots, tilted wide-angle shots |
| The line | Who it is for and what is visibly true | The best, a dream home, guaranteed |
| The picture | Rooms as shot, a generated cast | Readable prices or addresses |
| The end card | Agency, property name, a button, edited per stage | Reshooting for each stage of the sale |

## FAQ

### How do I make a real estate video with AI?

Add up to four photos of the property on SceneMixer’s Ad Video tab, write one line about who it is for, pick a tone, length and format, and press Make my ad. The rooms are kept as shot and the ad is set inside them.

### Will the video show my actual property?

Yes, in the sense that SceneMixer keeps your photos and uses them as the sets; people in them are removed. It is a staged film made from your photos, so check small details against the real rooms.

### Can I make a video tour of several rooms?

Yes. Up to four photos become up to four sets on SceneMixer; a 60-second ad has four segments, each in one place.

### Can I control the order of the rooms?

The first photo is the hero; for a fixed order choose Take the wheel in SceneMixer, where you can edit the script and each shot.

### Can the agent appear in the video?

No. SceneMixer removes the people in photos and generates the cast; the agency is named on the end card.

### Can I put the price and the address on screen?

SceneMixer does not put them in the picture. Put them in the listing, the caption and the end card within its limits.

### Can I change the end card for an open house?

Yes. Edit the end card in SceneMixer and rebuild the final cut; no reshoot.

### Can I make it in Spanish or another language?

Yes. SceneMixer’s spoken lines and end card follow the language you choose, so a second language is a second ad from the same photos.

### What does a listing video cost?

One figure is quoted on SceneMixer before it starts and it is a ceiling; length and quality set it. A 15-second teaser is the cheapest, a 60-second tour the most.

Updated November 3, 2026


---

# AI Short Drama in 15 Languages: Interface and Dialogue | SceneMixer

URL: https://scenemixer.com/guides/ai-short-drama-15-languages/
What "15 languages" means in SceneMixer: the interface, the storyboards you edit and the dialogue your cast actually speaks in the video. No dubbing pass.

Guide · Languages

# AI Short Drama in 15 Languages (2026): Interface, Scripts and the Dialogue Your Cast Actually Speaks

Published September 12, 2026

Quick answer: SceneMixer works in 15 languages on three layers at once. The interface is available in English, Chinese, Portuguese, Korean, Spanish, Japanese, Indonesian, Russian, French, Italian, Dutch, Arabic, Turkish, German and Ukrainian. The storyboards, synopses and character descriptions you read and edit are shown in your interface language. And the spoken dialogue of the generated video is set per project to any of those 15 languages: the video model performs the line in that language while it renders the shot, so there is no separate dubbing step and no lip-sync patch afterwards.

## The three layers, and which one you are usually asking about

| Layer | What it covers | Where you set it |

| Interface | Menus, buttons, wizard steps, error messages, emails — 15 languages | Language menu in the top bar; first visit follows your browser |

| Working documents | Script analysis, episode outlines, storyboard scripts, character and scene descriptions | Follows the interface language automatically; edits you make in your language sync back to the production script |

| Spoken dialogue | The lines the cast says out loud in the generated clips | Step 1 of the project wizard, "Dialogue language", before the script is parsed |

Most questions about "multilingual AI video" are really about the third layer. A translated interface is table stakes; a translated storyboard is convenient; what sells a series in a new market is a cast that speaks the audience's language natively, with the timing and mouth shapes of that language, not a voice track glued on top of an English performance.

## How the dialogue layer works

- One production script, one dialogue language. The pipeline keeps a working script in English or Chinese for the machine steps (analysis, storyboarding, consistency checks). When you pick a dialogue language, the outline is written with the lines in that language and every storyboard copies them verbatim — the storyboard step is not allowed to rewrite dialogue.

- The video model performs the line. Audio and picture are generated together, so pacing, breath and mouth movement come from the same pass. Nothing is dubbed afterwards.

- SceneMixer computes timing per language. Shot lengths are derived from how fast people actually speak: about 3 words per second for English and about 5 characters per second for Chinese, with the other languages mapped onto those scales. A line that needs 4 seconds gets a 4-second shot.

- A guard before you pay. If a segment's lines are not in the project's dialogue language (for example after a manual edit), the submit is stopped and you are asked to confirm before any credits are frozen.

- Same price in any language. The dialogue language does not change the credit cost of a segment; the translated working documents are free.

## What to expect per language (checked 2026-09-12)

| Language | Dialogue quality | Notes |

| English, Chinese | Baseline | The production languages; the general sample series on this site use one of them, and the language samples below add one native-dialogue series per language |

| Portuguese, Korean, Spanish, Japanese, Indonesian | Native-level | Verified with native speakers on multi-line scenes |

| Russian, French, Italian, Dutch, Turkish | Native-level | — |

| Arabic | Good | Occasional single-word pronunciation flaws; regenerate the segment if one lands on a key line |

| German, Ukrainian | Good, with a caveat | Long sentences can produce a mispronounced verb ending or past-tense form; the storyboard prompt already keeps lines short for these two, and a flagged segment can be regenerated on its own |

These are working observations from SceneMixer's own test scenes, not vendor claims. When a line comes out wrong, the fix is local: rewrite that segment's script or simply regenerate it, and you pay only for that segment.

## Localizing more than the words

A script written for one market often needs more than translation. SceneMixer's target-region setting adapts the cast's names and ethnicity, the street scenes and props, and the language of any on-screen text to a Western, East Asian, Southeast Asian, Latin American or Central Asian audience during script analysis, so the Korean-dialogue version of your story is not a Western cast reading Korean. See how to choose the dialogue language and localize the cast for the workflow, and native dialogue vs. dubbing for why the performance is generated rather than dubbed.

## Hear it: one sample series per language

Reading about dialogue quality is one thing; hearing a cast perform a scene in your language is another. For each of the 13 languages beyond English and Chinese, one complete series was produced by the same SceneMixer pipeline with the dialogue language set in Step 1 and the target region matched to that market, so the names, streets and on-screen text belong to it too. Open any of them to see every segment's storyboard next to its video, or browse the full samples page for all 15 languages.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## Inside SceneMixer

- Step 1: paste or upload the script, pick the visual style, aspect ratio and the dialogue language. On the international site the interface follows your browser; the dialogue language is yours to choose.

- Step 2: character sheets, scene plates and props are generated; their descriptions appear in your interface language.

- Step 3: storyboards per episode, then video per segment. Each segment can be regenerated or its script rewritten; version history keeps every take.

- Ads: the same language layers apply to the ad-video workflow, including the end card text.

## FAQ

### Does SceneMixer dub the video, or does the cast speak the language?

The cast speaks it. The video model generates picture and audio together with the line already in the chosen language, so pacing and mouth movement match that language. There is no dubbing pass and no separate voice track to align.

### Which languages can the dialogue be in?

Any of the 15 interface languages on SceneMixer — English, Chinese, Portuguese, Korean, Spanish, Japanese, Indonesian, Russian, French, Italian, Dutch, Arabic, Turkish, German and Ukrainian. The language is set per project in Step 1, before the script is parsed.

### Does choosing another dialogue language cost more credits?

No. A segment on SceneMixer costs the same in any language, and the translated storyboards and descriptions you read in your interface language are free. You only ever pay per generated segment.

### Can the interface be in one language and the dialogue in another?

Yes. A producer working in German can make a series whose cast speaks Indonesian: SceneMixer’s interface and working documents follow the interface language, the spoken dialogue follows the project setting.

### Can I watch a sample in my language before I start?

Yes. SceneMixer’s samples page for 15 languages lists one finished series per language, each with the cast speaking that language natively and every segment's storyboard shown next to its video. Open the one in your language to hear the timing and mouth shapes before you spend any credits.

### What if a line is mispronounced?

Rewrite that segment's script or regenerate the segment; only that segment is charged again on SceneMixer. German and Ukrainian are the two languages where an occasional mispronounced word ending shows up on long sentences, so keep lines short there.


---

# AI Short Drama Character Consistency: 3 Layers | SceneMixer

URL: https://scenemixer.com/guides/ai-short-drama-character-consistency/
Why faces drift between shots in AI short drama, and the three layers that lock them: reference sheets, per-shot image references and storyboard rules.

Guide · Production

# AI Short Drama Character Consistency (2026): Why Faces Change Between Shots, and the Three Layers That Lock Them

Published September 5, 2026 · Updated November 3, 2026

Quick answer: Character consistency in AI short drama means one character keeps the same face, the same wardrobe and props, and the same art direction across the dozens to hundreds of shots in a series. In 2026 it is the first reason platforms and viewers reject AI-made series. It is not solved by prompt wording. It is locked at three separate layers: the character sheet (white background, directional light, identity features first), the scene plate (empty set, three planes of depth, light source described rather than where the light lands), and shot-to-shot continuity inside the storyboard (actions and positions carried over, never naming people who are not in frame). Face drift has three root causes and each maps to one layer: text descriptions get diluted, references carry baked-in light, and shots have no anchor between them. Below: what to do at each layer, a reference-input table, and a 5-shot acceptance checklist.

## What "consistent" means: three layers

- Identity features: face shape, brows and eyes, nose and lips, skin tone, hair, build, apparent age, plus one instantly recognisable anchor (a scar, a mole, a piece of jewellery, a haircut). Carried by the character sheet.

- Wardrobe, makeup and props: what this character wears and holds in this scene, and what the set contains. Carried by the per-character costume description from script analysis and by the scene plate. Costume is per character and per social class, not one style tag for the whole series.

- Light and art direction: colour discipline, materials, set-dressing density, where the light source is. Carried by the scene plate and by the lighting lines in each shot. The reference images themselves must stay neutral.

## Why faces drift: three root causes

| Symptom | Root cause | Layer |

| The same character gets a new face every few segments and slowly turns generic | Diluted description: the face description sits in the middle of long action and technical instructions, and the model falls back to its "attractive face" prior | Layer 1: identity is not first in the sheet prompt, or the sheet is not attached as a reference on every generation |

| Right face, but light and colour temperature jump between shots of the same scene | Baked-in light: the sheet or the plate carries dramatic lighting that every shot inherits and then fights with the storyboard's own lighting | Layer 2: references are not neutral |

| Positions, actions and props do not connect inside one segment; people who are not in the scene walk in | No anchor between shots: continuity is not written, or the storyboard names absent people in negative phrasing ("X exits frame") | Layer 3: continuity |

## Layer 1: the character sheet

The sheet exists for recognisability, not mood. Four hard rules:

- Pure white background, as its own clause. State it as a standalone constraint and repeat it in the negative list. Buried in another paragraph, it gets overridden.

- No flat light. "Even lighting, no shadows" is the recipe for a plastic face: flat light erases skin texture and bone structure. Write one directional key light at roughly 75 degrees with fill and rim, and ask for shadow under the nose and cheekbone, a catchlight, and gradual falloff.

- Identity first. Fixed prompt order: subject identity features → identity lock (no influencer face, no smoothing of distinctive features) → technical block → framing constraints. Script analysis produces very detailed appearance text (bone structure, moles, scars); it has to lead, not sit in the middle.

- Face > posture > costume, and the expression should act. Declare that this is an actor's look test, not an e-commerce outfit shot; forbid pulling back to show the outfit at the cost of the head. Do not write "neutral expression": that deletes the very quality (a blade-like stare, say) that makes the character recognisable.

Known failure: characters in long coats or robes get swallowed by the garment in a 9:16 full-body frame, the coat filling seventy percent of the image and the face shrinking. That is a physical limit of the framing. The fix is a separate half-body sheet that carries the face, at the cost of one more image per character.

In SceneMixer the appearance field produced by script analysis becomes the subject of the sheet prompt directly. Each character gets one 9:16 full-body sheet on white, with face and half-body crops cut automatically, and every later video generation attaches that character's references. Before judging sheet quality, confirm which image model produced it: output size and prior differ a lot between models, and the evidence is the image dimensions and the task record, not a guess.

## Layer 2: the scene plate

A scene image is an empty "plate" the video model references. The standard is again neutrality:

- No people. Express scale with the building's own parts (a door two people tall, a table one arm wide).

- Three planes of depth. Foreground, mid-ground and background each with explicit materials and dressing, so the model has room to place a camera.

- Describe where the light comes from, not where it lands. Write "a warm-white pendant throws soft light down from upper right", not "light brightens the tabletop". Do not write a shadow terminator line, and avoid saturated colour words: 3200 K on skin is warm white, and "yellow" gets you yellow paint.

- No atmosphere grading. No mood grading, silhouettes, fog or light beams, and none of the old trigger words (cinematic, dramatic, 8K). Fog in the plate appears in every shot of that scene.

Art direction on SceneMixer is a project-level setting applied once, at script analysis: it shapes every scene description and storyboard, and is never appended to the character sheet or any other image prompt, otherwise the white-background sheet gets lit like the set.

## Layer 3: continuity in the storyboard

- Long segments, short shots. One segment is one generation call and should run close to the 15-second ceiling; raise pace by shortening individual shots, not by slicing the segment. Every extra segment restarts positions from a new first frame and throws away continuity.

- Carry over inside the segment. Where the character sits, what they hold and what they were doing at the end of the previous shot is where the next shot starts.

- Only write who is in frame. Never name absent people. "Mark exits" or "Anna is not present" raises the chance they appear, because nouns inside a negation act as attractors.

- Lighting by cause. Same rule as the plate. Light position, medium and grade each stated once, with no repeats and no conflicts.

## Reference inputs by model

| Model | Reference inputs | Clip length | Notes | Checked |

| Seedance 2.0 (Volcano Engine Ark) | Text plus image, video and audio references; images can be first frame, last frame or character reference | 4–15 s | Native audio; images up to 30 MB and 300–6000 px per side; one of the video paths in SceneMixer | 2026-09-05, per Ark API docs |

| Wan 3.0 (Alibaba Cloud Model Studio) | Up to 10 reference images, 5 reference videos and 5 reference audios per request; prompt up to 20,000 characters | 2–30 s | 480P / 720P / 1080P; failed generations are not billed; the other video path in SceneMixer | 2026-09-05, per Model Studio API docs |

| Other video models | Most accept image references or first/last frames | Varies | Limits change often; whatever the model, the white sheet plus neutral plate method is the same | 2026-09-05 |

Only verified cells are filled; blanks are blanks, not guesses. What decides consistency is not whether the cap is 9 references or 50, but whether the few you attach are neutral.

## Acceptance checklist: five shots of the same character side by side

- ☐ Face shape, brows, eyes, nose and lips read as one person across all five.

- ☐ The visual anchor (scar, mole, jewellery, haircut) is present and in the same place in every shot.

- ☐ Within one scene, light direction and colour temperature match; no warm shot next to a cold one.

- ☐ Costume and hand props are continuous inside the segment, nothing appears or vanishes.

- ☐ No absent character walks in; no second "lead" appears.

- ☐ Skin has texture and gradual shading, not a flat plastic surface.

When a shot fails, go back to the layer it maps to and fix that layer. "Strengthen the prompt" is not a fix.

## Further reading

How the script itself sets up consistency (one name per location, no more than three or four sets per episode) is in Vertical Short Drama Script Format and Hook Structure. Disclosure and likeness rules are in AI Short Drama Compliance Checklist 2026. To watch the method on a finished series, open the SceneMixer sample dramas: every one shows the cast sheet and each segment's script and video.

## FAQ

### Why do character reference images need a white background?

Because whatever light and background are baked into the reference is inherited by every shot that character appears in. A sheet shot under cold night light makes daytime scenes drift cold; a prop in the background gets repainted as part of the character. A white background with one directional key light at about 75 degrees only records the structure of the face and carries no scene information, so any shot's lighting can override it.

### How do I fix inconsistent characters in AI video?

Diagnose by layer before touching prompts. Different faces across shots is a character-sheet problem: check for flat lighting or a face that is too small in frame. Right face but jumping costume light is the scene plate or the shot lighting fighting each other. Positions and actions that do not connect inside one segment is missing continuity in the storyboard. Each layer has its own fix; rewriting one big prompt makes all three worse.

### Can I keep a character consistent just by describing the face in every prompt?

No. Across dozens of generations the description gets diluted by each segment's action, setting and lighting instructions and the model falls back to a generic attractive face. The stable method is to attach the same character sheet as a reference image to every generation and let the text carry only that segment's action and emotion.

### How many reference images does each character need?

At minimum one 9:16 full-body sheet on white. Lead characters benefit from an additional face crop for close-ups. Characters in long coats or robes tend to be swallowed by the costume in a full-body frame, so a separate half-body sheet that carries the face is worth the extra image.


---

# AI Short Drama Compliance Checklist 2026 | SceneMixer

URL: https://scenemixer.com/guides/ai-short-drama-compliance-2026/
Disclosure labels, EU AI Act Article 50, YouTube and TikTok rules, likeness and copyright: a 2026 compliance checklist for publishing AI short drama.

Guide · Compliance

# AI Short Drama Compliance Checklist (2026): Disclosure Labels, EU AI Act Article 50, Platform Rules, Likeness and Copyright

Published September 5, 2026 · Updated September 20, 2026

Quick answer: An AI-generated short drama published in 2026 has four compliance surfaces: disclosure (the EU AI Act's Article 50 transparency duties apply from 2 August 2026, and YouTube and TikTok already require AI-content labels), likeness and voice rights (no real person's face or voice without written permission), copyright (AI output without human authorship is not protected in the US, so the human-authored layers are what you own), and platform content rules (each app's own prohibited-content list). None of it needs a lawyer to get right for a typical series; it needs a checklist run before upload. Below: the disclosure rules by jurisdiction and platform, the likeness and copyright positions with sources, and a 12-point self-check. This is general information, not legal advice; rules change, and every source below carries the date we checked it.

## Disclosure: who requires a label, and where

| Rule | Who it binds | What it requires | Source (checked 2026-09-05) |

| EU AI Act, Article 50 (transparency obligations), applicable from 2 August 2026 | Providers of generative systems and deployers who publish deepfakes; in practice the studio publishing the series in the EU | AI-generated or manipulated image, audio or video that appreciably resembles real persons, places or events must be disclosed as artificially generated; providers must mark outputs in a machine-readable way | Regulation (EU) 2024/1689, Art. 50 |

| YouTube altered or synthetic content disclosure (from March 2024) | Every uploader | Disclose realistic content that was meaningfully altered or synthetically generated; YouTube shows a label, more prominently for sensitive topics | YouTube Help |

| TikTok AI-generated content policy | Every uploader | Label AI-generated content that shows realistic scenes or people; TikTok also auto-labels media carrying C2PA content credentials | TikTok Transparency Center |

| Short-drama apps (ReelShort, DramaBox and peers) | Studios and creators submitting content | Governed by each app's content policy and studio agreement rather than public law; expect a declaration of AI use and a ban on real-person likenesses in the submission terms | Check the current terms of the app you submit to; they are not published as a single public document |

### How to place the visible label

- Opening frame: a legible "AI-generated" notice on the first frames of every episode, at least caption size and at least two seconds long, is the common practice; a notice only in the end credits is not a disclosure.

- Upload declaration: tick the platform's AI-content option on every upload; the platform's label is what moderation checks first.

- Metadata: keep whatever content credentials or generation metadata the tool writes. Re-encoding or re-exporting through an editor that strips metadata removes the automatic layer, and then the visible notice is your only evidence of compliance.

In SceneMixer, the visible watermark on composited episodes is controlled by the "AIGC watermark" switch in settings and is on by default. If your publishing platform adds its own label, you can switch the in-app watermark off to avoid a double label.

## Likeness and voice

- Faces: generating or editing a real person's likeness without written permission exposes you to right-of-publicity claims (state law in the US, personality rights in the EU) and, for realistic video, to the deepfake disclosure rules above. This applies to living and deceased public figures alike. SceneMixer's character sheets are generated from the written description that script analysis produces, and real photographs are not accepted as character references. That is a compliance posture, not a limitation.

- Voices: cloning an identifiable voice is treated as a likeness issue in several US states with digital-replica statutes passed since 2024 (Tennessee's ELVIS Act, March 2024, is the best known) and in the platform rules. Use a preset voice library or your own recordings.

- Source material: adapting a novel, comic or existing series requires the adaptation right from the rights holder; a paid fan-fiction adaptation is still an adaptation.

## Copyright: what you own in an AI series

The US Copyright Office's January 2025 report on the copyrightability of AI-generated material restates the line: works produced by AI without sufficient human authorship are not protected, while human contributions, including the script, the selection and arrangement of generated shots, and substantive human edits, can be (copyright.gov/ai, checked 2026-09-05). For a series that means the protectable layer is your script, storyboard and edit decisions. Keep them on record, dated, so the human authorship of the series is provable if it is copied.

## 12-point self-check before upload

- ☐ 1. AI-generated notice visible on the opening frames of every episode.

- ☐ 2. Platform AI-content declaration ticked on every upload.

- ☐ 3. Export pipeline preserves content credentials and metadata.

- ☐ 4. No real person's face used as a reference or generated; no cameo of a public figure.

- ☐ 5. No cloned identifiable voice; presets or licensed recordings only.

- ☐ 6. Adaptation rights for any source novel or IP in writing.

- ☐ 7. Music, fonts and stock elements licensed for the platforms you publish on.

- ☐ 8. Content checked against the app's prohibited-content list (sexual content, graphic violence, hate, self-harm, minors in unsafe contexts).

- ☐ 9. Script, storyboard and edit decisions archived with dates (your protectable layer).

- ☐ 10. EU release: Article 50 disclosure satisfied for realistic scenes, people or events.

- ☐ 11. Paywall and purchase prompts comply with the app's consumer rules for the markets you sell in.

- ☐ 12. A contact for takedown or complaint requests is published with the series.

## Next

Compliance is the entry ticket, not the ceiling. For the craft: Vertical Short Drama Script Format and Hook Structure covers writing for the 2026 apps, and Character Consistency covers keeping the lead's face through sixty episodes.

## FAQ

### Do I have to label an AI-generated short drama as AI-made?

In most of the markets that matter, yes. In the EU, Article 50 of the AI Act applies from 2 August 2026 and requires AI-generated or manipulated video that resembles real people, places or events to be disclosed as artificially generated. YouTube requires creators to disclose realistic altered or synthetic content, and TikTok requires labelling AI-generated content and auto-labels media carrying content credentials. Even where no law applies, an unlabelled realistic AI series is one report away from removal.

### Can I copyright an AI-generated series?

Only the parts a human authored. The US Copyright Office's position, restated in its January 2025 report on copyrightability, is that material generated by AI without sufficient human authorship is not protected, while the human-written script, the selection and arrangement of shots, and human edits can be. Keep your script, storyboard and edit decisions on record; they are the protectable layer.

### Is it legal to use a real actor's face or voice in an AI drama?

Not without written permission. Right-of-publicity and likeness laws in most US states, and personality rights in the EU, cover a person's face and voice; several US states have added explicit digital-replica rules since 2024 and the EU AI Act treats such deepfakes as content that must be disclosed. The safe practice is to generate characters from written descriptions and never use a real person as a reference.

### Where does the disclosure label have to appear?

Use both layers: a visible notice in the video itself (opening frame or persistent caption) and the platform's own AI-content declaration when you upload. Keep any content credentials or metadata the generation tool writes; re-encoding that strips metadata can drop an automatic label, which then makes the visible notice the only thing protecting you.


---

# AI Short Drama Genres: How to Make Each One, and What Works Best

URL: https://scenemixer.com/guides/ai-short-drama-genres-how-to-make-each-one/
Werewolf, billionaire, revenge, horror, telenovela, sci-fi and more: what each genre needs, the technique that makes it work, and which genres AI handles best.

# AI Short Drama Genres: How to Make Each One

Quick answer: Every genre is made the same way in SceneMixer — a script, a cast with reference images, a storyboard, video segment by segment, a composite — and every genre is priced the same way. What changes is the handful of choices that make a genre look right: the scene and costume styles in step 1, which looks become states, how the signature scene is written into shots, and what to keep off screen. The genres AI video handles best are the ones built on faces and dialogue in a few recurring rooms: romance, revenge, family drama, mystery, comedy. The hardest moments in any genre are large crowds, long continuous fights, readable text and gore, and each genre guide below shows how to work around them.

Pick the genre closest to your story below, read its guide, and copy its example brief into AI Create to see a first draft.

## Every genre, and the one thing that makes it work

| Genre | The technique that matters most |
| Werewolf romance | The wolf form as a state of the character; the shift filmed in two shots |
| Vampire romance | Hunger as a state; blood implied, never shown; candlelight written as a source |
| Billionaire / CEO romance | Wealth described by look, never by brand; the gala gown as a state |
| Mafia romance | Power through rooms and silence; a gun is a prop, not an act |
| Revenge and second chance | The past-life look picked shot by shot; evidence as recurring props |
| Telenovela and soap opera | Native Spanish or Portuguese dialogue; twins as two characters |
| Regency and Victorian romance | Period sets and costumes; the ball filmed in two-shots |
| Xianxia and wuxia | Spells written as light; the drawn sword as a prop state; duels as exchanges |
| Sci-fi | Helmets as states; screens as light; battles seen from the bridge |
| Horror | A light source in every dark shot; scares in the sound |
| Mystery and detective | Clues as props that never change; flashbacks as states |
| Christmas romance | Snow and string lights written as light; the festival crowd in the background |
| Sitcom and sketch comedy | Punchline and reaction in separate shots; one living room every episode |
| Fruit drama | Each fruit with its own colour, shape and accessory; three named per shot |
| Cat and pet drama | Real animals with an inner voice, or cartoon animals who talk |
| Short film | The Film type and 16:9; longer takes with one movement each |
| Animated short film | The animation style picked first; one distinguishing trait per character |
| Book trailer | A one-page trailer script, or the cover with an end card |
| Movie trailer | The three-act trailer in short shots; the title spoken, not drawn |
| K-drama | East Asia as the target region, Korean dialogue, the slow beats given their own shots |
| Heiress and hidden identity | A ladder of who-knows-what; two names and two looks of one woman; the reveal told in reaction shots |
| High school drama | The country comes from the school scene style and uniform; romance stays innocent |
| Joseon palace drama | Joseon Korea scenes, hanbok worn by rank, poison and punishment told off screen |
| Ottoman palace drama | Tiled halls and court dress, Turkish dialogue, power exercised by order, not by fight |
| Regency romance | English Manor, high-waisted gowns, and a ballroom staged without a hundred dancers |
| Pirate or Viking adventure | Decks, harbours and longhouses instead of open sea; storms set on the segment header |
| Mermaid fantasy | One character with a tail look and a shore look, kept fully dressed |
| 1980s small-town mystery | Period in the props and clothes, danger told without harm on screen |
| Folk horror | An invented ritual, a place with rules, the fear without the gore |
| Cozy farm life | Small stakes, a field with two seasons, a gentle look |
| BL romance | Two leads made different at a glance; intimacy kept to looks and closeness |
| History | An invented witness inside a real event; famous figures kept off camera |
| Bible stories | A narrator carries it; one story beat per episode; miracles as short shot sequences |
| Bedtime stories for kids | A calm look, a gentle narrator and a quiet ending instead of a cliffhanger |
| Your manga or webtoon as anime | Chapters written as scripts; your own designs uploaded as the reference sheets |
| Faceless YouTube channel | A recurring cast in a continuing series, with Shorts cut from each episode |
| Zombie apocalypse | Infection as a state of the character; the horde as background; the bite always off-screen |
| Dragon fantasy | The dragon as a character with its own sheet and a state for each age |
| Princess or royal drama | Court dress scaled by rank; the crown as a prop; the ball as background guests |
| Police or crime drama | Forensic Cold White rooms; every clue a prop; the crime told through what is left behind |
| Ghost or haunted house | The ghost as a character with a ghostly state; sound before sight; a light in every dark shot |
| Medical drama | Hospital Ward and uniforms by role; emergencies on faces and monitors, not bodies |
| Family inheritance drama | One dinner table as the stage; the will as a prop; the outsider's before-and-after states |
| Robot or android drama | Machines as characters with a build description, a preset voice and a revealed state |
| K-pop idol drama | One mark per member; stage outfits as states; the practice room as a location |
| Alien or first contact | Aliens as characters described plainly; disguises as states |
| Space opera | Starship Uniform ranks; the bridge as a recurring location; battles in short cuts |
| Thai lakorn-style drama | Thai Royal Palace or Southeast Asian Town; dialogue in English or another available language |
| Military drama | Modern Military with invented insignia; combat before and after |
| Demon king dark fantasy | Dark Asura or Gothic Dark; the true form as a state; menace without gore |
| Prison drama | The cell, yard and visiting room as locations; the jumpsuit as a state |
| Sports drama | Sportswear kit, team colours and numbers; the game in one-action shots |
| Courtroom drama | Every exhibit a prop; the jury as background; long speeches split across shots |
| Rebirth revenge | Two lives as two states of one face; the death never shown |
| Time travel romance | One era as the project's style, the other through locations and states |
| Survival drama | Worn states as time passes; storms and fog set on the segment |
| Bollywood-style family drama | Indian Bazaar Streets or Indian Royal Palace; wedding outfits as states |
| Single dad and nanny romance | The child as a character with a sheet; small domestic scenes |
| Cinderella story | Before and after as states; an original gown, not the famous one |
| Noir detective | Black & White Film, rain on the segment, the detective's voice-over |
| Witch or magic academy | Magic Academy and Academy Robes; spells as light and effect |
| Filipino teleserye-style drama | Two families, Target Region Southeast Asia, dialogue in English or another available language |
| Superhero drama | The suit as a state; powers as light and effect; an original design |
| Elf or fae fantasy | Elven Forest; pointed ears on the sheet; glamour as a state |
| Bodyguard romance | One mark for the guard; the threat as a prop; danger shown before it lands |
| Turkish dizi-style drama | Turkish dialogue; the long look given its own shot |
| Western | Wild West and Western Cowboy; horses as characters; the showdown cut before the shot |
| Spy thriller | The cover as a state; dead drops as props |
| Chef or cooking drama | The signature dish as a prop; steam and light written in |
| World War II drama | WWII Front with invented insignia; invented people; war before and after |
| Desert romance | Gulf Metropolis or Silk Road City; Target Region Middle East; Arabic dialogue |
| Action movie | One move per shot; chases built from segments; stunts before and after |
| Monster movie | The creature as a character; scale against familiar things; a late reveal |
| Post-apocalyptic | Wasteland or Frozen Wilderness; one mark per faction; worn states over time |
| Psychological thriller | One room seen again and again; one detail changed; a narrator who may be wrong |
| Second chance romance | Past and present as two states of one face; the place they met in both times |
| Kung fu and martial arts | Retro Hong Kong or Ink Jianghu; one technique per shot; training as states |
| Medieval knights | Plate Armor with one heraldic colour each; jousts one action per shot |
| War god returns | A humble and a revealed state; the proof as a prop; the reaction shot as payoff |
| Dinosaur adventure | Each named dinosaur a character; realistic or cartoon by audience |
| Greek mythology | Greco-Roman worlds; gods with a mortal and a divine state |
| Cyberpunk | Cyber Metropolis and Neon Nights; implants on the sheet; screens kept wordless |
| Samurai or ninja | Japanese Machiya and Samurai Armor; duels in stillness, draw and aftermath |
| Enemies to lovers | Rivals staged as mirrors; banter timed in shot seconds |
| Divorce revenge | Before and after as states; the ring and the papers as props |
| Heist | The plan told over empty rooms, then the job in the same rooms |
| Office romance | The elevator, the meeting room and the roof; colleagues in the background |
| Hockey romance | Numbers and colours on every sheet; the romance off the ice |
| Contract marriage | The contract as a prop; wedding, public and at-home states |
| Love story | Two leads drawn once; the places that matter; one romance shape |
| Dark romance | A dangerous lead; gothic and noir rooms; light written per shot |
| Secret baby | The shared feature in both sheets; the years as states |
| School bully revenge | Four school settings; the crowd and the aftermath; a comeback that exposes |
| Gangster or yakuza | Rank and tattoos on the sheet; codes as rituals; violence off the frame |
| Boxing and fight | One punch per shot; the corner as the drama; training as states |
| Racing | Invented cars as props; the race in short tracking shots |
| Nollywood-style | Target Region Africa; African Prints; accents in the voice description |

## Anime, cartoons and formats

Not every series is a genre. These guides cover drawn looks, children’s channels and ways of using a series, from a festival short to a language lesson.

| Format | The technique that matters most |
| Anime series for YouTube | One anime look for every episode; the cast drawn once |
| Anime fight scenes | One move per shot; power-ups as states |
| Shojo romance anime | Watercolor Anime, a Japanese school year, small moments |
| Anime movie | Film type in a drawn style, sequence by sequence |
| Kids cartoon YouTube channel | Family 3D or 3D Toy Cartoon; a small cast children recognise |
| 3D cartoon movie | The 3D visual styles, Film type, an original world |
| Claymation short | The Claymation style, small sets, gentle comedy |
| Fairy tale videos | Public-domain tales with your own designs |
| Educational cartoon | One question per episode, one sentence to remember |
| Reddit-style story drama | Your own story, rewritten as scenes; real people kept out |
| Talking baby videos | The baby as a character with a grown-up voice |
| Talking animal videos | Real animals on four legs, cartoon animals with hands |
| Festival short film | Film type, Enhance, a badge-free download, SRT |
| TV pilot or pitch trailer | A pilot episode and a one-minute trailer with the same cast |
| Language-learning drama | One situation per episode, in any of the 15 dialogue languages |
| Isekai anime | The other world as the project's look; the old one through a few locations |
| Shonen anime | Hero and rival as mirrors; training arcs as states |
| Slice of life anime | A few rooms drawn once; seasons set on each segment |
| Tabletop campaign series | Session notes rewritten as scenes; one episode per session |
| Corporate training drama | The wrong way and the right way in the same room |
| Mockumentary | Talking heads to camera, then the observed truth |
| Podcast to drama | The host as narrator; the story acted |
| Stage play to short film | The set opened out; long speeches split across shots |
| Brainrot videos | Invented hybrid creatures; a narrator; 15 seconds each |
| True crime style | Narrator, reconstruction, interviews and evidence, around an invented case |
| Documentary style | Narration over B-roll; interviews to camera; nature and history |
| Birthday story | The birthday person as the hero of a one-minute adventure |

## What AI video does well, in any genre

- Faces that hold. Each character on SceneMixer has a reference image used in every shot they are named in, so the same cast carries a whole series. Same face in every shot.
- Rooms that hold. Each recurring location has its own image. Same room in every episode.
- Dialogue in the picture. Lines are performed with the shot, mouths moving to the words, in any of the dialogue languages. Dialogue and sound in one generation.
- Close-ups and two-shots. Confrontations, confessions, reactions: the core of nearly every drama genre.
- Light and weather. Night, rain, snow, candlelight and neon, written as causes in the shot.

## What every genre has to work around

| Hard thing | Where it comes up | The workaround |
| Big crowds | Galas, balls, battles, festivals | Name only the people the scene is about; the rest go in the background line. Background people |
| Long continuous fights | Action, xianxia, werewolf | Single exchanges with reactions between them. Fights and crowds |
| Readable text | Contracts, letters, screens, clues | Show the object, say the content. Text in the picture |
| Gore and extreme violence | Horror, mafia, revenge, vampire | The moment before and the aftermath. Blocked by moderation |
| Brands, celebrities, franchises | Billionaire, fan stories, parody | Describe by look; invent names. IP suspected |
| A body or outfit that changes | Werewolf, vampire, period, sci-fi | A state of the character. States |

## How to start, whatever the genre

- Get a script. Paste your own on Upload Script, or write one or two sentences in AI Create with the genre, the lead and what they want, the opponent, the ending and the tone. Writing the one-line brief.
- Set step 1 for the genre: Type (Short Drama, TV Series or Film), aspect ratio, visual style, scene style, costume style, region and dialogue language. The visual style is fixed once you continue. Visual style presets.
- Build the cast in step 2 and add the states your genre needs. The first 15 setup images on an account are free.
- Storyboard episode 1 (free on your first project) and rewrite the genre’s signature scene with the technique from its guide.
- Generate, fix, composite, post. SceneMixer's default tier is 10 credits a second. How to make an AI short drama.

Visual Style: Shot on Film

Visual Style: Anime

Visual Style: Watercolor Anime

Visual Style: Family 3D Animation

Visual Style: Claymation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

| If your story is… | Type | Aspect ratio |
| A feed series with a hook every episode | Short Drama | 9:16 Portrait |
| A slower series with room for scenes | TV Series | 16:9 or 9:16 |
| One piece, a short film or trailer | Film | 16:9 Landscape |

## FAQ

### Which genres work best with AI video?

Genres built on faces and dialogue in a few recurring rooms: romance, revenge, family drama, mystery and comedy. Large crowds, long fights, readable text and gore are the hard parts in any genre, and each has a workaround.

### Does the genre change the price?

No. Every genre on SceneMixer is priced the same way: the script read, setup images, the storyboard and video by the second at the tier you choose.

### Can I mix genres, like a werewolf mystery?

Yes. Use the settings of the main genre in step 1 and borrow the techniques of the other: states for the wolf, prop images for the clues.

### What is the most popular AI short drama genre?

Romance in its many forms (billionaire, werewolf, vampire, second chance) and revenge dominate vertical drama catalogues, with comedy and animal stories strong on social feeds.

### Can AI Create write a script in any of these genres?

Yes. Name the genre in the brief along with the lead, the opponent, the ending and the tone. Each genre guide has an example brief.

### Can I make an animated version of any genre?

Yes. Pick a 2D, 3D or stop-motion visual style in step 1; the rest of the workflow is the same.

### Can the dialogue be in another language?

Yes. Choose the dialogue language in step 1; the cast speaks it natively rather than being dubbed.

### Which genres get blocked most?

Those that lean on gore, explicit romance, real people or famous franchises. Each genre guide shows how to keep its signature scenes clear of the content check.

Updated November 3, 2026


---

# AI Stories: Ten Finished Examples and the Shape They Share

URL: https://scenemixer.com/guides/ai-stories/
Ten finished AI stories from the SceneMixer samples: how each one opens, turns and ends, the shape they share, and how to write your own in the same mold.

# AI Stories: Ten Finished Examples, and What They Have in Common

Quick answer: Ten of the sample stories on this site are in English, and each was made from a script into a finished episode. Laid side by side they are small and they share one shape. Each has four to nine characters and one to three places. None spends its opening on set-up; each starts with something already happening, turns on a reveal, and ends on a reversal or an open question rather than an answer. Nine of the ten also hang the plot on one physical object. This guide lists all ten with how they open, turn and end, shows what the shape means for a story of your own, and points to where to build one.

## The ten, in one table

The descriptions below are paraphrased from each story’s own sample page, where you can also watch the episode. Every one is a first episode, and several are written as the start of something longer.

| Story | What it is about | Where it opens | Where it turns | Where it ends |
| The Veil at Dawn | A woman wakes with memories of a past life and walks out of a wedding she knows will cage her | Mid-confrontation in the wedding suite | The walk-out, then a chance meeting in a garden | On a hint of a larger mystery around another character |
| The Violin on Mercer St | A violinist returns to the neighbourhood she fled seven years earlier | A direct confrontation, no warm-up | Each episode flips one assumption about guilt and loyalty | On a single piece of evidence that rewrites what came before |
| The Eagle’s Vow | A lone soldier carries a legion’s eagle standard along a dangerous road to a frontier camp | On the forest road, already under pressure | Suspicion inside the camp | On the identity of the unknown man who steps into the tent |
| Rust City Knockout | A night chase through a rain-soaked harbour district ends with a crate and a chain | Already running | The crate opens on a surprise and the chain bonds itself to the heroine’s arm | Mid-standoff, as armour spreads up her forearm and masked strangers arrive |
| The Glass Floor | A firing on a glass-walled executive floor turns into a public accusation | At the moment of the firing | Two reveals: a signature page and a buyer’s name | On a buyer line that carries the heroine’s own name |
| Neon Debt | A courier runs a data case through a night market with a pursuit team behind her | Mid-chase | The delivery is an erasure order, and a sister she thought lost turns out to be on file, alive | On a trap: the man helping her is a signed asset of the corporation |
| Crimson Vow | A wet-alley rescue becomes a family-secret mystery | On a rescue in progress | A locket bears the heroine’s name and tonight’s date | On a third voice that already knows her surname |
| Rocket Pup | A family adventure that starts with a dog riding a rocket | On the hook: why is a dog riding a rocket? | The mayor’s vanity, a letter mystery and a cat invasion | A double cliffhanger: an empty cockpit with a ringing phone, and a tower key with a cat-face side |
| The Last Envelope | A wartime courier in a bombed, blacked-out city | Pursuit | Infiltration, then an encounter | On a reversal: the messenger is the target |
| Starfall Command | A defector steals a command core to turn a station’s defences on the fleet besieging it | Mid-action in a burning docking bay | Backstory arrives through confrontation, never through exposition | On a twist: the enemy admiral is the man she thought was lost |
The samples page has all of them with the finished episodes: SceneMixer samples. There are twenty-three in total, thirteen of them in other languages.

## What the ten have in common

They begin in the middle. Read the “opens” column again: not one starts by introducing a world. The first thing on screen is a confrontation, a chase, a rescue or an action already under way, and the background arrives later, through the people, not before them. A video model draws what is visible, so a story that opens on an event gives it something to draw from the first second.

They are small. The casts run from four to nine characters and the stories from one to three places. That is not a limit on what SceneMixer can hold; it is the size at which every character’s face and every room stay the same in every shot and one person can finish the episode. Why the cast stays consistent explains what that size buys.

Nine of the ten hang on an object. A contract that is torn upstairs and sealed at the curb, an envelope, a locket, a data case, a legion’s standard, a violin with a carved thorn, a command core, a tower key, a crate and a chain. Only The Veil at Dawn turns on something that cannot be held. An object is the easiest thing for a camera to follow, it can change state as the story turns, and a reveal can be a close-up of it. Same prop in every hand covers how an object keeps its look.

None of them ends on an answer. The endings are an identity, a name, a reversal, a trap, a piece of evidence or a hint of something bigger, and two of them stack two cliffhangers. Every episode ends one beat before the answer, which is what makes the next one worth opening. Hooks and cliffhangers lists the kinds of ending to rotate.

## Ten engines, not ten genres

A genre label says what the story looks like; the engine says what keeps the viewer watching. The same few engines run under very different covers, so they are a better thing to choose first.

| Engine | The question the viewer carries | Seen in |
| Escape | Will she get out, and what is she running toward? | The Veil at Dawn, Neon Debt, Rust City Knockout |
| Return | Why did she leave, and what happens now she is back? | The Violin on Mercer St |
| Secret | What is in the contract, or the locket, and who else knows? | The Glass Floor, Crimson Vow |
| Mission | Will the message, the core or the standard arrive, and who is waiting? | The Last Envelope, Starfall Command, The Eagle’s Vow |
| Mischief | How did this get out of hand, and who is behind it? | Rocket Pup |
Choosing an engine first also tells you what to cast: an escape needs a pursuer, a secret needs a keeper and a seeker, a mission needs a carrier and someone in the way. For the settings that suit each genre on top of the engine, see AI short drama genres.

## Write yours in the same mold

Fill these six lines before you write a word of the story. They are the ten tables above reduced to a form.

- Who wants what, by when? One person and one want that has a clock.
- Who or what is in the way? A person is best.
- The object. One thing that can be held, shown and changed as the story turns.
- The first image. Something already happening, not a place being introduced.
- The turn. One reveal that changes what the object means.
- The last beat. A question, stopped one beat before its answer.
A worked line for a story that is not on the list: a piano tuner is let into a house she has never visited and finds a framed photograph of her own family on the mantel. The want is to finish and leave before the owner comes back. The thing in the way is the owner, who is already on the stairs. The object is the photograph. It opens with the tuner’s hand already on the lid, it turns when the back of the photograph names a date, and it ends as the owner says her name from the doorway. Three characters, one house. It fits the same mold as the ten, and none of it needs a sentence of set-up.

## Turn a story like this into a video

Three routes lead to the same SceneMixer project. Paste the story as a script. Give AI Create one or two sentences, the fixed line from the form above, and read what it drafts. Or write the first episode yourself and let the site continue it later. Four ways to start a project compares them; Writing the one-line brief shows how to ask for the six lines directly. The first 15 setup images on an account are free, and the first drama’s episode 1 storyboard is free, so a story of the size above is a cheap place to begin.

Once the story is written, the checks that decide whether it films well are in AI story: writing one a camera can use, and who decides what while it is being written is in AI story writer. The machine-written draft is covered in AI story generator.

## FAQ

### What is an AI story?

A story written for, or with the help of, an AI model. On SceneMixer it is a story meant to be filmed: it is read into a cast, places and episodes, then turned into video.

### Are these the only AI stories on the site?

No. SceneMixer has twenty-three finished samples, ten in English and the rest in thirteen other languages. The ten here are the English ones.

### How long is each of these stories?

Each is one episode between about 1 minute 46 seconds and 3 minutes 30 seconds, from four to nine characters in one to three places.

### Do all the stories have to end on a cliffhanger?

A series that is watched one episode at a time usually does, and all ten of these do. A film can end quietly. The type you choose in step 1 sets how the script is cut.

### Why do so many of them rely on an object?

A camera can follow an object and show it changing. A feeling cannot be drawn, so these stories give the feeling a thing to be about.

### Can I use one of these as my own story?

No. They are samples to watch and learn from. Write your own, or give SceneMixer’s AI Create a line and edit what comes back.

### How many characters should my story have?

The ten use four to nine. Fewer than that is easier to keep consistent. Every named character needs a picture and a voice that stay the same.

### Can an AI story be in another language?

Yes. SceneMixer sets the dialogue language in step 1, and the sample list includes stories in thirteen other languages.

Updated November 2, 2026


---

# AI Story: What Makes a Story Easy to Turn Into Video

URL: https://scenemixer.com/guides/ai-story/
How to write an AI story that becomes a watchable video: a lead who wants something, an opponent, what a camera can see, few places and a turn each episode. With numbers.

# AI Story: Writing One a Camera Can Use

Quick answer: An AI story is only as good as what a camera can do with it. A video model draws what is visible and audible, so a story that works on screen has a lead who wants something, someone or something in the way, events you can see rather than thoughts you cannot, a few places, and a turn before each episode ends. This guide turns that into six tests and a stopwatch: about three spoken words a second, two to eight seconds a shot, two to four scenes an episode, and an open question at both ends.

Try it: read your story’s dialogue aloud with a stopwatch. If one scene runs past the seconds in the table below, it is two scenes.

## Six tests a story has to pass

| Test | A story that passes | A story that fails | The fix |
| The lead wants something | Mara must get a letter back before midnight. | Mara is sad about her life. | Finish the sentence “she wants… before…” |
| Someone is in the way | The harbour master wants the same letter burned. | A story about a lighthouse. | Give the want an opponent; a story with no opponent drifts |
| It can be seen or heard | He tears the letter in half and drops it. | He remembers his father. | Turn the thought into an action, a spoken line or a voice-over |
| Two to four places | A stairwell and a balcony. | Nine locations in three minutes. | Merge places; each location is a picture that has to stay the same |
| A turn about every 30 seconds | The letter is addressed to someone else. | Twenty seconds of walking, then a speech. | Add new information, a shifted loyalty or rising danger |
| It ends open | Her phone lights up with her sister’s name. | They make up and smile. | Stop one beat before the answer |

The last two tests are the ones short formats live on. Writing hooks and cliffhangers covers the kinds of ending to rotate, and the one-line brief shows how to ask for them up front.

## The stopwatch: seconds, words and shots

Video is bought by the second, so a story has to be counted in seconds before it is counted in pages. The numbers to plan with:

| What | Number | Why it matters |
| Spoken English | About three words a second | A twelve-word line needs about four seconds of shot, plus any movement |
| A look, a turn, a small gesture | 2–3 seconds | Shorter than that reads as a flicker |
| One line and its reaction | 3–5 seconds | The commonest shot in a drama |
| Crossing a room, an action beat | 6–10 seconds | Give movement the room it needs |
| One line of dialogue | Under about 20 words | Longer lines are cut across several shots |
| A 1–2 minute episode | 180–360 spoken words in 2–4 scenes | Where a vertical drama usually lands |
| Episode ceiling | Short Drama 240 s, TV Series 300 s, Film none | Chosen in Step 1 before the read |
So a 60-second SceneMixer episode is four segments of 15 seconds, and with three to five shots a segment that is twelve to twenty shots. If your scene needs forty, it is a different, longer episode. Shot seconds and timing · Vertical short drama script format · Short Drama, TV Series or Film.

## The same scene, rewritten to be filmed

Before (reads well, films badly):
Mara had dreaded this for years. She climbed the stairs feeling that nothing would be the same, thinking about the day her father left, and knocked on the door, hoping he would not be home.
After (everything is something a camera can show or a voice can say):
INT. STAIRWELL, NIGHT. Mara, in a wet jacket, checks a door number against a letter and knocks twice.
Teo (through the door): Who is it?
Mara (steady): Your father’s courier. I have something you were not supposed to get.
The chain slides. A crack of light falls across her shoes.
What changed: the dread became an action (checking the number) and a line (“you were not supposed to get”); the memory of her father became a connection the viewer learns from the dialogue; the ending is a door opening, not a feeling. Lines written as Name (delivery): words are carried into the storyboard word for word, while speech folded into narration is only summarised. Write dialogue so it is picked up.

## A three-episode shape to copy

| Episode | Opens on | Turns | Ends on |
| 1 | A ticket scanner beeps: name already used | The passenger knows a nickname only the deckhand’s late brother used | A reveal: the ticket was bought yesterday |
| 2 | The deckhand follows the passenger onto the car deck | The passenger is not alone; the ferry captain is in on it | An arrival: a phone rings in the brother’s old locker |
| 3 | The locker, open | What the brother hid is not money but a recording | A turn: the voice on the recording is the captain’s |

Notice that each episode opens in the middle of the conflict and ends on a different kind of question, and that the whole story needs one ferry and two or three people. That is a small cast on purpose: every character needs a picture and a voice that stay the same, and every extra place needs its own. The same face in every shot.

## What to leave out of an AI story

- Readable text as a plot device. A letter, a sign, a text message, a price: video models do not draw legible words reliably. Let a character say it. Text on screen: fixes.
- Real people and famous characters. The IP check refuses a picture or a clip that resembles a real person, a brand mark or a known character, so invent your own cast.
- Crowds and brawls. AI video does best with a few named people and one clear action a shot; wide shots of many people tend to blur, so cut between close shots. Writing fights, crowds and group scenes.
- Time jumps carried by a caption. A caption line is kept in the storyboard but SceneMixer does not draw it in the clip at the moment, so say the jump in a line (“Three years later, and you still have the key”).
- Thoughts. Use a voice-over line or a look held a beat longer.

## Letting AI write it, then checking it against the tests

You can write the story yourself, paste it, or give SceneMixer’s AI Create one or two sentences and read what comes back. The six tests work the same on a machine draft: read it for the want, the opponent, the places and the ending before you press the button that reads it. If a test fails, the quickest fix is a sharper brief or an edit to the script, both of which cost far less than a video. AI story generator shows what the draft turns into, and AI story writer shows who decides what.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: check that the two or three people you wrote are the whole cast, and that each place is one you meant. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

## FAQ

### What is an AI story?

A story written with the help of an AI model, here meant to be filmed: it becomes a script, a cast and a storyboard, then a video.

### What kind of story works best as a video?

One with a lead who wants something, an opponent, events a camera can show, two to four places and a turn before each episode ends.

### How long should a scene be?

A vertical drama episode of one to two minutes usually holds two to four scenes and 180–360 spoken words, at about three words a second.

### Can I use my own story?

Yes. Paste it into SceneMixer or upload a .txt, .docx or .md file. Dialogue written as Name: line is kept word for word; speech inside narration is summarised.

### Why does my story need so few characters?

Every SceneMixer character gets a picture and a voice that are used in every shot, and each extra place needs its own picture. Two or three characters carry a short story.

### Can a story rely on a letter or a sign?

Not reliably. Video models do not draw legible text, so put the words in a spoken line.

### Does the story have to end on a cliffhanger?

In a short drama series every episode except the last should end on an open question. A film can end quietly.

### Is it free to try a story?

SceneMixer prices writing by length, reading is 4 credits per 1,000 characters, and the first drama gets 15 free setup images and a free episode 1 storyboard.

Updated November 2, 2026


---

# AI Story Continuer: What It Reads and How to Steer It

URL: https://scenemixer.com/guides/ai-story-continuer/
How the AI continuation works in SceneMixer: the five things it reads, the three controls (direction, episodes, length), what it costs and how to check the draft.

# AI Story Continuer: What It Reads, What It Writes, How to Steer It

Quick answer: The AI story continuer is the Let AI continue tab on the Add an episode card at the bottom of step 3. It does not re-read your whole story. It is handed a short brief: a recap of every earlier episode, the closing hook of the last one, your cast and places by name, the last 600 characters of your script and, if you type one, a direction of up to 500 characters. From that it writes one to three new episodes in the language and format of your script, then stops and shows you the draft. Nothing is added to the project, and no reading is charged, until you press Add and read. This guide says exactly what it reads, what it cannot see, and how to steer it.

## Where the continuer lives

Open a SceneMixer project whose script has been read, go to step 3 and scroll below the episode cards. The Add an episode card has two tabs: Paste the script and Let AI continue. The second is the continuer. It is only offered where there is something to continue: a project of your own whose reading has finished. Ad projects, remakes and the sample projects do not have it. Making the next episode covers the whole flow, including what to do when the next episode is already in your script. This page is about what the continuer itself reads and writes.

## What it reads: five things

| It reads | Exactly what | How much |
| 1. A recap of the story so far | The title, then every earlier episode as its number, title and short synopsis | Up to 1,500 characters for the last episode, 200 for each of the others, 6,000 for the whole recap. If the recap runs over, the earliest episodes shrink to titles first |
| 2. The last episode’s closing hook | The cliffhanger the reading recorded at the end of your last episode, which is where the new episode must pick up | One line |
| 3. Your cast and places | Characters first, then places, then props, each by name with the start of its description | Up to 40 entries, 60 characters of description each |
| 4. The way your script ends | The last stretch of your own script, cut at a sentence boundary where it can be | The last 600 characters |
| 5. Your direction | What you type in A direction (optional) | The first 500 characters |
The names are the part that matters most. SceneMixer’s continuer is told to use the cast and places exactly as listed, not to rename anyone and not to invent a second character with the same job. It may bring in new people. The leads and the people already in the project keep their names and their roles.

## What it cannot see

The brief is short on purpose, so a few things are outside it, and it helps to know which.

- The full text of your earlier episodes. SceneMixer shows it the recap and the last 600 characters, not every line. A detail that lives only on page 12 may not make it into the new episode, so put anything that must carry over into your direction.
- The storyboards and the clips. What was shot, and how it came out, is not part of the brief. The continuer writes script, not shots.
- Your pictures and voices. It sees the written descriptions of the cast, not their sheets or voices.
- Your style choices. Visual style, scene style and costume style are fixed for the project and are not something the writer changes.
Everything it writes is then read again as a normal script, so the cast, places and props of the new pages are checked against the existing project. A person who appears for the first time is added; a person who is already there is matched.

## The three controls

Direction. A sentence or two of what should happen next, in plain events. The field shows an example along the lines of the old man takes her to the pier; the black car follows. It is optional: left empty, the continuer is simply told to carry on from the hook and choose the direction itself.

Episodes to write. SceneMixer offers one, two or three. Each option shows what it will target (“About N characters”) and its price before you press anything, so you can compare writing one episode against three.

Length. There is no length box. Each new episode aims at about the average length of the episodes already in the project, held between a floor and a ceiling so that one very long or very short opening episode does not set the pace for the rest. To get longer episodes, write or paste one by hand. To get shorter ones, write fewer episodes at a time and trim the draft.

The draft is written in the language and format of your script: the same scene headings, the same way of marking who speaks, a numbered episode heading to begin each new episode. If your interface is in a language that uses the same writing system as your script, it is written in that language; otherwise it follows the script.

## Steering that works

The continuer starts at the closing hook, so the most useful thing you can give SceneMixer is the next event, not a mood. Compare:

| Direction | What the continuer does with it |
| “Make it more exciting” | Nothing to build on. It will choose its own event, and may choose a quieter one |
| “She finds out the truth” | A reveal with no object or person attached. Better: say which fact, and who tells her |
| “Mara takes the letter to the harbour master; the black car follows her onto the pier” | Names a person from the cast list, a place, a prop and a pressure. This is the shape that survives into a storyboard |
| “Two episodes: the first ends with the car stopping; the second ends with the door opening” | A direction can set the end of each episode, which gives the draft a hook to write toward |
Three habits help. Use the names exactly as they appear on the character cards. Ask for one episode first, read how it sounds, then ask for more once the voice is right. And treat the hook as a promise: a direction that contradicts how the last episode ended makes the continuer choose between your note and the recap.

## Read the draft, then add it

While it writes, the card says Writing the next part… and counts the characters so far. When it is done, a banner says the draft is in and to read it, change anything and then add it. The draft opens in the paste tab, so you edit it like any pasted script. Discard draft throws it away. Try again runs it once more with the same or a new direction.

The money is in two steps and both are visible. SceneMixer prices the writing by what it writes, at your text model’s rate: on the default text model that is 2 credits per thousand characters, and English counts 3,333 characters as 1,000, the same rule the reading uses. A draft that fails is refunded under the usual rules. Only when you press Add and read is the new text read, priced like any reading of that length, 4 credits per thousand characters on the default. Earlier episodes, their storyboards and clips are not touched.

After the reading, a window lists any character, place or prop that appears for the first time, and any new state of an existing one such as an outfit change or a ruined room, with the price of drawing them. The first 15 setup images on an account are free, and they count if you still have them.

## Limits worth knowing

| Limit | What it means |
| Three episodes per draft | To get more, run it again after you add the first draft |
| Fifty episodes per project | The same ceiling as the first reading. At fifty the card stops offering more; start part two as its own project and bring the cast with Pick existing |
| One draft at a time | While a draft is being written the card asks you to wait for it, or stop it, before starting another |
| Room in the project | If the script is already close to the longest the reading accepts, there may be no room for another episode. The card refuses instead of writing something that could not be read |
| No re-reading as a whole | Once something has been added, the project cannot be read again from the top. To cut the same script differently, start a new project with it |
Part two as a separate project is explained in Reuse a character, scene or voice across projects, and the habits that keep a long run steady are in Making a multi-episode series.

## FAQ

### What is an AI story continuer?

A tool that writes the next part of a story from where it stopped. In SceneMixer it is the Let AI continue tab on the Add an episode card in step 3. It writes script text that you review before it becomes episodes.

### Does it read my whole story?

No. SceneMixer gives it a recap of the earlier episodes, the closing hook of the last one, the cast and places by name, the last 600 characters of your script and your direction.

### How many episodes can it write at once?

SceneMixer writes one, two or three. Run it again for more.

### Can I control where the story goes?

Yes, with SceneMixer’s direction field, up to 500 characters. Name the people from your cast list and the event you want. Leaving it empty lets the continuer choose.

### Will it keep my characters’ names?

It is told to use the names in your cast list exactly and not to create a second character with the same role. It can add new people.

### Can I edit what it writes?

Yes. The SceneMixer draft lands in the paste tab as ordinary text. Change anything, or discard it, before pressing Add and read.

### Does it cost credits?

SceneMixer prices the writing by length at your text model’s rate and the price is on each option. The reading of the added text is charged separately, only when you add it.

### Can I continue an ad, a remake or a sample?

No. The continuer works only on your own SceneMixer projects that were made from a script and have finished reading.

### What if it contradicts an earlier episode?

Edit the draft before adding it. Anything that must carry over, such as a fact that appears only on one page of an early episode, is safest stated in your direction.

Updated November 2, 2026


---

# AI Story Generator With Pictures: What You Get and What Each Is For

URL: https://scenemixer.com/guides/ai-story-generator-with-pictures/
A story in SceneMixer comes with sheets, cover art and moving shots, not storybook pages. What each is for, how many pictures a real story needs and which ones are free.

# AI Story Generator With Pictures: Which Pictures You Get

Quick answer: A story in SceneMixer comes with three kinds of picture, and none is a page-by-page storybook illustration. Sheets are plain reference pictures of each character, place and prop, made so every shot can copy them. Cover art is the one picture made to be looked at: three poster options with your title painted in. And the shots themselves are the moving pictures, drawn from the story text and the sheets. The ten English samples need between 5 and 11 sheets before props, so the first 15 setup images on an account cover a story that size. This guide says what each kind is for and where you can change it.

## Three kinds of picture

| Picture | What it shows | What it is for | How many |
| Character sheet | One person, full length, on a plain white background | The reference every shot with that person copies | One per character who is seen |
| Outfit or state picture | The same sheet, changed in one thing: an outfit, an injury, a drawn blade | A reference for the scenes where that change applies | One per state you need |
| Place sheet | The empty location with its layout readable, no people | The reference for every shot set there | One per place |
| Prop sheet | The object alone on white | The reference when the object is in a shot | One per prop that matters |
| Cover art | A poster with your title painted into it; the site adds the tagline and episode numbers | The one picture made to be looked at: for the project, for sharing | Three options per set |
| Shots | The moving pictures themselves | The video | Drawn from the story text and the sheets above |
What none of them is: a picture for every page of a story. If you want a storybook with one illustration per page, this is not that. The pictures that tell the story are the shots, and the sheets exist to make sure the same faces and rooms appear in all of them. AI story video generator covers the way a story becomes a video.

## Why the sheets look plain

A character sheet shows the person full length on a plain white background; a place sheet shows the room empty; a prop sheet shows the object alone. That is not a limit of the picture maker, it is the job. Every shot is made by showing the model the sheets of everything in it, and the model copies what it sees, mood and background included. A sheet with a smoky background brings the smoke into a sunny scene. So the sheets stay neutral and the mood is written into each shot. Why character, scene and prop sheets look plain has the whole argument.

## How many pictures a story needs

Count the cast who are seen and the places, then add the props and the outfit changes that matter. Here are the ten English samples, with the cast and places each page lists.

| Story | Characters | Places | Together |
| The Veil at Dawn | 6 | 2 | 8 |
| The Violin on Mercer St | 5 | 1 | 6 |
| The Eagle’s Vow | 9 | 2 | 11 |
| Rust City Knockout | 6 | 1 | 7 |
| The Glass Floor | 5 | 2 | 7 |
| Neon Debt | 7 | 2 | 9 |
| Crimson Vow | 4 | 1 | 5 |
| Rocket Pup | 7 | 3 | 10 |
| The Last Envelope | 6 | 2 | 8 |
| Starfall Command | 7 | 2 | 9 |
The biggest of all 23 samples, in another language, has 11 characters and 4 places: 15 entries. These counts do not include props or outfit states, which depend on the story. Three kinds of entry do not need a picture: a character marked Off-screen (a narrator, a voice on a phone) is heard and never seen; a crowd is not cast, it is described in the shot; and a prop gets a sheet only where it matters to a shot. Leads, supporting cast and voice-only characters.

The first 15 setup images on an account are free, in any mix of characters, places and props. After that a picture on the default image model is 7 credits. The reading that finds your cast is priced by the script’s length, 4 credits per thousand characters on the default, and English counts 3,333 characters as 1,000.

## Changing a picture

| You want | Do this | Cost |
| A different face or outfit | Edit the Appearance or Clothing text on the card, then Regenerate | Priced like any image; the free allowance covers it while it lasts |
| One detail changed | Ask the card for an AI edit, for example an outfit, hair or expression. AI edit a sheet | Priced like an image, and a membership feature on the current plan settings |
| Your own art | Upload & Replace on the card (a paid-plan feature). A plain, well-lit picture on a simple background works best | Free; counts toward storage |
| The same person in another outfit | Add a state to the character instead of redrawing the sheet. Outfit changes and scene states | A state is one image |
| A different image model | Pick the image tier in settings; it applies to new sheets, not existing ones. Which image tier for sheets | Per image, at the tier’s price |
| Keep a copy | Download on the entry in the asset library. Regenerating replaces the image on the card, so download first if you may want it back | Free |
You do not have to make every sheet before moving on: SceneMixer’s step 2 lets you continue without generating all of them. Regenerate one sheet and describe it.

## The picture made to be looked at: cover art

Open the Cover art tab in SceneMixer’s step 2 once at least one lead has a character image. Type the title (twelve characters or fewer works best, because the title is painted by the picture model), choose the lettering, the genre mood and up to two cover characters, add an optional tagline, and press Generate 3 options. The three posters are built from your leads’ sheets, so the faces and costumes are the ones in the story. Pick one with Use this one; the site lays the tagline and the episode numbers on top, and makes the other shape, portrait or landscape, from it. The first set of three on an account’s first project is free. Make a promo cover poster for your AI drama.

## Where the story’s own pictures happen

The SceneMixer storyboard is the story written as shots. Each shot has a size, an angle, who is in frame, what happens, the line and the seconds. You can read all of it before a single frame is drawn, and editing the text costs nothing. The clip for a segment is then made from that text and the sheets of everything it names. So the way to get the picture you imagine is in two places: the sheets for who and where, and the shot text for what happens. How to write one shot and the episode editor tour show the screens.

For where the story itself comes from, see AI story generator and AI story.

## FAQ

### Does the generator draw an illustration for every page?

No. SceneMixer draws reference sheets for the cast, places and props, a cover poster on request, and the moving shots. It does not make a picture per page.

### How many pictures will my story need?

The cast who are seen plus the places, then the props and outfit states that matter. The ten English samples need between 5 and 11 before props.

### Are the pictures free?

The first 15 setup images on a SceneMixer account are free, in any mix. After that each is charged at the image model’s price, shown on the button.

### Why do the character pictures have a plain white background?

Every SceneMixer shot copies its sheets, background and mood included. A plain sheet can be used in any scene.

### Can I use my own drawings?

Yes, on a paid SceneMixer plan. Use Upload and Replace on the card. A plain, well-lit picture works best.

### Can I download the pictures?

Yes. Every entry in SceneMixer’s asset library has a Download button.

### Is there cover art for my story?

Yes. SceneMixer’s Cover art tab in step 2 makes three poster options with your title painted in. The first set on your first project is free.

### Can I change the picture style?

SceneMixer asks for the Visual Style in step 1 and fixes it once you continue. The image model for new sheets can be changed in settings.

Updated November 2, 2026


---

# AI Story Writer: What It Drafts and What You Decide

URL: https://scenemixer.com/guides/ai-story-writer/
How an AI story writer shares the work: you set genre, lead, conflict, ending and tone; it drafts the script; you rewrite lines, shots and endings by hand or on request.

# AI Story Writer: You Decide, It Drafts

Quick answer: An AI story writer drafts and you decide. In SceneMixer the decisions that make a story yours happen at three moments: in the brief, where you set the genre, the lead, the conflict, the ending and the tone; after the draft, where you write again, edit the script yourself or have the site continue it; and after the reading, where every character description and every shot is text you can rewrite by hand for nothing, or by instruction for 1 credit a revision. Below is who decides what, where each change is made and what it costs.

## Who decides what

| Decision | You | The writer | Where |
| Genre, lead, conflict, ending, tone | State them in the brief | Drafts to them | AI Create box |
| How much story | Pick Mini, Brief, Short, Medium or Long | Writes to that length; options too long to read afterwards are not offered | Length menu |
| How many episodes | Say it in the brief | Splits the script by story, not by clock | Brief, then the read |
| Language of the script | Switch the interface language first | Writes in it | Language button or Settings |
| Language the cast speaks | Choose the dialogue language | Does not follow the script language automatically | Step 1 |
| Names, beats, dialogue wording | Steer with the brief, rewrite after | Invents what it needs | Script, then storyboard |
| What a character looks like | Edit the description, regenerate the picture | Writes a first description | Step 2 |
| The wording of one shot | Edit by hand, or say what to change | Revises under your instruction | Storyboard card |
| What happens after the last episode | Give a line of direction, or none | Drafts the next episodes from where the story stands | Step 3, Add an episode |

The two language rows in SceneMixer trip people up: a Korean script does not by itself give Korean speech in the video, so pick the dialogue language in Step 1 as well. Writing the script in your language.

## The first draft: how to steer it

Put five things in one or two sentences: the genre, the lead and what they want, the conflict, the ending you want, and the tone. The genre sets the rhythm and the kind of hook each episode ends on; the ending, even one word, keeps the episodes pointed somewhere. Inspire Me fills the box with an example you can change.

Leave out camera directions, shot lists and character sheets, which come later; long backstory, which the writer will invent as needed; several stories at once; and real people or brands. Writing the one-line brief has before-and-after briefs.

If the draft is off, write again with a sharper brief. Each writing is charged on its own when it is delivered, a writing that fails is not charged, and every draft stays in Writing History, where any of them can be started later. The window itself does not edit the text: copy it, change it anywhere and paste it into Upload Script, or take it into the project and make small changes after the reading, where they are free. After the script is written.

## The second pass: rewrite by hand, or by saying what you want

| The writer’s job | Do it by | Why |
| Fix one word, a number of seconds, a name | Hand (free) | It is faster than describing it, and exact |
| Make a whole segment colder, or quieter | Instruction | The change touches many lines; one sentence covers it |
| Take a character out of three shots | Instruction | Tedious by hand; the revision rewrites the shots without them |
| Add a voice-over line in the right place and format | Instruction | e.g. “in shot 1, add a voice-over in Mara’s voice” |
| Add, drop or retime a shot | Either | Ask in the same box; the review shows the old and new shot count and total seconds |
| Put in a character chip from the project | Hand, with the @ menu | Instructions only use characters already in the project |

By instruction means SceneMixer’s AI edit on a segment’s card: write what to change in up to 500 characters, press Revise, and read the review, with removed text struck through and new text highlighted. Nothing is saved until you press Save. A revision that comes back costs 1 credit; one that fails, is refused or changes nothing costs nothing, and the result is checked before you see it: one that fails the checks (names that are not in the project, a different language, broken shot numbers) is dropped at no charge. There are eight starters to finish: change a line, an action, the mood, the time and weather, the camera, remove someone, add a voice-over, change the location. The eight starters, with examples · AI edit or edit by hand.

## The safety net: text history

Writers rewrite and regret. SceneMixer keeps History for the storyboard text: the original, which is never overwritten, plus the text from before each of your last five changes, labelled Before an AI edit, Before a manual edit or Before a restore. View shows where it differs; Restore puts it back, and the current text goes into the history in turn, so even a restore can be undone. It is free. What wipes it is regenerating that episode’s storyboard, which rebuilds the segments from scratch, or deleting the segment; copy anything you want first. Text history.

## Carrying the voice into later episodes

When the script ends and the story does not, Add an episode at the bottom of SceneMixer’s Step 3 offers Let AI continue. Give it a sentence of direction, or none, and choose how many episodes. It picks up after the last episode with the same cast, in the same language and format as your script. The draft is priced like AI Create, by the length it writes. You read it, change anything, and only then add it; nothing is read or charged for the reading until you do. Making the next episode.

## What the writer will not do

SceneMixer will not change how a character looks by editing the story: the face, hair, build and outfit come from the character’s picture, so change that. Voices are chosen on the character, music has its own switch, and the spoken language is a project setting, none of which an edit to the text moves. Real people’s likenesses and famous characters are caught by the IP check when pictures and clips are made. And an instruction cannot bring in a character or a place that is not already in the project; add it to the cast first. AI story generator shows what the finished draft turns into, and AI story shows what makes a draft filmable.

## FAQ

### What does an AI story writer do?

It drafts a script from a short brief. In SceneMixer you then read and rewrite it, and the script goes on to become a cast, a storyboard and video.

### What should my brief include?

The genre, the lead and what they want, the conflict, the ending you want and the tone, in one or two sentences.

### Can I edit what it wrote?

Not in SceneMixer’s writing window. Write again with a sharper brief, or edit the text elsewhere and paste it. After the reading, everything is editable: by hand for free, or by instruction for 1 credit a revision.

### What if I do not like the draft?

Write again. SceneMixer charges each draft on its own, failures are not charged, and every draft stays in Writing History.

### Can I change the ending afterwards?

Yes. Edit the last shots by hand or by instruction, and regenerate the segments to see it on video. Earlier text is kept in history.

### Which language does it write in?

SceneMixer writes in the language your interface is in when you press Start Writing. The cast’s spoken language is a separate choice in Step 1.

### Can it keep writing after the script ends?

Yes. Let AI continue drafts the next episodes with the same cast; you read and edit the draft before it is added.

### Can I undo a rewrite?

Yes. SceneMixer’s History keeps the original and the text before your last five changes, and restoring is free.

Updated November 2, 2026


---

# AI Text to Video Generator: The Controls That Shape the Result

URL: https://scenemixer.com/guides/ai-text-to-video-generator/
The same text can make different videos. Every control in the generator, which level it sits on (project, cast, episode, segment, shot) and what changing it does.

# AI Text to Video Generator: What You Control

Quick answer: The same text gives different videos depending on five levels of control, and knowing which level a setting lives on saves money. The project decides the type, the look, the shape and the spoken language, and is fixed once the script is read. The cast decides who people are. The episode decides how long a segment may be. The segment decides the tier and the price. The shot decides seconds, framing and who is in frame. Everything above the segment is text or settings and free to change; the segment is where each attempt is charged.

## Five levels of control

| Level | What you control | Set when | Can it change later? |
| Project (Step 1) | Type, Visual Style, Scene Style, Costume Style, shape, Target Region, dialogue language | Before the script is read | No. Start a new project with the same script |
| Cast (Step 2) | Names, descriptions, pictures, voices, outfit states, your own photo (members) | After the read | Yes. New shots follow; old clips keep the old look until regenerated |
| Episode | Segment length: Standard 15 seconds or Long take 30 seconds | When the storyboard is generated | Per episode |
| Segment | The tier (resolution, look, price per second) | When you press generate | Yes, per attempt; every take is kept |
| Shot | Seconds, who is in frame, framing, camera, mood, the line | Any time before you regenerate | Yes, in the text, for free |

A rule of thumb: change things from the top down only while it is free. The project row is decided by one read; the shot row is decided by typing. Settings, item by item.

## The look: Visual, Scene and Costume Style

Visual Style decides how every picture and every shot is drawn. The presets are grouped under 2D Animation, 3D Animation, Live-action and Stop-motion, and there is a custom style. Scene Style sets the world a story lives in and Costume Style what people wear; both are read with the script. All three are fixed once you continue. A few of the cards as they appear in Step 1:

Visual Style: Shot on Film

Visual Style: Family 3D Animation

Visual Style: Claymation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
Visual style presets, one by one · Custom visual style · Choosing visual style settings.

## Shape and length

Shape. 16:9 Landscape is the default and the recommended choice for film and large screens; 9:16 Portrait suits a phone-first vertical drama. It governs the video only: character pictures are always portrait and scene pictures always widescreen. Shape does not change the price, and one project cannot mix both. In practice a 9:16 shot is framed close to fill a phone, and a 16:9 shot is staged wider with more of the room, which also suits more than two people in a shot. Vertical or landscape.

Length. A SceneMixer segment is up to 15 seconds or, on the tiers that allow it, 30; the shots inside add up to the segment, and a segment over its tier’s cap is refused before you pay. The segment ceiling is a ceiling, not a target. Dialogue needs about three English words a second. Shot seconds and timing · 15 or 30 second segments.

## Keeping the same characters

In SceneMixer, each character is one entry with one approved picture and one voice, and every shot that shows them is made from that entry. The control that matters most is in the shot text: name who is in frame with an @ chip. A line that says “a woman turns to the window” gives the shot nothing to hold on to; one that says @Mara does. For a new outfit or an injury, add a state to the character and pick it for the segments where it applies. The same face in every shot · Costume changes and scene states.

## Resolution and price

In SceneMixer, resolution belongs to the tier, not to a separate setting. The default tier is 480P at 10 credits a second; Wan 3.0 also offers 720P (recommended) and 1080P, and the other families give a different look, some with shorter segments. Work this way: draft on the default tier, fix the text, then regenerate the segments you will deliver on the tier you will deliver on. A higher tier does not fix a bad shot; the text does. If you mix tiers inside one episode, the join has to re-encode onto one canvas and loses a generation of quality; if every segment came from the same tier, the streams are copied. Which quality tier · What comes out.

## Same text, different settings

| Same shot text | Setting changed | What you see |
| Two people argue across a kitchen table | 9:16 instead of 16:9 | Tighter framing on faces; less of the room |
| The same | Digital HD instead of Claymation | A different drawing of every cast picture and every shot |
| The same | Long take 30 s instead of Standard 15 s | Fewer, longer segments; more of the action in one generation |
| The same | A higher tier | More pixels or a different look; the same action |
| The same | @Mara written into the shot | The shot is made from her picture instead of an unnamed face |

That is why this page’s advice is the same as the editor’s: settle the project row first, because it cannot be changed, then work down. Text to movie · Text to video AI: writing the brief.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Medieval Castle; Costume Style Medieval Nobility; Visual Style Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: check each lead has the look you want before the storyboard; the cast is what every shot is made from. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

## FAQ

### What can I control in an AI text to video generator?

SceneMixer lets you control the type, look, shape and spoken language of the project; the cast; the segment length of each episode; the tier of each segment; and the seconds, framing and people in each shot.

### Can I change the visual style after the script is read?

No. SceneMixer fixes Visual Style once you continue. Start a new project with the same script and choose another.

### Does 9:16 cost more than 16:9?

No. SceneMixer prices video by the second, not by the shape of the frame.

### How do I keep a character the same?

Name them with an @ chip in every shot they are in. SceneMixer treats each character as one entry with one picture and one voice.

### Does a higher tier improve the shot?

It changes resolution and look, not what the shot shows. Fix the storyboard text first.

### Can I mix tiers in one episode?

Yes. Each SceneMixer segment uses the tier selected when you generated it, but a mixed join is re-encoded and loses some quality.

### How long can one generated clip be?

A SceneMixer clip runs up to 15 seconds, or up to 30 on the tiers that allow it. Longer video is segments joined.

### What does it cost?

SceneMixer charges 10 credits a second for video on the default tier, and joining is 1 credit a second; the price shows on the button before you press it.

Updated November 2, 2026


---

# AI UGC Ad Generator: UGC-Style Ads From a Product Photo

URL: https://scenemixer.com/guides/ai-ugc-ad-generator/
Make UGC-style ads with AI: vertical, phone-like, one generated person with your product. The look, six spoken hooks, a three-version test and what it cannot be.

# AI UGC ad generator

Quick answer: UGC means content a customer made. A generator cannot make a customer, so what it can make is a UGC-style ad: vertical, shot like a phone video, one person talking about your product in an ordinary place. In SceneMixer that is a product photo and one line, a portrait ad in the A day in the life or Show how it works tone, a generated person who speaks in the buyer's language, and your name on the end card. It has no library of creators, no face swap and no voice cloning, so call it an ad and label it as one. For a precise phone-camera look, take the wheel and edit the shot lines.

One line for the box next to the upload: A fragrance-free night cream in a glass jar; for people with dry skin who want one step before bed; filmed like a friend's phone video. Tone: A day in the life; portrait; 15 seconds.

## UGC-style, not UGC

The things that make a video read as user-generated are mostly things that are missing: no tripod, no lighting kit, no announcer voice, a place that looks lived in, a product being handled. Those can be written. What cannot be produced here is the part that gives UGC its trust: a real customer. People in your photos are removed, the cast is generated, and there is no face swap and no voice cloning. So:

- Call it a UGC-style ad, and let the platform's own made with AI setting say so. On SceneMixer, free accounts also carry a small AI label in the corner of the video.
- Do not write lines that claim a purchase, a result or a review. The script removes superlatives and unverifiable claims on its own, but that is a wording filter; the claims you make are yours. How do I make AI UGC-style ads?

## The look, written as shot lines

A one-click ad takes the tone and your one line as hints and writes the shots itself. For an exact phone-camera look, choose Take the wheel: you pay for the photo cleanup and the written script, land in the editor with the ad set up, and can rewrite the script and edit any shot. The look is then a handful of words per shot:

| UGC feature | Write it as |
| Selfie framing | Close-up, handheld, selfie-style framing; she holds the phone at arm's length |
| Natural light | Daylight from a window on her left; no studio light |
| A lived-in place | A small kitchen, a cluttered shelf, a bed with a rumpled throw |
| The product handled | She holds the jar up to the window, turns it so the label faces us |
| Casual speech | (casual, a little rushed): short lines, a pause, a half laugh |
| No polish | Fixed or handheld camera; no push-ins, no sweeping moves |
A 15-second example of three shots:

| Shot | Write |
| 1 | [Shot 1] 4s: Close-up, handheld, selfie-style framing. @Dana holds a glass jar up to the window light in a small kitchen. @Dana (casual): Okay, I wasn't going to post this, but… (voice: Dana) Ambient: a kettle clicking off. |
| 2 | [Shot 2] 6s: Medium close-up, handheld. @Dana scoops a little cream onto two fingers and rubs it on the back of her hand, turning the jar toward us. @Dana (easy): It's the only thing I put on before bed now. (voice: Dana) |
| 3 | [Shot 3] 4s: Close-up, fixed. @Dana smiles and sets the jar on a shelf next to a mug. Ambient: the room is quiet; a clock ticks. |
About three English words a second: lines that fit the seconds sound natural; lines that do not sound rushed. Shot seconds and timing · Writing dialogue in the storyboard.

## Six spoken hooks, with no results promised

| Hook | Example line | Tone |
| Confession | Okay, I wasn't going to post this, but… | A day in the life |
| Problem first | I keep losing my lunch box lid, so I tried this. | Show how it works |
| Show first | Watch what happens when I fold it flat. | Show how it works |
| Reaction | Wait, it fits in my coat pocket? | Funny |
| Routine | This is what I do before work now. | A day in the life |
| Question | Why did nobody tell me you could do this? | Twist |
Hooks state what a person does or sees, not what the product guarantees. Two to four spoken lines fit a 15-second ad. Hooks and cliffhangers.

## Three versions from one set of photos

UGC works by testing. The same one to four photos make three separate ads, each with its own quote; tone and format do not change the price on SceneMixer, length and quality do.

| Version | Tone | Length | What it tests |
| A | Show how it works | 15 seconds | Does the demo hold attention? |
| B | Funny | 15 seconds | Does the reaction beat the demo? |
| C | A day in the life | 30 seconds | Does a routine sell better than a trick? |
Post them a day apart and keep what holds attention. The same product can speak another language: a Spanish version is a second ad from the same photos. TikTok Shop videos · Posting to TikTok, Shorts and Reels.

## A store's product page

For a store page (a Shopify shop or any other), the input is the photos already on the product page: the product alone, in use, a detail, the pack. The first is the hero and the end card is built around it. Choose portrait for feeds and stories, landscape for the product page itself, and put the brand, the product name and the button on SceneMixer’s end card. AI product video generator · An ad for a physical product.

## What it will not do

- No creators or avatars. There is no list of faces to pick from and no way to put a real person on screen.
- No real reviews. The speaker is generated; it cannot be a buyer.
- No captions in the picture. Add them in the platform's own tool; nothing legible is drawn into the video. Subtitles.
- No audio upload. Only text, up to four photos and one video go in. What you can feed in.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Part | UGC-style choice | Why |
| Format | Portrait 9:16 | Phone feeds |
| Tone | A day in the life or Show how it works | A person living with the product, or handling it |
| Length | 15 seconds | One hook, one beat |
| Control | One-click, or Take the wheel for exact framing | Edit shot lines for handheld, selfie-style |
| Label | Say it is an ad; use the platform's AI setting | The speaker is generated |

## FAQ

### How do I make a UGC-style ad with AI?

Upload one to four photos of the product on SceneMixer, write one line, choose portrait, 15 seconds and A day in the life or Show how it works. For exact phone-style framing, take the wheel and edit the shot lines.

### Is it real user-generated content?

No. SceneMixer generates the person, and no real customer is involved, so it is a UGC-style ad. Label it as an ad.

### Can I pick a creator or an avatar?

No. SceneMixer has no creator library; the cast is generated to fit the buyer you describe.

### Can it clone a voice or swap a face?

No. SceneMixer removes people from photos, and there is no face swap and no voice cloning.

### What length works for a UGC-style ad?

Fifteen seconds is one hook and one beat; thirty seconds holds a routine.

### Can I test several versions?

Yes. The same photos on SceneMixer make separate ads with different tones and lengths; each has its own quote.

### Can I make it in another language?

Yes. The spoken lines and end card on SceneMixer follow the language you choose; a second language is a second ad.

### Can it work for a Shopify product page?

Yes. Use the photos from the product page in SceneMixer; choose landscape for the page and portrait for feeds.

### What does it cost?

One figure is quoted on SceneMixer before it starts and it is a ceiling. A 15-second ad on the default quality is the cheapest; ads are never the free first project.

Updated November 3, 2026


---

# AI Video: What It Is, How It Is Made and What It Can Do in 2026

URL: https://scenemixer.com/guides/ai-video/
What an AI video is, how text or photos become moving pictures, why clips are short, what is hard to get right, and how a finished episode is built from segments.

# AI Video: What It Is and How It Is Made

Quick answer: An AI video is a video whose pictures, and often its voices and music, are generated by a model from text, images or both instead of being filmed. Most models make a short clip at a time, so a video of any real length is a set of clips joined together, and the hard part is making the clips look like they belong to the same world. In SceneMixer a clip is a segment of 15 or 30 seconds; the cast, the locations and the props are fixed once, every segment is generated from that same set, and the segments are joined into an episode.

Try it: write a courier delivers a letter to the wrong flat on New Year’s Eve and stays for the countdown, make it at the shortest length, read the cast and the storyboard it produces, and use the free 5-second preview to see one real shot.

## How an AI video gets made

- An idea in words. A sentence, a script, or a novel. Photos or a clip can start it too.
- Words become shots. The text is broken into scenes, and each scene into shots with a camera framing, who is in frame, what happens and how long it lasts.
- Each shot is generated as a clip. A video model draws the moving picture; in a good setup it also generates the voices speaking the lines.
- The clips are joined. Music goes under the voices, subtitles can go over the picture.
- You fix what missed. A shot that came out wrong is changed and made again on its own.
Steps 2 and 5 are where a person spends their time, and in SceneMixer they are text: you read the storyboard and edit the words, which is free. Step 3 is the one that costs money, because every attempt is a separate call to a model. What is inside a project.

## Why clips are short, and how a long video is built

A video model holds a few seconds of motion together well and loses its grip as the clip gets longer. So a 60-second video is four 15-second segments, or two 30-second ones, joined. On the Segment length setting SceneMixer offers Standard 15 seconds or Long take 30 seconds; the price per second is the same either way, and 30-second segments can only be made by the models that support them. More, shorter segments give you finer control when something misses; longer ones hold one stretch of action in a single generation. 15 or 30 second segments.

An episode is segments joined, and a series is episodes: a short script gives a few episodes, a long one gives many. Making a multi-episode series · Long AI video generator.

## The hard part: staying the same

Every clip is drawn separately, so left alone a face drifts, a jacket changes colour and the room rearranges itself. The fix is not a better prompt, it is a shared set of references. In SceneMixer the script is read once into a cast, a prop list and a set of locations, you approve them, and every shot of every episode is generated from that same set. The system also builds an automatic AI 3D model of every scene, so people stay where they sat and the cuts make sense. Why a whole series stays consistent.

It is not perfect. When a face changes or characters swap sides between cuts, there is a checklist: face changed · swapped sides.

## What AI video does well, and what it does not yet

| Does well | Does not do (yet) |
| Worlds that would cost a fortune to build: a palace, a spaceship hangar, a rainy street at night | Draw legible text. Prices, dates and signs are not drawn reliably, so they go on an end card or in the spoken line |
| Anime, cartoon, claymation and film looks from the same script | Hold one unbroken take longer than a segment (up to 30 seconds) |
| The same scene in another language, with the lines spoken in it | Keep a real person’s likeness: photos of real faces and famous characters are refused |
| An ad from one to four photos | Be instant: a generation takes minutes, because the model takes minutes |

Why a generation takes time · Text on screen and wrong dialogue.

## Rights, labels and the badge

A video made on a free SceneMixer account carries a small AI badge in the bottom-right corner; turning it off is a membership benefit, and a shared link always carries it. Removing the badge does not remove the rules of the platform you post to: TikTok, YouTube and others have their own AI-content labels. Do AI videos need a label? · Can you monetize AI videos?.

## Three ways to start

| You have | You get | Read |
| A script, a novel or a one-line idea | A drama with a consistent cast and spoken dialogue | Upload a script |
| One to four photos of a product or a place | An ad of 15, 30 or 60 seconds | An ad from one photo |
| A clip you shot | A remake with new characters in a new look | AI video remake |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with one mark each and a voice; the stairwell and the balcony as locations; the letter as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Term | What it means in a SceneMixer project |
| Project | The settings, the cast and the list of episodes. Everything below inherits from it |
| Cast | Characters, locations and props, each with a description and a picture |
| Episode | A numbered part of the story with its own storyboard |
| Segment | One generation: about 15 or 30 seconds of finished video, with the shot text that produced it |
| Take | Every time you generate a segment the result is kept; one is current, and switching between them is free |
| Tier | The model and quality you generate on; the price per second depends on it |
| Join | Putting the current take of every segment into one episode file |

## FAQ

### What is an AI video?

A video whose pictures, and often voices and music, are generated by a model from text, images or both rather than filmed.

### How are AI videos made?

Words are broken into shots, each shot is generated as a short clip by a video model, and the clips are joined with music and sometimes subtitles.

### Why are AI video clips so short?

Models hold a few seconds of motion together well and lose consistency as a clip grows. Longer videos are made of several clips joined.

### How do AI videos keep the same character?

By generating every shot from one fixed set of references. SceneMixer reads the script into a cast, locations and props once, and every shot uses that set.

### Can AI video include speech and music?

Yes. SceneMixer speaks dialogue inside the video in any of 15 languages, one voice per character, and each episode is scored for you.

### How long does it take to make an AI video?

Writing and checking take minutes on SceneMixer; each generated segment takes minutes too, because the model takes minutes.

### How much does an AI video cost?

SceneMixer charges 10 credits a second for video on the default tier, and joining is 1 credit a second, so a 60-second episode is about 660 credits before retakes.

### Do I have to label AI video?

The platforms you post to have their own labels and you should use them. SceneMixer puts a small AI badge on by default for free accounts, and turning it off is a membership benefit.

Updated November 3, 2026


---

# AI Video Automation Software: What Runs Itself, What Needs You

URL: https://scenemixer.com/guides/ai-video-automation-software/
Which steps of making an AI video run by themselves, which take one click, and which still need you. From script to a joined episode, with the price of each.

# AI Video Automation Software

Quick answer: Most of the work in making an AI video can run without you: reading a script into episodes, building the cast and the locations, writing the storyboard, giving every character a voice, scoring the music, and generating every segment with one click. Three things stay with you on purpose, because they decide whether the money is well spent: approving the cast, reading the storyboard, and checking the first segment before you generate the rest. Posting the finished video is also yours.

Try it: paste a short script into SceneMixer, then use Generate all only after you have read episode 1’s storyboard and made one segment on its own.

## What runs without you

- Reading. A script, a novel or a one-line idea is split into episodes. Long books are read in two passes. Long scripts, read in two passes.
- Cast and locations. Characters, locations and props are found, described and drawn once, then reused in every shot.
- Storyboard. SceneMixer writes a shot list with seconds for each episode.
- Voices and music. One voice per character; each episode is scored; subtitles can be switched on.

## One click for the rest: Generate all

In SceneMixer’s episode editor, in the top bar next to Composite, Generate all appears while some segments have no clip. It shows the total price; press it and a dialog lists how many segments will be generated, their total seconds and the model. The segments are submitted one by one, each charged at its own price and refunded if it fails, in timeline order, on the model you picked. Segments beyond your account’s concurrency limit wait in line and start by themselves. Generate all · Generate all with a low balance.

## What does not run by itself, and why

| Step | Why it stays with you |
| Approving the cast | A wrong face or age is cheap to fix before video and expensive after |
| Reading the storyboard | It is what the video is made from; editing it is free |
| The first segment | One attempt tells you whether the other segments will hold; generating them all first can cost a lot |
| Choosing a tier | Draft low, finish high; the price per second depends on it |
| Posting | Each platform has its own labels and rules. Posting to TikTok, Shorts and Reels |

## How many jobs run at once

Your SceneMixer account runs a fixed number of generations at the same time, depending on your plan. Start more than that and the extra ones park in a waiting state and are promoted automatically as running jobs finish. Parking is not a failure and the credits are held, not spent. No plan changes how long a single generation takes, which is the model’s. Why a generation takes time.

## Automating a series, and a second language

Because SceneMixer fixes the cast once, a series stays consistent without per-episode setup. To publish in a second language, make a second-language version of the finished drama instead of starting again. Why a whole series stays consistent · Make a second-language version.

## What automation costs

On SceneMixer, reading is 4 credits per 1,000 characters, video is 10 credits a second on the default tier, and joining is 1 credit a second. A 60-second episode is about 660 credits before retakes. The one expense you can automate away is the wasted retake: fix the text before you regenerate. Spend fewer credits.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with one mark each and a voice; the stairwell and the balcony as locations; the letter as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Task | Automatic? | Where you step in |
| Split a script into episodes | Yes | Choose the script length and read the result |
| Cast, locations and props | Yes | Approve once; replace a picture if wrong |
| Storyboard | Yes | Read it; edit any line for free |
| Voices | Yes | Pick a different voice if you want |
| Music | Yes | Switch on or off |
| Generate every segment | One click | After the first segment is right |
| Join into an episode | One click | Choose which takes are current |
| Post to a platform | No | You, with the platform’s labels |

## FAQ

### What can be automated when making an AI video?

SceneMixer reads the script into episodes, the cast and locations, the storyboard, the voices, the music and generating every segment. Approving the cast and reading the storyboard stay with you.

### Can it generate a whole episode with one click?

Yes. SceneMixer’s Generate all submits every segment that has text and no clip, each at its own price and refunded if it fails.

### Why not automate approving the cast too?

A wrong face is cheap to fix before video and expensive after, so the cast is shown once for you to approve.

### Will it post for me?

No. Posting stays with you because each platform has its own labels and rules.

### What if my balance runs out halfway through Generate all?

The dialog shows the total before you start, and the guide on low balances explains what is left out.

### How many segments can run at the same time?

A number set by your SceneMixer plan; the rest wait in line and start by themselves.

### Can I run a whole series this way?

Yes. SceneMixer fixes the cast once, so every episode uses the same faces and places.

### Is it cheaper to automate?

Automation saves time, not credits. Savings come from fixing text before regenerating and checking one segment first.

Updated November 3, 2026


---

# AI Video Creation Platforms: Ten Platform-Level Things to Check

URL: https://scenemixer.com/guides/ai-video-creation-platforms/
Ten platform-level checks for an AI video creation platform: accounts, ownership, simultaneous generations, billing, storage, export, content rules, labels.

# AI Video Creation Platforms: Ten Things to Check Beyond the Clip

Quick answer: An AI video creation platform is more than the model behind it. Once you make more than one video, the platform layer decides your week: how you sign in, who owns a project, how many generations run at once, how credits and plans behave, what counts against storage, how you get files out, what the content rules do to a batch, and what is stamped on the result. This page lists ten checks at that level, with SceneMixer’s answer to each in the table and the details that need more than a line underneath. It does not rank platforms; it gives you the questions, and the numbers for the one you can try here.

Try it: take the ten checks in the table and put a tick or a question mark beside each for every platform on your shortlist. The question marks are where to read the help pages before you pay.

## Accounts and projects

Sign-in. There is no password. You continue with Google, with Apple, or with a six-digit code sent to your email, and a new address gets an account on the spot. Google can be linked later in Settings so that either route reaches the same account. Signing up and logging in.

Who owns a project. A project, its cast, storyboards and videos live in one account, and only that account can edit or generate. There is no team workspace and no shared editing. Reviewers watch through share links that open without an account and cannot edit. A share link does not expire by itself and has a daily play limit; it stops working when the segment, episode or project it points to is deleted. A project cannot be moved between accounts: hand over the script, the character pictures and the downloaded videos. Using SceneMixer as a team · Share links.

## Concurrency: how many things run at once

| Plan | Generations at once | Also |
| Free | 3 | Extras wait with their credits held and start by themselves |
| Pro | 5 | A paid plan raises the number, not the speed of one generation |
| Studio | 8 | Same model, same run time |
| Enterprise | 10 | The top three plans also join a priority queue ahead of other users |

SceneMixer’s limit counts every generation, not only video: segments, sheets, storyboards and composites. Three short interactive steps are never parked: writing a script with AI, suggesting a style and matching a voice. The pricing page lists the larger plans above these. Generating many segments at once.

## Credits, plans and what a bill can surprise you with

SceneMixer holds credits when you press a button, spends them when the result arrives, and returns them if a confirmed failure means nothing was produced; every movement is a line in the credit history. A plan grants credits at the start of each period into the same balance as packs, and they do not expire at the month’s end. A generation costs the same number of credits on every plan, there is no per-generation member discount, and there is no unlimited plan. Upgrading takes effect at once; downgrading mid-subscription is not possible; cancelling stops the renewal and you keep the plan until the period you paid for ends, with no pro-rata refund. What a paid plan changes.

## Content rules and what a batch can trigger

Each video generation blocked by the model provider’s content check counts in a 24-hour window. Past 12 blocks, video submissions from the account are refused for 1 hour; blocks within the number are refunded automatically and later ones are held for review. Only video is paused; sheets, storyboards and edits continue. If you run batches, this is the number to know. How the brake works.

## Storage, export and what is stamped on the result

Only what you upload counts against storage, and generated pictures, takes and episodes never do. The asset library shows everything across projects in six tabs and every entry can be downloaded; uploading an image there is a paid-plan feature. Deleting is final. The asset library.

A SceneMixer result carries a small AI badge and a metadata tag on a free account, off by default on a paid plan, and a shared link always carries the badge. Languages are three separate things: the interface language, the language your documents are written in, and the language your characters speak, which can be any of 15. Interface, script and dialogue language · AI-generated video: how it is marked.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Suburban Mansion; Costume Style Loungewear; Visual Style Realistic HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the three siblings with a description each; the kitchen and the study as locations; the will as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Check | Why it matters | SceneMixer |
| 1. Sign-in | Lost passwords and shared logins | Google, Apple or an emailed code; no password |
| 2. Project ownership | Who can edit, who can only watch | One account per project; share links for review; no shared editing |
| 3. Simultaneous generations | A batch queues behind itself | Per plan; extras wait and start by themselves |
| 4. Credits and plans | Surprise bills, expiring balances | Held then spent; plan credits do not expire; no unlimited plan |
| 5. Storage | Quotas spent on your own output | Only uploads count |
| 6. Export | Leaving should be easy | Full-quality download per segment and episode |
| 7. Content rules | A blocked batch can pause the account | A daily number of blocks, then a short pause for video only |
| 8. Labels | What is stamped, and who can remove it | Badge and metadata tag by plan; shared links always badged |
| 9. Help | Where an answer comes from | Ask inside the site, answering from its guides, and feedback |
| 10. Languages | Interface, script and speech are different | Three settings; speech in 15 languages |

## FAQ

### What should I look for in an AI video creation platform?

Beyond the clip: sign-in, who owns a project, how many generations run at once, how credits and plans behave, what counts against storage, how you export, the content rules, and what mark the result carries.

### Can a team share one project?

Not on SceneMixer: a project belongs to one account, and reviewers watch through share links without signing in.

### How many videos can I generate at once?

SceneMixer runs 3 on a free account, 5 on the first paid plan, 8 and 10 on the next two, and more on larger plans. The extras wait and start by themselves.

### Is there an unlimited plan?

No. Every SceneMixer plan works in credits. A plan grants credits each period and raises how many generations run together.

### Do plan credits expire?

No. SceneMixer puts them into the same balance as credits you buy, and they stay there.

### Can I export my project to another platform?

You can download every segment and episode from SceneMixer in full quality, and the character pictures. A project cannot be moved to another account.

### What happens if my videos are blocked by moderation?

SceneMixer refunds blocks within a daily number; past it, video submissions are paused for a short time. Only video is paused.

### Does the platform mark AI videos?

Yes: SceneMixer adds a small AI badge and a metadata tag on a free account, off by default on a paid plan, and a shared link always carries the badge.

Updated November 2, 2026


---

# AI Video Creation Tool: Six Things to Decide Before You Open It

URL: https://scenemixer.com/guides/ai-video-creation-tool/
Before opening an AI video creation tool, settle six things: material, size, languages, shape, names and budget. Why each changes what follows, with a filled-in sheet.

# AI Video Creation Tool: What to Prepare Before You Open It

Quick answer: The costly mistakes in an AI video creation tool are made in the first two minutes, before anything has been generated: the wrong shape, the wrong dialogue language, a story too big for one read, or a cast whose names clash. Text is cheap to change afterwards; several first-screen settings are not. So decide six things on paper first: what you start from, how big the story is, which languages it speaks, what shape and look it has, what its cast is called, and how many seconds of video you can afford. Each section below says what happens if you skip it, and the last one is a filled-in sheet you can copy.

Try it: before you open the tool, write one line for each of the six headings below. Any line you cannot write in under a minute is the one to settle first.

## 1. What you start from

SceneMixer has four doors, and each wants a different thing ready. Words: a script or novel as .txt, .docx or .md, or pasted text; the floor is about a paragraph, roughly 333 characters of English prose, because below that there is nothing to split into episodes. A link is not fetched, so paste the text itself. An idea: one or two sentences that carry genre, a lead who wants something, an opponent, the ending and the tone. Photos: one to four of the same subject for an ad. A clip: one MP4 or MOV for a remake.

If you skip this decision you will stand at the first screen deciding whether to write, paste or upload while the clock runs. If the material is not yours, stop here: you need the right to use everything you upload. Four ways to start · Writing the one-line brief · What you can feed in.

## 2. How big the story is

SceneMixer counts everything in equivalent characters, and English counts 3,333 characters as 1,000 because it needs about three times as many letters as Chinese for the same amount of story. A script under 10,000 counted characters is read in one pass; a longer one is read in two passes up to 100,000 counted characters, which for English is 300,000 raw characters. Beyond that the tool refuses before you pay rather than cutting the end silently. The read costs 4 credits per 1,000 counted characters, once for the whole project, and the length menu on AI Create prints its price beside each option.

Size also sets how many episodes you get: more characters mean more episodes of a minute or two, not longer episodes. If you skip this decision you either pay to read a book when you wanted a test, or find your script refused. A good first size is the “Short” option on AI Create, enough to see read, sheets, storyboard and video once. Script length caps · Choosing the script length.

## 3. Which languages it speaks

A SceneMixer project has more than one language. The interface is menus and buttons. The working text, the language your cast, scenes and storyboards are produced in, follows the interface when you confirm Step 1. The dialogue language is what the characters actually say in the clips, chosen in Step 1 from the 15 the site speaks. AI Create writes in the language you are using the site in.

Target region, dialogue language and content type shape what the read writes, and once the read has finished they cannot be changed in that project. The ways out are to edit the descriptions and lines, or to start again from the same script. If you skip this decision you can end with an English script spoken in the wrong language. Interface, script and dialogue language · What cannot change after the read.

## 4. The shape and the look

In SceneMixer, choose 9:16 Portrait for a vertical feed and 16:9 Landscape for a channel, a site or a screen. The shape governs video only and is fixed once you continue past Step 1; to make the other shape you start a new project from the same script. Character sheets are always 9:16 and location and prop plates always 16:9, whatever the project is.

The look is split in three on purpose: visual style (the medium), scene style (what the set is built from) and costume style (what people wear). Visual style is fixed after Step 1; scene and costume style are applied once, at the read, and after that you change them by editing the descriptions in Step 2 and regenerating those sheets. If you skip this decision, the first thing you will notice is a project in the wrong shape. Vertical or landscape · Choosing your style settings.

## 5. What the cast is called

Names are the glue between your text and your pictures. No two assets in a SceneMixer project may share a name, not two characters and not a character and a prop, and one spelling should be used for every person and place throughout the script. Use invented names: a famous character, a celebrity or a brand mascot can be refused, and a reference photo that looks like a real person is flagged. If you skip this decision you meet a rename later, which the tool handles for you across the storyboards, but not before the read has already given two people the same label. Why cast names must be unique · Celebrity, brand or franchise flags.

## 6. How many seconds you can afford

Decide seconds, not credits. Video is the largest line by far, charged per second, and everything else is rounding: SceneMixer’s default tier is 10 credits a second, and a 60-second episode with the join is about 660 credits, $4.11 at the lowest price. Plan a first episode, add room for retakes (a shot that misses is regenerated alone, not the episode), and draft on a low tier before finishing on a higher one. The free pieces cover the cheap steps: 15 setup images, the first storyboard and one 5-second preview. A new account starts with 50 credits.

If you skip this decision you press Generate all on a full episode before you know what one segment costs. The tool does check the balance first: if it does not cover the batch, nothing is submitted. Where credits go · Spend fewer credits.

## A filled-in sheet

Here is the sheet for a small first project. Copy the left column and fill in your own.

| Decision | Example answer | Can it change later? |
| Start from | A one-line idea turned into a script by AI Create | Yes: the draft and the script are editable |
| Size | The Short option; three episodes expected | Yes, by reading it again in a new project; the read is priced by length again |
| Languages | English interface, English dialogue | No: dialogue language is fixed once the script is read |
| Shape and look | 9:16 Portrait; Digital HD; one scene style; one costume style | Shape and visual style: no. Descriptions: yes |
| Cast names | Four invented names, one spelling each | Yes: rename updates the storyboards |
| Budget | One 60-second episode first, about 660 credits, plus room for retakes | Yes, one segment at a time |

You do not need to prepare camera directions, character art, voices or music. SceneMixer’s read writes the cast, the voices are matched for free and the episode is scored for you. Your first five minutes.

| Decision | If you skip it | Where it is set |
| Material | You stall at the first screen | Home page tabs |
| Size | A refusal, or a bigger read than you wanted | Counter under the script box; AI Create length menu |
| Languages | Wrong dialogue language | Step 1, before the read |
| Shape and look | A project in the wrong shape | Step 1; fixed after you continue |
| Names | Two people with one label | Your script; Step 2 |
| Budget | A batch you did not price | Button prices; balance in the top bar |

## FAQ

### What do I need before using an AI video creation tool?

A story or an idea, a rough size, the dialogue language, the shape you will post in, a few invented names, and a number of seconds you are willing to pay for. Camera work and character art are not needed up front.

### Can I change the language after the script is read?

Not the dialogue language, target region or content type: SceneMixer fixes them once the read finishes. You can edit lines and descriptions, or start a new project from the same script.

### Can I change vertical to landscape later?

No. SceneMixer fixes the shape once you continue past Step 1. Start a new project from the same script; cast can come across from your library.

### How long can my script be?

Under 10,000 counted characters SceneMixer reads it in one pass, and up to 100,000 in two passes. English counts 3,333 characters as 1,000, so the English limit is 300,000 characters.

### Do I have to write camera directions?

No. SceneMixer writes the storyboard from your script, and you can edit it for free. A one-line idea is enough to start.

### What is free before I spend anything?

SceneMixer gives 50 credits on sign-up, 15 setup images for the account, the first storyboard of your first drama and one preview of up to 5 seconds.

### Can I use real people or famous characters?

Avoid them. Famous characters, celebrities and brand mascots can be refused, and a photo of a real person used as a reference can be flagged.

### Do I need to install anything?

No. SceneMixer runs in the browser; the heavy work happens on servers.

Updated November 2, 2026


---

# AI Video Creator: How One Person Plans a Week of Episodes

URL: https://scenemixer.com/guides/ai-video-creator/
What limits how much video one person can make with AI: credits, simultaneous generations and your reading time. Per-plan numbers, a worked week, reusing a cast.

# AI Video Creator: How One Person Plans a Week

Quick answer: A solo creator’s output is limited by three things, and typing is not one of them: credits, how many generations your plan runs at once, and your own time spent reading storyboards. On the default tier a 60-second episode costs 660 credits, a segment takes ten to twenty minutes, and a plan runs from 3 to 10 generations at the same time or more. Put together, that gives a rhythm: do the thinking and the reading in text, which is free, start the rendering in waves, and use the waiting time for the next episode. This page does that arithmetic per plan, sets out a working week, and shows how reusing one cast is what turns a pile of clips into a series.

Try it: write down how many 60-second episodes you want this month and multiply by 660 credits. Compare that with the credits your plan grants, then ask how many segments you need running at once. Whichever is tighter is your real limit.

## The three limits, in the order you hit them

1. Credits. SceneMixer charges video by the second and nothing else comes close. At 10 credits a second plus 1 credit a second to join, a 60-second episode is 660 credits, about $4.11 at the lowest credit price, before the read and before retakes. A plan grants its credits at the start of every billing period, into the same balance as packs you buy; they do not expire at the end of the month.

2. Simultaneous generations. Your account runs a fixed number at once and parks the rest. A 60-second episode is at least four 15-second segments, so at today’s limits a free account needs a second wave for the fourth segment, and any plan that runs four or more at once starts all four together. Each wave is typically ten to twenty minutes. A plan changes how many run together, not how long one takes.

3. Your own reading time. The storyboard is the document you are paid to check. Reading and fixing an episode’s shots costs no credits but real minutes, and it is the step nobody can do for you.

| Plan | Credits each period | Generations at once | Waves for a four-segment episode |
| Free | 50 credits once, at signup | 3 | Two: three start, the fourth waits |
| Pro | 700 | 5 | One |
| Studio | 2,000 | 8 | One |
| Enterprise | 5,000 | 10 | One |

The pricing page lists every plan, including the larger ones above these, and the top three plans also put your jobs ahead of other users’ in the shared queue. What a paid plan changes.

## A working week

This is a pattern built from the product’s rules, not a promise of output. The idea is that every step which waits on a model sits next to a step that waits on you.

| Day | Your hands | Running on the server |
| Mon | Write or paste the script for the next series. Read the cast. Fix descriptions. Pick the three styles on purpose | The read, then the setup pictures |
| Tue | Storyboard episode 1 and read every shot. Shoot the first segment and check faces, room and line | Storyboard; one segment |
| Wed | Start the remaining segments of episode 1 in one go. Meanwhile storyboard episode 2 | A wave of segments; the next storyboard |
| Thu | Join episode 1, fix single shots by editing their text, regenerate only those | Join; individual segments |
| Fri | Download, add titles in your editor, post. Start episode 2’s segments | Episode 2’s wave |

Two rules carry the week. Finish episode 1 completely before you run the rest, because a cast problem found in episode 1 costs one regeneration and found in episode 8 costs seven episodes of them. And draft on a low tier, finish the keepers on a higher one, using the same tier for every segment of an episode so the join copies the streams rather than re-encoding. Making a multi-episode series.

## Reuse is what makes it one person’s work

Inside a SceneMixer project every episode shares the same cast, locations and props, so nothing is redrawn. Across projects, Pick existing in step 2 lists every finished character, scene and item from your other projects. Choosing one moves everything over: the name, the description, the pictures, the voice and the state pictures, and the new name is written everywhere the old one stood. It is free, needs no plan, and can be undone with Restore the original. A copy keeps the visual style it was drawn in, so reuse within one style, or regenerate the entry in the new project’s style.

A second season can be a new project that reuses the cast, or more pages added at the bottom of step 3 with Add an episode, which reads only the new pages and reuses the cast of the earlier episodes. Reusing a character, scene or voice · Make the next episode.

## Posting and review without a team

A project belongs to one account and there is no shared editing, so a solo creator’s pipeline is simple: you own the project, reviewers watch a share link without signing in, and finished episodes leave as downloaded files. Your downloads carry the AI badge on a free account and not by default on a paid plan; a shared link always carries it. To lay your own music under an episode, turn the Music switch off in the editor and download the version without it. Using SceneMixer as a team.

For channel formats see running a faceless YouTube channel, a kids cartoon channel and posting to TikTok, Shorts and Reels.

## Where the hours really go

Not in the rendering. The waits are minutes per wave and they overlap with your reading. The hours go to three places that are all free to repeat: the cast descriptions (fix a face once), the storyboard text (name who is in frame, one action per shot) and the single shots that miss. A creator who keeps those three tidy spends credits on video and almost nothing on rework. Spend fewer credits on an AI drama · My character’s face changed: a checklist.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Cyber Metropolis; Costume Style Near-Future Techwear; Visual Style Game Engine Realism; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the courier and the hacker as the two leads; the night market and the server room as locations; the sealed drive as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Limit | What sets it | What you can do about it |
| Credits | Seconds rendered and the tier | Draft on a low tier; fix text before re-rendering; shoot one segment first |
| Simultaneous generations | Your plan | Start everything and let the extras wait; they start by themselves |
| Reading time | How many episodes you check | Finish episode 1 first; keep the cast tidy |
| Cast | One entry per character | Pick existing from earlier projects; it is free |

## FAQ

### How much video can one person make with AI?

On SceneMixer it is set by credits, by how many generations your plan runs at once, and by your reading time. A 60-second episode is 660 credits on the default tier, and a segment takes ten to twenty minutes.

### How long does one episode take to render?

A 60-second SceneMixer episode is at least four 15-second segments. If your plan runs four at once they render together, in typically ten to twenty minutes, then the episode is joined.

### Can I make a series with the same characters?

Yes. Every episode in a SceneMixer project shares one cast. For another project, Pick existing in step 2 moves a character, scene or item over for free.

### Can I work with an editor or a teammate?

A SceneMixer project belongs to one account. Reviewers watch share links without signing in, and an editor works from downloaded files.

### Do plan credits expire?

No. SceneMixer grants them into the same balance as credits you buy, and they stay there.

### Does a bigger plan make a video faster?

No. A bigger SceneMixer plan raises how many of your generations run at once; one segment takes the same time on every plan.

### What is the cheapest way to plan a week?

Do the reading and fixing in text, which is free, draft on a low tier, finish episode 1 completely before running the rest, and regenerate single shots, not whole episodes.

### Can I add my own music?

Turn the Music switch off in SceneMixer’s editor, download the version without it, and lay your own track in your editing software.

Updated November 2, 2026


---

# AI Video Editing: Change a Line, a Shot or a Whole Episode

URL: https://scenemixer.com/guides/ai-video-editing/
AI video editing by size of change: one spoken line, one shot, or a whole episode. How each is done, what gets made again, what it costs and what is kept.

# AI Video Editing in Three Sizes

Quick answer: The useful way to think about AI video editing is by the size of the change. A single spoken line is fixed on the finished clip by redubbing it. A single shot lives inside a segment, and fixing it means editing that shot’s text and making the segment again, or applying a trim or one of eight picture changes to the clip. A whole episode is changed by one written request run across every segment, and only the segments that changed are made again. This guide takes the three sizes in turn: how to do each, what is re-made, what it costs and what you keep.

## Three sizes at a glance

| Size of change | The unit that changes | How | What is re-made | Cost shape |
| A line of dialogue | One segment’s spoken lines | Edit this clip, Redub; or edit the line and regenerate | Sound only with Lip sync; the whole clip with Re-perform or a regeneration | 1 credit a second of clip for lip sync; 2 times the tier price for re-perform |
| A shot | A segment, because a segment is one generation holding several numbered shots | Edit that shot’s text and regenerate; or Trim; or AI visuals | That segment only | The segment’s seconds at the tier price; trim 5 credits |
| A whole episode | The text of every segment you pick | AI edit then Apply to the whole episode… | Nothing until you regenerate; then only the segments that changed | 1 credit per revised segment, then the segments you choose to make again |
SceneMixer’s three edit sizes are nested. A change to the whole episode is a set of changes to segments, and a change to a segment can be as small as one line. The skill is choosing the smallest size that does the job, because each step up re-makes more and costs more.

## A line: redub it, or rewrite and regenerate

The fastest fix for a wrong word in SceneMixer is on the finished clip. Select the segment, open Edit this clip, choose Redub, and every spoken line is listed with its speaker, a voice pill and a text box. Change the words, check the voice, choose a method and press Redub into a new version.

| | Lip sync | Re-perform |
| Picture | Unchanged; the mouths are moved to the new audio | Played again with the new lines; can change a little |
| Sound | All the segment’s lines are voiced again from the start, with short pauses. The clip’s own ambience is not kept | Timing and delivery fit the words |
| Price | 1 credit per second of the whole clip | 2 times the clip tier’s per-second price |
| Limit | The new lines must fit the clip: too long and it stops before charging | Clips of 15 seconds or less only |
Two details change how you work. Because lip sync voices every line, a one-word fix still re-voices the segment, so read the other lines in the new version. And the changed words are written back into the storyboard text, so if you regenerate the segment later it says the new line. A box you empty is not a deleted line: the original words come back when you click away. If the delivery is wrong rather than the wording, rewrite the line in the storyboard and regenerate, which costs the segment’s full price but can change the performance. Change a line after the video is made.

## A shot: the unit is the segment

A SceneMixer segment is one generation of 15 or 30 seconds, and inside it the storyboard numbers the shots with durations that add up to the segment’s length. You cannot regenerate shot 3 alone; you edit shot 3’s text and make the segment again, and the others in the episode are untouched. Every take is kept, so you can compare and switch back.

| The problem is | Use | Why |
| The clip starts early, runs on, or stumbles in the middle | Trim, Keep or Remove | Cut to the exact frame; nothing new is generated; the original stays |
| The right shot, but the colour, time of day, weather, background people, lights, a scar, an expression or who plays a part is wrong | AI visuals | A fixed menu of eight changes that come out reliably; everything else about the shot carries over |
| The framing, camera angle or where people stand is wrong | Edit the shot text, then regenerate the segment | No fix on a finished clip can change them |
| A person who should not be there, or words that appeared in the picture | Edit the shot text and regenerate | Same reason |
| A good clip, just soft | Enhance | The same take, one resolution step up |
On a 30-second clip the picture changes apply to the first or the second half, and you pay for that half only. Each kind of fix keeps one version per segment, a second trim replaces the first trim, and the generated clip underneath is kept for as long as a fix made from it exists. Edit this clip · How the versions fit together.

## A whole episode: one request, every segment

Some notes apply from the first segment to the last: a way of speaking, a word someone keeps using, a sound that should sit under a whole scene, a pace that drags. For those, open SceneMixer’s AI edit on any segment, write the request and press Apply to the whole episode…. Everything it can revise is listed and ticked; untick what should stay. The Start button carries the most the run can cost.

- The request runs on each picked segment against that segment’s own text. It never rewrites the episode as one block.
- Each row comes back as revised, no change needed or failed. A segment that needs no change is not charged, and nor is a request the text cannot fulfil.
- Open a row to see the difference line by line. You can stop at any time; segments not started are dropped without charge.
- Save every revision at once, or save row by row. Nothing reaches your script until you save, and History can put the old text back.
- Saving changes the text only. The segments still play the old takes, so regenerate the ones you care about, and then export the episode again.
The price is 1 credit per revised segment, then the segments you regenerate at their normal price. Revise the whole episode: one request, every segment.

## The note decides the size

| The note says | Size | First move |
| “She should say Friday.” | A line | Redub with Lip sync |
| “He looks terrified in the second half.” | A shot | AI visuals, the expression change, on the second half |
| “Cut the pause before the last line.” | A shot | Trim with Remove |
| “It should be raining in this scene.” | A shot, or several | AI visuals for the weather; or a text request over the scene |
| “Everyone sounds too formal.” | An episode | Apply to the whole episode… with one request, then review |
| “The ending lands wrong.” | A segment or two | Edit the text of the last segments, then regenerate those |
| “Make it a different genre.” | A new project | The type is fixed after the read; start again with the same script |
The last row is the edge of the system. A SceneMixer project’s type and aspect ratio are decided before the text is read and cannot be edited afterwards, because the episodes were cut and written for them. For another shape, start a new project with the same script; the cast can come across from the asset library.

## The safety nets

Three things make larger edits less risky than they look. Every generation is kept as a take, switching between them is free, and takes that are not active are cleaned up only after 90 days. The text history keeps the original and the text from before your last five changes. And an AI revision is shown to you, with a line-by-line difference, before it reaches the script.

In SceneMixer, the one button that throws work away is regenerating an episode’s storyboard. It rebuilds every segment, discards the finished videos and their takes, and is not refunded. If the problem is in the wording, edit the wording. Regenerate a storyboard: what you lose · Text history.

## FAQ

### Can AI edit a video after it is made?

Yes, in set ways in SceneMixer: trim it, change one thing in the picture from a menu of eight, change a line or a voice, raise the resolution one step, or revise the storyboard text and make the segment again.

### What is the smallest edit?

A single spoken line, fixed with a lip-sync redub, or a trim that cuts a moment. Both keep the picture you already have.

### Can I fix one shot without remaking the whole segment?

Not by regenerating. A SceneMixer segment is one generation. You can trim it, apply one of eight picture changes to it, or edit the shot’s text and make the segment again.

### How do I change something across the whole episode?

Open SceneMixer’s AI edit, write one request and press Apply to the whole episode. It runs on each segment you pick, shows the differences and saves nothing until you do.

### Does editing the text change the video?

No. The clip keeps playing the old take until you regenerate that segment.

### What does editing cost?

Editing text by hand is free in SceneMixer. An AI revision is 1 credit. A trim is 5 credits, lip sync is 1 credit a second, and picture changes cost 2 times the tier’s per-second price.

### Can I change the genre or the aspect ratio of a finished project?

No. SceneMixer fixes both once the text has been read. Start a new project with the same script.

### Will I lose my earlier version if I edit?

No. Every SceneMixer fix is a new take and the clip you started from is kept. Only regenerating the whole storyboard removes finished videos.

Updated November 2, 2026


---

# AI Video Editing Software: In the Browser, and Out to Your Editor

URL: https://scenemixer.com/guides/ai-video-editing-software/
Where AI video editing happens here, what it covers compared with timeline software, and exactly what to download to carry your work into another editor.

# AI Video Editing Software: What Happens Where

Quick answer: In SceneMixer the editing happens in a web page, not in software you install. The episode editor lets you rewrite a segment’s text, trim a clip to the frame, change one thing in the picture, redub a line and switch between takes, all on servers, from a computer or a phone. It is not a timeline editor, so colour grading, titles over footage, transitions and mixing in your own clips are done elsewhere. This guide separates the two: what the in-page editing covers, and a hand-off sheet for what to download, in which form, to finish the video in other software.

## Where the editing happens

The SceneMixer editor is one page with three working areas. On the left, the characters and places the episode uses. In the middle, the storyboard text of the segment you selected, shot by shot with durations. On the right, the player, with the edit controls under it, and along the bottom a timeline of the episode’s segments. On a phone the same things stack into views you swipe between. Nothing is installed, and the work runs on servers, which is also why a generation carries on if you close the tab. A tour of the episode editor · Cloud or local, and where your files go.

Because the editing is attached to generated segments, its unit is the segment, a take of 15 or 30 seconds, not the frame on an open timeline. That is the difference between this and timeline software, and the next section lists it job by job.

## Timeline software jobs, one by one

| A job in timeline software | Here | Notes |
| Cut the start or end of a clip | Trim with Keep | Cut to the exact frame, picture and sound together; 5 credits; the original stays |
| Remove a moment from the middle | Trim with Remove | Same price. A segment keeps one trimmed version, so a second trim is made from the trimmed picture |
| Replace a line of dialogue | Redub | Lip sync keeps the picture; Re-perform acts the clip again |
| Swap in a different take | Version history | Free. The active take plays, downloads, shares and joins |
| Change a colour, the weather or the time of day of a thing in the shot | AI visuals | One of eight changes, not a free-text box |
| Raise the resolution | Enhance | One step up on the same take |
| Add captions | Subtitles | Burned in or as an SRT file; the words are your script’s |
| Colour grading, LUTs, keyframed effects | Not offered | The look is set by the visual, scene and costume styles before generation |
| Titles and lower thirds over footage | Not offered | Video models do not write legible text. An ad has an end card for the name and the call to action |
| Transitions and effects between clips | Not offered | Cuts inside a segment are written into the storyboard text; segments are joined as they are |
| Your own clips on the timeline | Not offered | An uploaded video can start a remake, which reads it and builds a new storyboard but never hands your footage to the video model |
Read down the right-hand column and the idea is clear. The editing is about repairing generated clips cheaply and exactly; the creative finishing that depends on your own footage, typography or grade belongs to software built for it, and the hand-off below is how you get there. Video editor with AI puts the fixes in price order.

## The hand-off sheet: what to download for what

| You want to | Download | How | Watch for |
| Finish a whole episode in another editor | The exported episode | Export episode once every segment is done, then Download in Exports | The segments are joined without re-encoding when they all came from the same tier. Mix tiers and the join re-encodes onto one canvas, which costs a generation of quality |
| Re-time or re-order clips yourself | The individual segments | Download under the player, or Assets → Videos for every clip across projects | It is the active take that downloads, so set your versions first |
| Use your own music | The episode or clip without the score | Turn the Music switch off before downloading | With it off, a clip that has the episode’s score as its own track downloads without it, and an exported episode gives the dialogue-enhanced version without music |
| Style the captions yourself | The episode plus the caption file | Tick Add subtitles on export, then Download SRT | The SRT is plain text with times, with your script’s words. Accounts that have bought credits or a plan only |
| Have captions already on the picture | The burned-in version | Download with subtitles | They follow the Music switch: with it on you get the version with the episode’s score |
| Use the pictures | Character, scene and item images | Assets, any tab, Download | Sheets are 9:16 for characters and 16:9 for places and props, whatever the project’s shape |
| Rework the text elsewhere | The storyboard text | Copy it out | Check the links after you paste it back in |
Download always gives the original, the full-quality file. The clip you see playing on the page is a lighter copy of the same picture made for quick streaming, so judge sharpness only after you have downloaded. Download: the original vs the play version.

## What travels with the file

The AI badge. A free SceneMixer account’s videos carry a small badge in the corner and a line of AI-generated-content information in the file. The badge is part of the picture, not an overlay, so it survives download, re-upload and re-encoding. On a paid plan neither is added by default, and you can switch them on. A link shared through the site always shows the label. Decide this before you generate, because it is applied at delivery: switching it later only affects videos you generate afterwards. Why shared videos carry the badge.

Resolution. It comes from the tier you generated on, shown in the picker beside the price, 480P, 720P or 1080P for the Wan family. Check frame rate, codec and audio layout in your editor’s file properties, because they are not promised on any page. What comes out.

Length. The export is priced by the finished length of each active take, so a trimmed segment is cheaper to join. A shared page is public and indexable, so treat a share link as publishing, not as a private hand-off; to move a file to one person, download it. Compositing and sharing an episode.

## A round trip that works

- Fix inside first. Repairs on a generated clip are cheapest and most exact inside SceneMixer: trim, redub, one picture change, or a text edit and a regeneration.
- Settle every segment on one tier if you will export the episode, so the join copies the streams instead of re-encoding.
- Decide on music and captions and set the Music and Subtitles switches to match what the other editor will add.
- Export, download, and finish outside: grade, titles, your own music, other clips.
There is no way back in. Nothing in the editor takes an edited file back as a segment, so keep the downloaded original of anything you care about and treat the outside edit as the last step.

## FAQ

### Is it AI video editing software I download?

No. SceneMixer’s editing is a web page and the work runs on servers. You can use it from a computer or a phone, and you download only the files you want to keep.

### Is there a timeline like in video editing software?

SceneMixer has a timeline of an episode’s segments, not a frame-level editing timeline. You edit one segment at a time. For frame-accurate cuts between clips, export and use an editor built for it.

### Can I add my own footage to the video?

Not into a SceneMixer episode. An uploaded video can start a remake, which reads it and builds a new storyboard, but your footage is not used as a clip.

### What file do I get when I download?

SceneMixer gives you the original, full-quality file of the active take, for a segment or for an exported episode. The page plays a lighter copy.

### How do I get the video without music?

Turn SceneMixer’s Music switch off before you download. A clip that has the episode’s score as its own track then downloads without it, and an exported episode gives the version without music.

### Can I download subtitles as a separate file?

Yes, as an SRT, if the SceneMixer account has bought credits or a plan at least once. Tick Add subtitles when you export the episode.

### Will the badge be on a file I take to another editor?

On a free SceneMixer account, yes, burned into the picture. On a paid plan it is off by default. It is applied when the video is generated.

### Can I bring my edited video back into SceneMixer?

Not as a segment. Finish the outside edit last.

Updated November 2, 2026


---

# AI Video for Authors: Book Trailers, Characters, Screenplay Scenes

URL: https://scenemixer.com/guides/ai-video-for-authors/
What a writer can make with AI video: a book trailer, a look-book of characters and places, a screenplay scene, a pitch teaser. File types, limits, what it cannot do.

# AI video for authors, novelists and screenwriters

Quick answer: An author's own text is the input, so a writer can do more here than most: a book trailer (a 60 to 90 second story trailer or a cover trailer with a pre-order button), a look-book of the characters and places as drawings, a screenplay scene played on screen to test pacing, and a pitch teaser. Paste text or drop a .txt, .docx or .md file (PDFs are refused); a project reads roughly 50,000 words of English at a time, so a novel goes in as volumes or as the chapters you choose. It does not narrate your book or clone your voice: it takes text, up to four photos and one video, never audio.

A first-page trailer script to paste into a new project: NARRATOR (voice-over): In the town of Halloway, nobody has opened the lighthouse door in forty years. Tonight, someone knocked from inside. Then three moments that raise a question, and a last line.

## Five things a writer can make

| Make | Start from | Result | Guide |
| A story trailer | A one-page trailer script you write | Your characters and settings filmed in 60 to 90 seconds; title spoken by a narrator | Book trailers |
| A cover trailer | Your cover photo, one line, an Ad Video project | 15, 30 or 60 seconds, ending on an end card with title, tagline and Pre-order now | An ad from one photo |
| A look-book | The character and place descriptions from your book | Character sheets and scene images you can show a cover designer or an editor | Character sheet template |
| A scene on screen | One scene of your screenplay or chapter | A storyboard with seconds and a moving version of each segment | Previs and animatics |
| A pitch teaser | A logline and a trailer script | A pilot or a one-minute trailer for a pitch | Pilot and pitch trailer |
Book to movie AI · Can AI turn a book into a movie?.

## For novelists: what goes in

- Text. Paste it into SceneMixer, or drop a .txt, .docx or .md file. A Word file keeps its text and paragraph order and loses styles, comments and images. Which script files work.
- A length the project can read. One project reads up to 300,000 characters of English, roughly 50,000 words; a script over the limit is split into volumes and each volume is its own project. Length caps.
- The chapters you choose, not the whole novel: pasting a whole novel gives you the whole novel, episode by episode. For a trailer, write a one-page trailer script instead.
- Narration kept as voice-over. Your narrator can be heard and never seen. Voice-over and captions.
Upload a script and parse it · Building the world before the storyboard.

## For screenwriters: a screenplay as input

A screenplay works as input in SceneMixer like any other script. Export it as plain text or Word (a Fountain file is plain text: rename it to .txt) or paste the pages. Keep the scene headings (INT. KITCHEN – NIGHT); they tell the reading where and when each scene happens. For lines that must be spoken word for word, put the name and the line on one line: MARA (quietly): Read it before you say no. Use one spelling for each name, because MARA in one scene and MRS. CHEN in another becomes two characters. A feature is longer than one project reads at once, so adapt it in parts, one sequence at a time. Using a screenplay as input · Screenplay template.

To visualise a screenplay, paste one scene, generate its storyboard (numbered shots with framing and seconds) and read the totals: they tell you early whether the scene runs long, and the seconds can be changed for free before anything is rendered. The clips are generated sketches, not exact lenses or performances. Shot list template · What is previs?.

## The look-book is the quiet win

Every character on SceneMixer is a full-length sheet on a plain background, drawn from the description you write: face first, one anchor nobody else has, then build and one outfit. The first 15 setup images on an account are free, so a writer can see the cast of a novel for nothing before any video is paid for. The same sheets then carry through every shot. Places are drawn the same way. Use them for a pitch deck, a cover brief or your own sense of the book. Describing a face · Story bible template · AI character sheet generator.

## A three-day plan for a book trailer

| Day | Do | Cost |
| 1 | Write the one-page trailer script: a narrator, three or four moments that raise questions, a last line. | Free |
| 2 | Cast the characters and the place from your own descriptions; read the free storyboard. | Free setup images, free first storyboard |
| 3 | Render at the default tier, fix single shots, put it together, post. | Video is 10 credits a second plus 1 credit a second to join |
A 60-second trailer on SceneMixer is four segments of 15 seconds, or two of 30. Post 16:9 for YouTube and your site, 9:16 for book communities on short-video feeds. Posting to TikTok, Shorts and Reels.

## What it will not do for a writer

- No audiobook. Audio does not go in and your voice is not cloned; text goes in and a cast is generated. AI audiobook to video.
- No book cover. The cover photo can be the hero of a cover trailer; designing a cover is a different job.
- No text in the shots. The title reaches the viewer by the narrator's voice or the end card, because generated video does not write legible text.
- No characters you do not own. Celebrity faces, brands and well-known franchises trigger an IP check; upload your own manuscript. Content rules.
- No exact lenses. The clips show intent; a real shoot still plans its own camera.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday, Smart Casual; Visual Style Shot on Film, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Rewrite each character's description to match the book (the scar, the coat, the colour of the eyes) and regenerate; the narrator is set to voice only. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| You are | Start with | Then |
| A novelist with a finished book | A one-page trailer script, not the whole book | Cast, storyboard, render; or a cover trailer from the cover photo |
| A novelist mid-draft | A chapter as a first episode, to see the cast | Use the look-book to find the characters |
| A screenwriter | One scene as .txt or .docx, headings kept | Read the storyboard seconds; adapt the feature in parts |
| A writer pitching | A logline and a one-minute trailer | A pilot, if the pitch asks for one |

## FAQ

### Can an author make a book trailer with AI?

Yes, two ways on SceneMixer: a story trailer from a one-page trailer script, or a cover trailer from a photo of the cover that ends on an end card with title, tagline and a pre-order button.

### Can I turn my novel into a video?

You can turn chapters into episodes with your own characters. A SceneMixer project reads up to 300,000 characters of English at a time, and longer text is split into volumes.

### What file types can I upload?

Paste text or drop a .txt, .docx or .md file. SceneMixer refuses PDFs: export to Word or copy the text.

### Can I use my screenplay?

Yes. Keep the scene headings, put each line as NAME: line, rename a Fountain file to .txt, and adapt a feature in parts.

### Will the characters look like the ones in my book?

As close as your descriptions: write the face, one anchor, the build and one outfit, then regenerate the sheet until it matches.

### Can it read my book aloud?

No. Audio does not go into SceneMixer, and your voice is not cloned; the characters speak with generated voices.

### Is it free to try?

New SceneMixer accounts start with 50 credits, the first 15 setup images are free, the first storyboard on your first drama is free and there is a 5-second preview.

### How much does a trailer cost?

Video on SceneMixer is priced by the second: 10 credits a second on the default tier plus 1 credit a second to join; a 60-second trailer is 60 seconds.

### Can I use someone else's characters?

No. Well-known characters and franchises trigger SceneMixer’s IP check; use your own manuscript.

Updated November 3, 2026


---

# AI Video for Small Business: What to Make From Four Photos

URL: https://scenemixer.com/guides/ai-video-for-small-business/
A small business does not need a crew: four photos and one line make a 15-, 30- or 60-second ad with your name on the end card. What to make first, by business type.

# AI video for small business

Quick answer: A small business needs three videos: who we are (the place), what we sell (the product) and why come now (the offer). All three can come from the same four photos and one line: the ad is written, shot with sound and finished on an end card carrying your name and a button, in about ten minutes for a 15-second spot. Start with a 30-second portrait A day in the life for a place, or a 15-second Show how it works for a product. The price is quoted before it starts and is a ceiling, and ads are never the free first project.

One line for the box next to the upload: A family barbershop on the corner of Elm Street; walk-ins welcome; for people who want a proper haircut without an appointment. Tone: A day in the life; 30 seconds; portrait.

## Three videos from one set of photos

| Video | Job | First photo | Tone | Length |
| Who we are | Make the place feel real | The shopfront with its sign, in daylight | A day in the life, Warm | 30 seconds |
| What we sell | Show one thing working | The product alone, large in the frame | Show how it works, Minimal | 15 seconds |
| Why come now | An offer with a date | The same hero photo; the offer goes on the end card | Funny, Twist, Warm | 15 seconds |
The third needs no new shots: edit the end card (Half-price Tuesdays) and rebuild the final cut. The end card · An ad from one photo.

## By business type

| Business | Photos | Tone | Length, format | End card button | Guide |
| Cafe, bakery, restaurant | The dish close; the room | Mouth-watering close-ups, Warm | 15 or 30 seconds, portrait | Book a table | Restaurants |
| Barber, salon, gym | The room from the door; the sign | A day in the life, Show how it works | 30 seconds, portrait | Book an appointment | Shops and venues |
| Boutique, gift shop | The shopfront; two products | A day in the life, Minimal | 30 seconds, portrait | Visit us | Clothing |
| Online shop | The product alone, in use | Show how it works | 15 seconds, portrait or landscape | Shop now | Product videos |
| Plumber, electrician, cleaner | The finished job, or the van with its sign (say in the line that it is a service, not a van for sale) | Show how it works, Warm | 30 seconds, landscape | Call today | What goes in |
| Consultant, B2B service | A clean screenshot of the result screen, or the office | Show how it works, Minimal | 30 seconds, landscape | Book a call | Apps and sites |
| Car dealer or garage | The car, front three-quarter | Epic, Minimal | 30 seconds | See the car | Cars |
The first photo is the featured subject and fills the frame; up to three more become angles, details or the places the ad is shot in. People in the photos (customers, staff) are removed; the ad's cast is generated. Which photos work.

## A church or a community group

An event or a service is best told as an invitation: a photo of the building or the hall, one line (Our community hall, Sunday gathering at 10, all welcome; coffee after), the Warm tone, 15 or 30 seconds, and Join us Sunday on the end card. A sermon is audio, and audio cannot be uploaded: only text, up to four photos and one video go in. If you have the text of a talk, it can be turned into an illustrated story with a cast, as the podcast and Bible-story guides describe. What you can feed in · From a transcript · A Bible story series.

## LinkedIn and other business posts

For a business feed, the video is an explainer: one promise, one person, one screen. Upload a clean screenshot of the result screen (one screen, roughly phone- or computer-shaped, no real faces), say who it is for in the one line, pick Show how it works, 30 seconds, landscape. Check the post composer on the day for the shapes it takes; the picture is chosen in the pickers, not afterwards. An ad for an app or website · A training video.

## A month from one shoot

| Week | What goes out | How it is made |
| 1 | A 15-second teaser of the hero product | One-click ad, 15 seconds, portrait |
| 2 | The 30-second place ad | One-click ad from the shopfront photos |
| 3 | The same place ad with an offer | Edit end card, rebuild the final cut; no reshoot |
| 4 | A second-language version | A second ad from the same photos |
Start by drafting the 15-second ad on the lowest quality, because the quote follows length and quality; then raise the quality for the one you will pay to promote. Length, format and quality · What you actually pay.

## What to have ready

- Your brand name legible on the sign or the package: it is copied letter for letter onto the end card.
- One line: what it is and who it is for; it wins over the photo when they disagree.
- A tagline and a button you would say out loud; superlatives are removed.
- Four photos of the same thing, not four different things.
- The right to use them: JPG or PNG under 10 MB each, of something you own or have permission to use.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Picker | Options | Changes the price? |
| Vibe | Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life | No |
| Length | 15, 30 or 60 seconds (one, two or four segments) | Yes |
| Format | Portrait 9:16 or Landscape 16:9 | No |
| Quality | 480P, 720P or 1080P | Yes |

## FAQ

### Can a small business make a video with AI?

Yes. Upload one to four photos of your shop or product and one line on the Ad Video tab; SceneMixer makes the script, cast, shots, sound and end card for you.

### How long does it take?

A 15-second one-click ad on SceneMixer is typically ready in about ten minutes; longer ads take longer because they have more segments.

### What should I make first?

A 30-second portrait A day in the life for a place, or a 15-second Show how it works for a product.

### Can I put my business name and phone number on it?

The brand, product name, a tagline and a button go on the end card. Readable text cannot be drawn into the picture, so the number belongs in your caption or listing.

### Can I update an offer without reshooting?

Yes. Edit the end card and rebuild the final cut with the same shots.

### Can I make a video for my church or community group?

Yes, as an invitation from a photo of the building or hall. Audio such as a recorded sermon cannot be uploaded; the text of a talk can be turned into an illustrated story.

### Can I make it for LinkedIn?

Yes, as a landscape explainer from a clean screenshot or a photo; check the post composer for the shapes it takes.

### Will my staff or customers be in the video?

No. People in your photos are removed and the cast is generated.

### What does it cost?

One figure is quoted on SceneMixer before it starts and it is a ceiling; length and quality set it. Ads are never the free first project.

Updated November 3, 2026


---

# AI Video Generation: How a Clip Is Made and Why It Takes Segments

URL: https://scenemixer.com/guides/ai-video-generation/
How AI video generation works: why one generation is a short clip, how faces and rooms stay put, what carries across a join, where time goes, how failures are handled.

# AI Video Generation: How It Actually Happens

Quick answer: AI video generation is one model call that turns a text description and a few reference pictures into a short clip. Anything longer is clips joined. A call has a fixed ceiling on length, it has no memory of any other call, and it takes minutes. So every system that makes minutes of video has to supply from outside what the model lacks: who the characters are (reference pictures), what the room is (location pictures and an automatic AI 3D model of the scene), what has already happened (a running record kept as text), and what to do when a call fails. This page follows one request from the button to the file and says, at each step, what the model does and what the site does around it.

Try it: generate one segment twice on the cheapest tier without changing a word, then change one sentence and generate once more. Two identical requests can give two different clips; the third shows how much a single sentence of text matters.

## One generation is one short clip

A video model takes a description and some reference pictures and returns a clip. It does that once per call, and the tier you pick decides the longest clip a call may return. On SceneMixer that ceiling is 15 seconds on some tiers and 30 seconds on the long-take tiers, and the storyboard is written to fit under it. One such clip is a segment. A minute of finished video is therefore at least four segments on a 15-second tier and at least two on a long-take tier, each made by its own call, never by one long one.

Inside a segment the story is cut into three to five numbered shots, and the seconds of the shots add up to the segment’s length. The ceiling is a maximum, not a target: a segment with two short shots may run only a few seconds. When the shots add up to more than the limit, the silent ones are tightened first and then the segment is split at a shot boundary. Nothing is cut off the end. Price follows seconds, so an episode costs the same cut into short or long pieces; what changes is how much you pay to redo one. 15 or 30 second segments.

## What goes into one call

A call is made of three things:

- The segment’s own text. In SceneMixer, a header names the one place and the time, then numbered shots with their seconds, the people and props named with @ chips, and the lines with a voice.
- Reference pictures. One for each person in the shot, one for the place, one for each prop that matters, and a state picture for anyone in a different outfit or condition.
- The tier. It fixes the resolution, the longest clip and the price per second.
The clip before it is not shown to the model. That is why a segment must stay in one place and one timeline: it is drawn from one set of scene pictures, so a shot that jumps to another room or another day is refused before you pay and has to become its own segment. A segment stays in one place and one timeline.

## Why faces change, and what holds them still

Because each call starts from nothing, anything that is only described in words is invented again every time. Most of what looks like “AI inconsistency” is that, and each case has a different supply.

| What drifts | Why it happens | What the site supplies | What you do |
| A face between shots | No memory of any earlier call, so the face is invented again | One approved picture per character, attached to every shot that names them | Name the character with the @ chip; fix the picture once in step 2 |
| A room between shots | Same reason, for places | One picture per location, and an automatic AI 3D model of each scene so who sits where is settled before the shots are made | Keep one place per segment |
| An outfit or an injury | Nothing carries a wound from one call to the next unless it is written again | A running record kept in text, and state pictures for looks that last | Write the change in the shot where it happens |
| A voice | No call remembers how a character sounded in an earlier one | One voice per character, chosen once and used for every line | Pick or upload the voice in the cast |

That is also why a higher SceneMixer tier does not fix a wrong face: resolution changes how sharp the picture is, not who is drawn. My character’s face changed: a checklist and why a whole series stays consistent go through the fixes.

## What carries across a join

In storyboards written or regenerated since late October 2026, the story is carried from one segment to the next by a running record that SceneMixer keeps while it reads your text: who is in which scene, whether they sit, stand, lie, kneel or crouch, what is in their hands, which look they wear, wounds and mud on the body, wet or dry, unconscious or not; and for each scene the time of day, the weather, the light, whether doors and windows are open, and whether a crowd is there. Each new segment starts from that record plus its own text and pictures.

| You write, in segment 3 | Segment 4 starts with |
| @Mara takes off her coat and lays it over the back of the chair. | Mara without the coat; the coat on the chair |
| @Leo is hit across the face; blood runs from his left eyebrow. | A cut on Leo’s left eyebrow, still there |
| @Leo washes his face at the sink. | The blood washed away |
The record holds what is still true when a shot ends, not what happens during it. It is not a picture of the previous clip: no last frame is attached, and regenerating one segment never changes the next. It also invents nothing you did not write. Storyboards from before late October 2026 work the older way, each segment from its own text and pictures. What the system keeps track of from one segment to the next.

## From the button to the file: where the minutes go

- Credits are held. Pressing Generate holds the segment’s SceneMixer price; it is spent only if the clip arrives.
- Your own limit. An account runs a fixed number of generations at once: 3 on Free, 5 on the first paid plan, more above that. The rest wait, credits held, and start by themselves.
- The shared limit. The whole site can have only so many jobs in flight at a model service, so when it is busy yours waits behind other people’s. The top plans put their jobs ahead in that line.
- The model runs. This is almost all of the time. An internal measurement in August 2026 averaged about a minute and a half for a setup picture and about nine and a half minutes for a segment; the card on screen shows a typical ten to twenty minutes for video.
- Delivery. The clip is stored, the AI badge is added unless your plan switched it off, and the card turns Done.
A cheaper, lower-resolution tier is usually also a faster one, which makes it the quickest way to find out whether a shot works. Why a generation takes time.

## When a call fails

Models fail in different ways and SceneMixer treats the ways differently. If the service says it is busy or rate-limited, the call is tried a second time on another provider account for the same model, with no change to what you asked for or the tier you picked. A moderation refusal and a parameter error are not tried again: a refusal is a judgement, and asking again gets the same answer.

What happens to the credits depends on what is known. If the model never received the request, or answered that it was busy, the credits are released back to you. If the call timed out or the connection broke in flight, nobody can tell whether the model ran and billed, so the credits stay frozen, tagged Under review in your credit history, until a person settles it. A balance that looks short after a failure is usually held, not gone. When a generation fails · Provider busy, not charged.

## Steady a generation before you pay for it

- Shoot one SceneMixer segment first. Check the faces, the room and whether the line is spoken before you start the rest.
- Name who is in frame. A shot that says “a woman turns to the window” gives the model nothing to hold; @Mara turns to the window does. Who is in frame.
- Say what is there, not what is not. “An empty street” works; “no people” invites people. Say what is there.
- Fix text before you re-render. Editing a storyboard is free; a regeneration is a full charge.
- Draft cheap, finish high. Draft at the lowest tier, finish at the highest.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Street Life; Costume Style Streetwear; Visual Style Realistic HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with a description each, then the alley and the rooftop as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Stage | What happens | Typical time | What you control |
| Read | Your text becomes a summary, an episode list and a cast, places and props in words | Minutes | The text; the three styles |
| Setup pictures | Each description becomes a reference picture | About a minute and a half each | The description; regenerate one |
| Storyboard | Each episode becomes segments, and each segment becomes numbered shots | A few minutes | The text, which is free to edit |
| Segment | One model call returns one clip | Typically ten to twenty minutes | The tier, the seconds, the shot text |
| Join | Finished segments become one file per episode | After the last segment is done | Using one tier for every segment avoids a re-encode |

## FAQ

### How does AI video generation work?

A model takes a text description and some reference pictures and returns a short clip in one call. Longer videos are made by joining many clips, with SceneMixer supplying the characters, places and story so far from outside.

### Why are AI video clips so short?

Each call has a ceiling set by the model tier, and on SceneMixer it is 15 or 30 seconds here. A longer video is several segments joined, not one long generation.

### Why does a character's face change between clips?

Each call starts with no memory of the others. A face that is only described in words is invented again; a face supplied as the same reference picture, with the character named in the shot, stays the same.

### Does the model remember the previous clip?

No. SceneMixer generates a segment from its own text and pictures. A running record of who is where and what they wear carries the story across the join, and no last frame is attached.

### How long does AI video generation take?

A SceneMixer segment typically takes ten to twenty minutes, depending on the tier and the queue. You can leave the page; the work continues on the server.

### What happens if a generation fails?

If the model never received the request, the credits are released. If the outcome is unknown, such as a timeout, they stay frozen until a person settles it. Delivered failures are not charged.

### Why does the same prompt give different results?

Generation is probabilistic: two identical requests can give two different clips. That is why the cheap tier is for finding out, and why editing one sentence is worth testing.

### Will a higher tier fix a wrong shot?

Rarely. The tier sets resolution and the longest segment, not who is drawn or what happens. The shot text and the character's picture decide that.

Updated November 2, 2026


---

# AI Video Generation Options: Every Setting and When It Locks

URL: https://scenemixer.com/guides/ai-video-generation-options/
Every option in SceneMixer's video generation: frame, type, region, language, styles, segment length, tiers, voices, switches, clip fixes. When each locks, and its price.

# AI Video Generation Options: Every Setting, When You Choose It and Whether It Can Change

Quick answer: Video generation here has options at five moments, and the useful thing to know about each is when you choose it, whether you can change it later and whether it touches the price. Step 1 holds the options that lock when the script is read: type, frame, region, dialogue language and three style lists. Step 3 sets the segment length per episode. Each generation takes a tier, which is the only option that changes the price of a second. Characters, scenes and props have their own options, and a finished clip can be trimmed, repainted, redubbed or sharpened. This page is a reference: one table per moment, with the choices, the lock and the price effect.

## Chosen in step 1: they lock when the script is read

These are chosen in SceneMixer before anything is generated. Most of them shape what the read writes into the cast and episode outlines, which is why they cannot be changed afterwards in the same project; for another choice, start a new project with the same script. What can and cannot change after the read.

| Option | Choices | Changeable later? | Changes the price? |
| Type | Short Drama (episodes up to 240 seconds), TV Series (up to 300 seconds), Film (no cap) | No. The episodes were cut for it | No |
| Aspect Ratio | 16:9 Landscape (default, marked Recommended) or 9:16 Portrait | No, once you continue past step 1 | No. Video is priced by the second, not by the frame |
| Target Region | Eleven choices: ten regions plus Follow script | No. Edit descriptions and names in step 2 instead | Not priced separately |
| Dialogue language | One of 15; the lines are written in it from the start | No. Edit lines in segments, or start again | No. A segment costs the same in any language |
| Visual Style | 21 presets in three groups (live-action, 3D animation, 2D animation); members can write their own | No. It shapes every sheet and clip | Not priced separately |
| Scene Style | 74 presets in 7 groups; members can write their own | Not after the read. Edit scene descriptions in step 2 and regenerate | Not priced separately |
| Costume Style | 90 presets in 9 groups; members can write their own | Not after the read. Edit character descriptions in step 2 and regenerate | Not priced separately |
| Let AI choose | If styles are left unpicked, AI can fill the rest from the script | Yes, before you continue | No |

The preset lists are described one by one in visual style presets and scene and costume styles. The scene and costume cards switch to 3D or 2D pictures when the visual style is 3D or 2D.

## Chosen in step 3: per episode

| Option | Choices | Changeable later? | Changes the price? |
| Segment length | Standard 15s (default) or Long take 30s | Per episode. Re-parsing re-cuts the episode and discards its clips | Not per second; a long take costs twice a standard segment to redo |
| Add or delete an episode | Add an episode at the bottom of step 3; Delete Episode on its card | Yes | Adding reads only the new pages; deleting removes the episode's clips |

In SceneMixer, Long take 30s can only be rendered by the tiers whose cap is 30 seconds; a segment above its tier's cap is refused before anything is charged. The shortest a segment can be is 4 seconds. 15-second or 30-second segments.

## Chosen on every generation: the tiers

A tier is the one SceneMixer option that moves the price. Each kind of generation has its own: a video tier per segment, an image tier for sheets and covers, a text tier for reading scripts and writing storyboards.

| Kind | Where you pick it | How many tiers | What changes |
| Video | The model button in the editor, per segment; the default is also in Settings | Several, from 480P up to 2K, priced per second | Resolution, longest segment and price per second. Each segment keeps the tier it was made on; tiers can be mixed in one episode, though the join then re-encodes |
| Image | Settings, Image Model | Several; the default is Qwen-Image 3.0 at 7 credits an image | How new sheets and covers look and what each costs; existing ones stay |
| Text | Settings, Text Model | Several, each with a rate per 1,000 characters | The price of reading and of storyboards, and how carefully a storyboard is structured |

Changing a tier never regenerates anything by itself. The price of a second per tier is in the price list, and which model to pick by purpose says what each family is for. A tier a plan does not include shows Available after upgrade.

## On characters, scenes and props

| Option | Choices | Changeable later? | Changes the price? |
| Voice, per character | Match a preset from the built-in library (free, playable), write a voice description, or upload your own (a plan feature) | Yes. It applies to shots you generate afterwards | Matching is free; the shots that use it are priced as usual |
| Looks, per asset | A state image for a changed outfit or scene, within caps: up to three for a character, two for a location, one for a prop | Yes | A state image is an image, with its own small free allowance on the first project |
| Own image as a sheet | Upload your own (a plan feature) | Yes | Uploads count against your plan's storage quota |
| Reuse from another project | Pick existing, in step 2 | Yes, with a one-step restore | No, on any plan |

## Switches: music, captions, badge, interface language

| Switch | Default | Who has it | What it does |
| Music | On | Everyone | Plays and downloads the episode's score; off gives the dialogue version without music. No track picker |
| Subtitles | Off | Accounts that have bought credits or a plan at least once | Burned-in download for a clip, and for a whole episode an SRT too. Free, with a daily limit |
| AI badge | On for free accounts, off for members | Members can switch it either way | A small mark in the bottom-right corner of new clips; it does not change clips you already have |
| Interface language | Your browser's | Everyone | One of sixteen. It changes how text is shown to you, not what the cast says |

## After a clip exists: fixes and versions

SceneMixer can correct a finished segment without making a new take. Every fix is priced on its button before you press it and is refunded if it fails.

| Fix | What it changes | Price |
| Trim | Cuts the start or the end | 5 credits each |
| AI visuals | Changes one thing in the picture: clothes, a scar, the weather | 2 the tier's per-second price, per second of clip |
| Redub: lip sync | Re-times the mouth to a line | 1 credit a second |
| Redub: re-perform | Has the character perform the line again | Same as AI visuals |
| Enhance | Raises resolution one step on a clip you already have | Per second, by option; on the card |
| Regenerate | A new take of the segment | The tier's per-second price, per second |

Every new version is kept in Version History, so you can switch back. What each clip fix costs and enhancing a clip have the details.

## Other project kinds have their own options

A SceneMixer ad project picks its length (15, 30 or 60 seconds), aspect ratio and resolution in the ad composer, per ad. A remake reads an uploaded clip of up to 5 minutes, fixes its dialogue language when the clip is read and has its read priced by the minute. Both end in the same editor and the same fixes as a story project. Four ways to start a project compares them, and ad type, length, aspect and resolution lists the ad options.

## FAQ

### Which options cannot be changed after the script is read?

Type, frame, target region, dialogue language and the visual, scene and costume styles. They shape what SceneMixer’s read writes. For another choice, start a new project with the same script.

### Which option changes the price?

On SceneMixer, the tier you generate on, and the seconds you generate. Frame, type, language and style lists are not priced separately.

### Can I change the video tier for one segment?

Yes. SceneMixer lets you choose the tier per segment with the model button in the editor, and each segment keeps the tier it was made on.

### Can I mix 15-second and 30-second segments?

A SceneMixer episode is cut at one length, 15 or 30 seconds, but you can split a segment by hand afterwards.

### How many styles are there?

21 visual styles, 74 scene styles in 7 groups and 90 costume styles in 9 groups, counted from the Step 1 pickers on November 2, 2026, plus members' own text for each.

### How many dialogue languages are there?

15. The language is chosen in step 1 and the lines are written in it from the start; the interface language is a separate setting.

### Is there a setting for music?

Music is automatic in SceneMixer. A switch turns it off for playback and download, but there is no track picker or volume control.

### What changes if I join a plan?

Monthly credits, more generations at once, your own uploads and your own style text, and the badge switch. The price per second of a tier does not depend on the plan.

Updated November 2, 2026


---

# AI Video Generator Free: Seven Hidden Costs and How to Check Each

URL: https://scenemixer.com/guides/ai-video-generator-free/
Seven things a free AI video generator can quietly cost you: a mark, a queue, a ceiling, rights, expiry, credits, a card. A quick test for each, and what SceneMixer does.

# AI Video Generator Free: Seven Hidden Costs to Check Before You Start

Quick answer: A free AI video generator is rarely free of conditions. The cost moves somewhere you do not look at first: a mark burned into the picture, a queue you cannot see the end of, a ceiling on length or quality, unclear rights to what you make, files that vanish, free credits that expire or buy almost nothing, or a card requested at sign-up. This page names those seven, gives a five-minute test for each that works on any tool, and then says what SceneMixer actually does on each point, including the ones that are not flattering: the badge on free downloads, a short queue, and no free video beyond one 5-second preview.

Try it: open the pricing, terms and sign-up pages of any free generator you are considering and look for the answer to each of the seven questions below. Whatever you cannot find in five minutes is the cost you will meet later.

## 1. A mark on the picture

Ask: is there a mark, where is it, does it survive a download, and does it appear on shared copies too? Test: make the shortest clip the tool allows, download it, and play the file full screen on a phone. A mark that only shows in the web player is a different thing from one burned into the file.

Here: a free account’s videos carry a small SceneMixer AIGC badge in the bottom-right corner and a line of AI-generated-content metadata in the file. The badge is part of the picture, so it survives a download and a re-upload. It is not a full-frame watermark. On a paid plan neither is added by default and you can switch them back on. A link shared from SceneMixer always shows the badge, whatever the plan. The AI badge switch · Why shared videos carry the badge.

## 2. A queue you cannot see the end of

Ask: how many jobs run at once on the free tier, what happens to the rest, and is anything of yours locked while it waits? Test: start four jobs together and read what the fourth one says.

Here: a free SceneMixer account runs 3 generations at once. Any more show Queued, start by themselves in the order you pressed, and hold their credits until they do. That held amount is the quiet cost of a queue: your balance looks lower than it is until the job runs. A paid plan raises the number that run together (5 on Pro, 8 on Studio, 10 on Enterprise, more above that). It does not make one clip faster: a segment takes typically ten to twenty minutes on every plan. Generating many segments at once.

## 3. A ceiling on length or quality

Ask: what is the longest clip, which resolutions are on the free tier, and is the better setting locked behind a plan? Test: open the quality picker and see which entries are greyed out for your account.

Here: SceneMixer’s picker offers the same tiers on every plan, from 480P at 10 credits a second up to a 2K tier, so quality is not locked; it is a budget. The longest segment is 15 seconds on some tiers and 30 seconds on the long-take tiers, and an episode is several segments joined. What a free account really lacks is seconds, not pixels. The only free video is one preview of up to 5 seconds on the lowest tier; every other second is paid.

## 4. Rights to what you make

Ask: may you use the output commercially, is output from free credits treated the same, is it exclusive to you, and who is responsible if someone objects? Test: find the clause on your content in the terms, not in the marketing.

Here: the terms give you what you generate to the extent the law allows, commercial use included, and output made with free credits counts the same as any other. Two limits come with it: SceneMixer cannot promise it is registrable as a copyrighted work, and it is not exclusive to you. It also does not indemnify you if someone claims infringement, and the model services underneath have their own terms. That is a description of the terms, not legal advice; read them before you build a business on a clip. What you get to keep.

## 5. Files that disappear

Ask: how long do generated files and uploads stay, what does deleting do, and do share links expire? Test: find the retention line in the privacy policy and the deletion line in the help pages.

Here: what you generate stays in your SceneMixer account until you delete it, and generated sheets and videos never count against upload storage. Deleting removes an item from your account at once; the files behind it are cleaned up about seven days later. Uploaded reference material is kept 30 days by default unless it sits in an active project, according to the privacy policy. A share link has no expiry date, is public to anyone who holds it, and has a daily play limit. Because of that last point, download the original file for anything you cannot afford to lose. The seven-day window · Share links.

## 6. Credits that expire, or buy almost nothing

Ask: is the free amount one-off or monthly, does it expire, and what does it buy in seconds of video? Test: divide the free balance by the price per second on the button.

Here: a new account gets 50 credits once, and they do not expire; credits you buy and credits a plan grants do not expire either. The same sum, 50 credits divided by 10 credits a second on the default tier, shows how little video that is on its own. So the free part of SceneMixer is built from the cheap steps: 15 setup images for the whole account, the storyboard for episode 1 of your first drama, one batch of three cover posters and the one preview. Used is used: deleting an image and making it again does not give the allowance back. What you get for free · Do credits expire?.

| Step | Free here? | What it costs after the free part |
| Reading a script | No | 4 credits per 3,333 English characters |
| Setup images | 15 per account | 7 credits each on the default image tier |
| Storyboard | Episode 1 of your first drama | 1 credit per 100 counted characters of the episode |
| Preview clip | Once, up to 5 seconds | Not repeatable |
| Video | No | 10 credits a second on the default tier, more on higher tiers |

## 7. A card on file, or something that renews by itself

Ask: does sign-up ask for a card, and does anything renew without you pressing a button? Test: go through sign-up up to the point where it asks for anything other than an email address.

Here: SceneMixer sign-up takes Google, Apple or a six-digit code sent to your email, and it asks for no card. Nothing renews unless you subscribe to a plan, which starts at $7.99 a month; credit packs, from $1.49, are one-off purchases. Cancelling a plan stops the next renewal and keeps the credits already granted. There is no unlimited plan: every plan works in credits. Signing up and logging in · What a paid plan changes.

## A five-minute habit that works on any tool

Keep one line per cost in a note and fill it in before you create an account: mark, queue, ceiling, rights, files, credits, card. A tool that answers all seven in plain words is easy to judge, whatever the answers are. A tool that hides three of them has told you where its free tier stops. If you only want to see how far free goes on this site, the free AI video generator page walks that path, and free AI text to video does it from a piece of text.

| Hidden cost | The question | Quick test | On SceneMixer |
| Mark | Is it in the file? | Download and play full screen | Small badge in the corner on free downloads; always on shared links |
| Queue | How many at once? | Start four jobs | 3 at once on free; extras queue and hold credits |
| Ceiling | Longest clip, quality locked? | Open the quality picker | Same tiers on every plan; 15 or 30 seconds per segment |
| Rights | Commercial, exclusive? | Read the content clause | Yours to the extent the law allows, not exclusive, no indemnity |
| Files | How long are they kept? | Read retention and deletion | Kept until you delete; about seven days to clean up; uploads 30 days unless in a project |
| Credits | Do they expire, what do they buy? | Divide balance by price per second | 50 once, no expiry; video is paid from the first second |
| Card | Asked at sign-up? | Stop at the first card field | No card; nothing renews unless you subscribe |

## FAQ

### Is there a free AI video generator with no limits?

Not one that can be trusted to say so. Every generator pays for each second of video it makes, so the limit shows up as a mark, a queue, a cap or an allowance. SceneMixer gives 50 credits and a set of free setup steps; video is paid from the first second.

### Does SceneMixer put a watermark on free videos?

A small SceneMixer AIGC badge in the bottom-right corner, plus an AI-generated-content metadata line in the file. It is not a full-frame watermark. Paid plans are badge-free by default, and shared links always carry it.

### Can I use videos made with free credits commercially?

Under SceneMixer’s terms, yes: output made with a free allowance is treated like any other, and commercial use is covered to the extent the law allows. It is not exclusive to you and the terms give no indemnity.

### Do the free credits expire?

No. The 50 credits on a new SceneMixer account stay until spent, as do purchased and plan credits. Offers such as a first-purchase bonus can have a deadline.

### Do I need a credit card to start?

No. SceneMixer sign-up uses Google, Apple or an emailed six-digit code. A card or PayPal is only asked for when you buy credits or a plan.

### Why is video not free when images are?

Video is the one step whose cost scales with the seconds you make. Text and image steps are one call each; a segment is a call per clip, charged by the second.

### Does a paid plan make generation faster?

No. A SceneMixer plan raises how many of your generations run at once. One segment takes the same time on every plan.

### What happens to my files if I stop using the site?

SceneMixer keeps generated files in your account until you delete them. Uploaded reference material outside an active project is kept 30 days by default, per the privacy policy. Download originals of anything you need.

Updated November 2, 2026


---

# AI Video Generator Free Unlimited: Every Limit, With Its Number

URL: https://scenemixer.com/guides/ai-video-generator-free-unlimited/
Free and unlimited do not go together in video: every generator limits credits, queue, length, size, storage and clicks. Each limit on SceneMixer, with its number.

# AI Video Generator Free Unlimited: Every Limit a Generator Has, With the Numbers

Quick answer: No AI video generator is free and unlimited at once, and a page that says it is has only moved the limit somewhere you cannot see. Video is paid for by the second, so a limit always exists: on credits, on how many jobs run together, on length, on size, on how long files are kept, on how fast you may click. This page lists every limit SceneMixer has, with the real number where there is one and the place you meet it, then gives four questions that turn any “unlimited” page into a list of numbers.

## Six kinds of limit, in one register

Money, time, size, storage, frequency and content are the six places a limit can live. Here is each, with the number on SceneMixer. Numbers marked as live are read from the site's price list when this page is built.

| Kind | Limit | The number here | Where you meet it |
| Money | Credits | No unlimited plan: every plan works in credits. 50 welcome credits once; credits do not expire | The price on every button; the balance in the top bar |
| Time | Generations at once | 3 on a free account, 5 on Pro, 8 on Studio, 10 on Enterprise, more above that | Cards that say Queued; they hold their credits and start by themselves |
| Time | How long one clip takes | Minutes, not seconds, for one segment, on every plan; the estimate on screen is an average of what recently finished, not a promise | A generating card with its progress |
| Size | Length of one segment | From 4 seconds up to 15 on some tiers and 30 on others | A segment over its tier's cap is refused before anything is charged |
| Size | Segments in an episode | Up to 60 | The timeline of an episode |
| Size | Length of an episode | Short Drama up to 240 seconds, TV Series up to 300, Film no cap | Step 1, Type |
| Size | Script length | Read in one pass below 10,000 counted characters, in two passes up to 100,000; for English that is 300,000 raw characters, because 3,333 English characters count as 1,000 | The counter under the script box |
| Size | Cast per project | Up to 100 characters and 100 props | Step 2 |
| Size | Picture quality | Tiers rise to a 2K tier, each at its own price per second | The model button in the editor |
| Free quotas | What is free | 15 setup images per account, the first storyboard of the first drama, one batch of cover posters, one 5-second preview per account | A free label on the button; used is used |
| Storage | Uploads | JPG or PNG up to 20 MB; a paid-plan feature with a quota on the plan card. Only uploads count; everything generated is stored free | The asset library |
| Storage | How long things stay | Generated files until you delete them; deleted files cleaned up about seven days later; uploads 30 days by default outside an active project | Deleting, and the privacy page |
| Storage | A share link | No expiry date, with a daily play limit per link and a daily limit for the whole site | A poster instead of the player, until the next day |
| Frequency | Clicks that cost the site money | Sign-in codes, image recognition and voice matching are limited per hour; the click is refused, nothing is charged | A message that tells you to wait |
| Content | Blocked video attempts | More than 12 in a day pauses video submission for 1 hour | A banner on the project |
| Source | A clip for a remake | Up to 5 minutes | The upload box |

Two of these are the ones people mean by “unlimited”, and both are real ceilings. There is no flat price for any number of seconds, because each second is a paid model call at the provider. And there is no count of generations that is free in a way that avoids that: the only free video is the one preview. What a paid plan changes and why unlimited plans disappeared explain the arithmetic.

## What stays the same on a free account

Free changes how much you can do at once and how much you have to spend; it does not change the rules of the register. The same tiers, the same length caps and the same script limits apply, and one segment takes the same time. What a free account lacks is seconds, not features of the picture: video is charged from the first second apart from the single preview, and a free account runs 3 generations at a time. Everything above that is a decision about how many credits you want to buy. What SceneMixer gives you for free lists the free quotas and their conditions.

The credits you start with last as long as you want them to: welcome credits, bought credits and plan credits do not expire, and an offer such as a first-purchase bonus is the only kind that can carry a deadline. Do credits expire?.

## Four questions that turn “unlimited” into numbers

Unlimited what? Seconds, generations, images, downloads or sign-ins are different things. A page that says unlimited without a noun has not told you anything. At what length and quality? A cap of a few seconds per clip, or a lower resolution for free, is a limit even if the count is unlimited. How many at once? One job at a time makes a long batch a long wait whatever the count. For how long, and where? Files that are removed after a day, or only kept while you are signed in, make the count irrelevant. If a page answers all four with numbers, it is telling you the truth even when the numbers are small. AI video generator free: seven hidden costs goes through the same test from the other side, and the free AI video generator page shows where free ends here.

## FAQ

### Is there a free AI video generator with unlimited use?

Not here, and not anywhere that charges by the second for its own costs. SceneMixer is free to start with 50 credits, free setup images and a free preview; video is paid from the first second.

### How many videos can I make for free?

SceneMixer gives one free preview of up to 5 seconds per account. After that a video is made from credits: the welcome credits cover one short segment, and more takes more credits.

### Do the free credits expire?

No. SceneMixer keeps welcome, bought and plan credits in your balance until you spend them.

### How many generations can run at once?

SceneMixer runs 3 on a free account, 5 on Pro, 8 on Studio and 10 on Enterprise, with more on the higher plans. Extra ones queue and start by themselves.

### What is the longest video I can make?

A SceneMixer segment is up to 30 seconds on the 30-second tiers, an episode holds up to 60 segments, and episode length is capped by type: 240 seconds for Short Drama and 300 for TV Series. Film has no cap.

### Is there a limit on how often I can click?

SceneMixer limits a few helper actions per hour, such as image recognition and voice matching. A refused click costs nothing. Generations are never refused for frequency; they queue.

### How long are my files kept?

SceneMixer keeps generated files until you delete them. Deleted files are cleaned up about seven days later. Uploaded reference material is kept 30 days by default unless it is in an active project.

### Why does a paid plan not make it unlimited?

Because each second of video costs SceneMixer a model call. A plan adds credits every month and lets more of your jobs run at once; it does not remove the per-second price.

Updated November 2, 2026


---

# AI Video Generator Software: Cloud or Local, and Where Your Files Go

URL: https://scenemixer.com/guides/ai-video-generator-software/
Cloud or local AI video generator software: what your computer needs, what runs in your browser, what is stored on servers, who can open a file, how to get your work out.

# AI Video Generator Software: Cloud or Local, and Where Your Files Go

Quick answer: AI video generator software comes in three kinds, and the difference that matters is where the model runs: on your own computer, on a server behind an app you install, or on a server behind a website. SceneMixer is the third kind. Nothing is installed and no graphics card is needed, because your browser only pastes, converts and uploads, plays a lighter copy, and downloads the finished file. The reading, the pictures, the video and the joining all happen on servers. The price of that convenience is that your material is stored there, so this page says what is stored, who can open a file, what counts against your storage, what deleting does and how to get your work out.

Try it: open Assets, pick a finished clip and press Download, then play the file from your own disk. That is the full-quality copy; the one the site plays in the page is a lighter one, and the gap between the two is the first thing to know about cloud video.

## Three places AI video software can live

| Kind | Where the model runs | You provide | Where your material goes |
| Runs on your own computer | On your own graphics card | The hardware, the setup, the model files and the updates | Nowhere, unless you upload it |
| An app you install that calls a cloud model | On a remote server | The install and its updates, plus a connection | Sent to the service for each generation |
| A website you open in a browser | On remote servers | A browser and a connection | Uploaded to, and stored by, the service |

None of these is better in general. Software on your own machine trades money and setup time for control of the files. A website trades control for having nothing to set up and for using models that no single computer could run. SceneMixer is the third row, so the useful questions are the ones below: what runs in your browser, what is stored, and how you leave.

## What runs where in SceneMixer

In your browser: pasting text; unpacking a Word file to plain text before anything is sent; converting a picture to JPEG before it goes up; sending an upload straight to storage; playing a lighter copy of a clip; downloading the full-quality file.

On servers: reading the script into a cast and an episode list; making the setup pictures; writing the storyboard; generating every video segment on a model service; adding the AI badge; and joining segments into an episode. Because of that split a phone can do everything a desktop can, and closing the laptop does not stop a generation. What runs underneath.

## What your computer needs

| Need | What it takes |
| Graphics card | None. Nothing is rendered on your machine |
| Install | None. A web browser on a computer or a phone |
| Connection | Enough to upload what you choose to upload, and to download finished files. A generation itself does not need you connected |
| Disk space | Only for the files you download. The full-quality download is a separate, larger file than the lighter copy the player streams |
| Image formats | Whatever your browser can decode: WebP, HEIC, PNG or JPEG. SVG, camera RAW and HEIC outside Safari are refused when you pick them |
| Script formats | Pasted text, or TXT, DOCX and Markdown. PDFs are refused: export to text or Word first |

## Where your files live

Uploads go straight to storage. Your browser asks for a short-lived permission, good for fifteen minutes and for one address only, then sends the file directly to object storage. It never passes through the application servers, and when the upload finishes the site checks the bytes that actually arrived rather than trusting what the browser said it was sending.

An address is unguessable, not private. Every picture and clip is shown through a long signed address with an expiry. The addresses are not published or listed, but whoever holds a copied one can open the file while it is valid. To show someone a finished clip, use Share, which makes a proper page; a shared copy always carries the badge.

References can travel. To generate from a reference, the site publishes a copy of it where the model service can read it, and that copy may sit in storage in the region the service runs in, not necessarily where your account’s data is. If material must not leave a particular jurisdiction, weigh that before uploading it.

Only uploads count against your storage. Every generated picture, take and episode is stored at the site’s expense. The quota is reserved when an upload is authorised and given back when you delete.

Deleting is final. A deleted asset, episode or project disappears from your account at once and its files are cleaned up afterwards, after a check that nothing else still uses them. There is no restore on the site. Where your files live has the full account.

## Getting your work out

- Video. Every SceneMixer segment and every joined episode has a Download button that gives the full-quality file. A free account’s files carry the small AI badge; on a paid plan it is off by default.
- Pictures. Characters, scenes and items can be downloaded from the Assets page, across all your projects.
- Text. Your script is yours already. Storyboard text can be copied out and pasted back in; check the links after you paste.
- A project as a whole. A project cannot be moved to another account. Hand over the script, the character pictures and the downloaded videos instead.
What you get to keep covers the terms: output is yours to use, commercial use included, to the extent the law allows. It is not exclusive, and keeping your own copy of anything important is sensible given that deleting here is one-way.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Nordic Minimal; Costume Style Outdoor Gear; Visual Style Painterly Anime; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the keeper and the stranger as characters, the lighthouse and the shore as locations; the lantern as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Question to ask any AI video software | Why it matters | SceneMixer |
| Where does the model run? | Decides what hardware you need | On servers; nothing to install, no graphics card |
| What leaves my computer? | Decides what you are comfortable uploading | Only what you choose to upload, plus the text you paste |
| Who can open a file? | Preview addresses are often mistaken for private | Unguessable addresses with an expiry; share page for showing others |
| What counts against storage? | Quotas can be spent on your own output | Only your uploads; generated files never do |
| Can I leave while it works? | Decides whether you must stay connected | Yes; generations continue on the server |
| How do I get the file out? | The exit matters more than the entrance | Full-quality download per segment and episode; no project export between accounts |

## FAQ

### Is AI video generator software cloud or local?

Both exist. Local software runs the model on your own graphics card; cloud software runs it on servers. SceneMixer is cloud: you use it in a browser and the models run elsewhere.

### Do I need a powerful computer?

Not for SceneMixer. Nothing is rendered on your machine, so a phone or a modest laptop is enough.

### Do I need to download or install software?

No. You open SceneMixer in a web browser. You only download the videos and pictures you want to keep.

### Can I use it offline?

No. SceneMixer needs a connection to upload, to sign in and to download. A generation, once started, continues on the server even if you disconnect.

### Where are my files stored?

SceneMixer keeps them in object storage, reached by long, unguessable addresses. Uploads go straight from your browser to storage. Only what you upload counts against your quota.

### Is a preview link private?

It is unguessable rather than private: anyone holding the address can open the file until it expires. Use Share to show a clip to other people.

### Can I move a project to another account?

No. A SceneMixer project cannot be moved; hand over the script, the character pictures and the downloaded videos.

### What happens to my files if I delete a project?

They vanish from your SceneMixer account at once and are cleaned up afterwards once nothing else uses them. There is no restore.

Updated November 2, 2026


---

# AI Video Generator Tool: What You Open and Every Screen on the Way

URL: https://scenemixer.com/guides/ai-video-generator-tool/
An AI video generator tool you open in the browser: what you need, the nine screens of one full use, where the price shows, and what keeps running when you close the tab.

# AI Video Generator Tool: Every Screen From Idea to File

Quick answer: SceneMixer is a tool you open, not one you install. It runs in a web browser, on a phone as well as a computer, and the heavy work happens on servers, so closing the tab or the laptop does not stop a generation. One complete use passes through nine screens in a fixed order, and two choices are locked on the way: the visual style and the aspect ratio. This page walks the screens in order, says what you do on each and what it costs, then lists the side screens (assets, credits, settings, help) so that nothing on the way is a surprise.

Try it: type a few lines into the script box on the home page and watch the price on the button follow the length. If a sign-in window appears when you press the button, what you typed stays in the box.

## What you need to open it

An account, with no password. Press Sign In and choose Continue with Google, Continue with Apple, or Continue with email: type the address, press Send code, type the six digits. A new address gets an account on the spot. Pressing any button that needs an account opens the same window over the page. Signing up and logging in.

A browser, and nothing to install. A few jobs are done by the browser before anything is sent. A Word file is unpacked to plain text in your browser. A picture in WebP, HEIC, PNG or JPEG is converted to JPEG in your browser. Files the browser itself cannot open, such as SVG, camera RAW or HEIC outside Safari, are refused the moment you pick them, not three steps later.

A phone works. It is the same SceneMixer account and the same projects. The episode editor stacks into three views (preview, timeline, storyboard text), and the model picker and the AI badge switch sit in a bottom sheet. Pasting a long script and editing storyboard text at length are easier with a keyboard. Using the site on a phone.

## The nine screens, in the order you meet them

| # | Screen | What you do there | Price, or what locks |
| 1 | Home | Pick one of four tabs (Upload Script, AI Create, Ad Video, Remake a video) and paste, write or upload | The price of the first step is on the button and follows what you type |
| 2 | Step 1: Set Up Your Project | Name it; pick Visual Style, Scene Style and Costume Style (or let AI choose); 16:9 or 9:16; target region and dialogue language | Visual style and aspect ratio are locked once you continue |
| 3 | Reading the script | Watch the stages: finding characters, scenes and props, splitting episodes, shaping the visual style, preparing episodes. Leave if you like | Charged by length, refunded if the read fails |
| 4 | Script Parsed | Read the summary: title, protagonist, era, how it was adapted, the episode list and the counts | If the split is wrong, fix the text and read again |
| 5 | Step 2: the cast | Characters, scenes and items, each with a description and a picture; edit a description, replace a picture | The first 15 pictures on an account are free |
| 6 | Step 3: the episodes | One card per episode; Parse Storyboard; choose Segment length 15s or Long take 30s | Episode 1 of your first drama is free; the length applies to the next parse |
| 7 | The editor | Three columns over a timeline: assets on the left, storyboard in the middle, preview on the right. Read it as Shots or as Text; edit; use AI edit | Editing text is always free |
| 8 | Generate | Pick a tier in the top bar and press the black button with the price on it, or Generate all | 10 credits a second on the default tier, per segment, held then spent |
| 9 | Join, download, share | Join the segments into one episode; Download the full-quality file; Share a link | Joining is 1 credit a second; a shared link always carries the badge |

The phrases in italics are the button and tab names as they appear on the site. Your first five minutes is the home page in detail, and the episode editor, a tour goes through screen 7.

## Three places where something locks or costs money

Step 1, when you press Next. SceneMixer’s visual style shapes every picture and video, so it cannot be changed after you continue; the aspect ratio is fixed too. If you are unsure, open a sample project in the style you are considering first: each sample is labelled with its style, so you can see what it does before you commit. Choosing visual style settings.

Parse Storyboard. Billed by the length of that episode’s text, and the same price the second time. Regenerating a storyboard also discards the finished video in that episode, so edit the text instead. Regenerating a storyboard: what you lose.

Generate all. The dialog states the plan before you commit: how many segments, their total seconds, the model, which segments already have a clip and stay as they are, and any that are longer than the model allows. Each segment is charged on its own and refunded if it fails. Generate all.

## The side screens

- Assets (sidebar): everything you made or uploaded, across projects, in six tabs: Characters, Scenes, Items, Videos, Novels, Other. Everything can be downloaded. Uploading an image there is a paid-plan feature and counts against your storage; generated files never do. The asset library.
- Credits (click your SceneMixer balance): Available, Frozen, Credit History and Buy Credits. Each task is one line in the history.
- Projects: cards that say Step 1/3, 2/3 or 3/3. Click one to resume where you stopped.
- Settings (the gear): including Model Preferences for the text and image tiers, and the replay of the quick tour.
- Ask (the question mark): answers questions about the site from its guides; feedback is sent from inside it.

## What keeps running without you

SceneMixer runs generations on the server. Close the tab, close the laptop, or start something on the phone and check on the desktop: the clip lands in the account either way. While anything in an episode is generating, the browser tab’s title gains a prefix such as (3/12) generating, and on a desktop browser the site can ask once for permission to send a notification when a segment finishes. Segments beyond your plan’s simultaneous limit show as Queued, hold their credits, and start by themselves. Progress bar stuck, page closed: does generation continue?.

## What this tool is not

- Not an editing suite. The editor changes one clip at a time: trim, change a detail of the picture, redub the talking, enhance. Titles and captions are added in your own editing software after you download.
- Not a team workspace. A project belongs to one account. Reviewers watch through share links without an account; they cannot edit. Using SceneMixer as a team.
- Not metered in secret. Every button that spends credits shows its price first, and editing text and switching between takes you already have are free.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Corporate Cold; Costume Style Power Tailoring; Visual Style Filmic Live-Action; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the two rivals with a description each, then the boardroom and the corridor as locations; the signed contract as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Where | Locked after you continue? | Why |
| Visual style (step 1) | Yes | It shapes every sheet and every video |
| Aspect ratio (step 1) | Yes | Every segment and the join are cut to one shape |
| Names, descriptions, pictures (step 2) | No | Editing them is free, regenerating a picture is priced |
| Storyboard text (editor) | No | Editing is free; regenerating the whole storyboard discards finished video |
| Tier (editor top bar) | No | Chosen per segment, so you can draft cheap and finish high |

## FAQ

### Do I need to install anything to use the AI video generator?

No. SceneMixer runs in a web browser, on a computer or a phone. The generation itself happens on servers.

### Do I need a password?

No. You sign in to SceneMixer with Google, with Apple, or with a six-digit code sent to your email; a new address gets an account automatically.

### Does it work on a phone?

Yes, with the same account and projects. The editor stacks into preview, timeline and text views; long scripts and long edits are easier on a desktop.

### What if I close the browser while a video is generating?

Nothing is lost. SceneMixer runs generations on the server, and the finished clip is in your account when you come back.

### Where do I see the price before I pay?

On the button. Each SceneMixer button that spends credits shows its price, and the price on the button is the price charged.

### Which choices can I not change later?

The visual style and the aspect ratio, once you continue past SceneMixer’s step 1. Everything else, including the cast and the storyboard text, can be edited.

### Can my team work in the same project?

No. A SceneMixer project belongs to one account. Share links let reviewers watch without signing in.

### Can I add titles and captions inside the tool?

Add them in your own editing software after downloading. The built-in SceneMixer editor trims, changes picture details, redubs and enhances one clip at a time.

Updated November 2, 2026


---

# AI Video Generator Tools: What Is Inside One, and What Each Costs

URL: https://scenemixer.com/guides/ai-video-generator-tools/
The tools inside one AI video generator: script reader, cast sheets, voices, storyboard, shooting, four clip fixes, music, subtitles, cover art, and what each costs.

# AI Video Generator Tools: What Is Inside One Generator

Quick answer: People search for AI video generator “tools” as if they were separate apps. In practice a generator is a bench of tools that hand work to each other, and the useful question is which tool sits at which step, what it takes in, what it gives back, and whether pressing it costs credits. SceneMixer has about twenty of them, grouped here by the stage where you reach for them: reading, casting, planning, shooting, repairing and finishing. Most of the text tools are free to use once the script has been read; the ones that cost are the ones that call an image or video model.

Try it: open any sample project, then find three things in it: the voice line under a character, the shot list of one segment, and the Edit this clip panel under a finished clip. That is a quarter of the bench.

## Reading tools: from words to a project

Script writer (AI Create). You give one or two sentences carrying genre, lead, conflict, ending and tone; SceneMixer drafts a script. It is charged by length, 4 credits per 1,000 counted characters at the default text tier, and every draft is kept in Writing History. Writing the one-line brief.

Script reader (the read). Paste or upload .txt, .docx or .md and the reader turns it into characters, locations, props and an episode split. It is charged once per project at 4 credits per 3,333 English characters, and after it everything it wrote is editable for free. Long scripts are read in two passes. Long scripts are read in two passes.

Remake reader. Upload one MP4 or MOV and it is read shot by shot into a cast, sets and a storyboard. It is priced by the clip’s length in minutes, rounded up, and the source is never handed to the video model. How to remake a video with AI.

## Casting tools: who and where things look like

Character, scene and prop sheets. In SceneMixer, one image per entry, made from its description, then used as the reference for every shot that includes it. The first 15 on an account are free; after that each is 7 credits on the default image tier. You can regenerate, edit the description, or replace the sheet with your own image on a paid plan. The same face in every shot.

State images. A second reference for a thing that changes shape: a full change of outfit, a burnt hall, a drawn blade. The script reader proposes them within caps (up to three looks for a character, two for a location, one for a prop) and works out which segment uses which. Creating or editing states by hand is a member feature. Costume changes and scene states.

Voice matcher. Each character starts with a voice written in words. A free pass can match it to a preset from the built-in library, which you can play before choosing, and an uploaded voice is a paid-plan feature. How character voices work.

## Planning tools: the storyboard and its editors

Storyboard generator. SceneMixer turns each episode into numbered shots with seconds, charged per episode at 1 credit per 100 counted characters; episode 1 of your first drama is free. Shots / Text editor. Rewrite a shot by hand: free. AI edit. Describe a change in plain words and the storyboard text is revised, 1 credit per revision that comes back. Text history. Restores any earlier storyboard text, free. AI edit the storyboard text · Text history.

## Shooting tools: from text to a clip

Segment generator. In SceneMixer, one press makes one segment, picture and sound together, at the tier in the top bar; the price is on the button before you press it. Default tier: 10 credits a second. Generate all submits every segment that has no clip yet in timeline order. Free preview makes one shot of up to 5 seconds, once per account. Join (Composite Episode) stitches finished segments into the episode at 1 credit a second. Generate all segments at once · Composite and share an episode.

## Repair tools: fix a clip without re-shooting it

Under every finished clip sits Edit this clip with four tabs. Each fix is priced on its button and refunded if it fails.

| Tool | What it changes | What it costs |
| Trim | Cuts the start or the end | 5 credits per saved trim |
| AI visuals | One change from a fixed menu of eight: clothes colour, time of day, weather, background people, lights, a scar, an expression, or swapping one character | 2× the per-second price of the clip’s tier, for the clip’s length |
| Redub | New lines or a new voice. Lip sync keeps the picture, re-voices every line and moves the mouths to match (the take’s own ambience is dropped). Re-perform plays the clip again with the new lines, so the picture may shift a little; clips of 15 seconds or less | Lip sync 1 credit a second; Re-perform priced like AI visuals |
| Enhance | Raises the clip one resolution step (480P to 720P to 1080P to 2K) | Per second, by option and target; shown on the card |

None of these SceneMixer repairs changes framing, camera or where people stand; those are a storyboard edit and a regeneration. What each clip fix costs · What AI visuals can change.

## Finishing tools: music, words, cover and sharing

Score. SceneMixer scores each episode for you from its storyboard and mixes it under the voices; the Music switch by the player turns it off. There is no track picker. Included in the video price. The built-in score.

Subtitles. A switch beside Music shows each line as it is spoken and offers a burned-in download and an SRT file for a whole episode. No credits, available once an account has bought credits or a plan. Add subtitles.

Cover art. Three poster options from your leads’ sheets, priced as images, with one free batch on your first drama. Make a promo cover poster. Share page and download. A share link is free and public; a download is the full-quality file. Original versus play version.

## What is not on the bench

There is no stand-alone text-to-speech, face-swap, upscaler or timeline editor to open on their own. Voices, lips and ambience come with each clip; Enhance works only on clips made on the site; and cutting a finished episode against your own footage, adding titles or colour grading is a job for an editor you already use after you download. If you need one of those, the honest route is to export the file and finish elsewhere.

| Stage | Tools | Cost to use |
| Reading | Script writer, script reader, remake reader | Charged by length; edits afterwards are free |
| Casting | Sheets, states, voice matcher | 15 free sheets per account; then per image |
| Planning | Storyboard, Shots/Text editor, AI edit, text history | Per episode; hand edits free |
| Shooting | Segment, Generate all, preview, join | Per second; join 1 credit a second |
| Repair | Trim, AI visuals, Redub, Enhance | On the button; refunded if the fix fails |
| Finishing | Score, subtitles, cover art, share, download | Score and subtitles included; cover art priced as images |

## FAQ

### What tools does an AI video generator need?

At minimum something to read your text into a cast and shots, something to draw reference images, something to turn shots into clips with sound, and a way to join the clips. SceneMixer adds voices, a score, four clip fixes, subtitles and cover art.

### Which tools are free here?

SceneMixer does these for free: editing text by hand, text history, matching a voice, switching between takes you already have, the share page, and the score and subtitles once you qualify. Anything that calls an image or video model is charged.

### Can I fix one shot without remaking the episode?

Yes. Edit the shot text and regenerate that one segment, or use SceneMixer’s Edit this clip for a trim, one of eight picture changes, a re-voice or an enhance.

### Is there a tool to change the camera angle afterwards?

No. SceneMixer changes framing, camera and positions only by editing the storyboard text and regenerating the segment.

### Is there a separate voice or text-to-speech tool?

No. SceneMixer speaks dialogue inside each clip. You can change a character’s voice and re-voice a finished clip, but there is no stand-alone narration tool.

### Can I pick the music?

No. SceneMixer chooses the score for each episode, and it can be switched off; you can then lay your own track in an editor after downloading.

### Do I need to install anything?

No. All of these SceneMixer tools run in the browser; the heavy work happens on servers.

### What do repairs cost compared with regenerating?

A picture fix costs 2× a regeneration of the same length, a re-voice costs a small fraction, a trim is a flat fee. Compare the two prices on screen before you choose.

Updated November 2, 2026


---

# How to Choose an AI Video Generator: Seven Questions to Ask

URL: https://scenemixer.com/guides/ai-video-generators/
Seven questions that separate AI video generators from each other, a ten-minute test for each, and the red flags to look for. No rankings, just what to check.

# How to Choose an AI Video Generator

Quick answer: AI video generators differ less in how good one clip looks than in what happens around the clip. Seven questions sort them out: how long a finished video can be, whether characters stay the same, whether sound is generated with the picture, which formats and quality you get, whether you can fix one shot, what you pay and when you learn the price, and what you can do with the file. Each has a ten-minute test below. SceneMixer’s answer to each is stated plainly, so you can hold any tool, ours included, to the same seven questions.

Try it: write the same three-shot scene, two named characters in one room, into every generator you are considering. Compare the second and third shots against the first, not the first against the others.

## 1. How long can a finished video be?

Ask for the longest finished video, not the longest clip. Most models make a clip of a few seconds to half a minute. A longer video is clips joined, so the real question is what holds the joined clips together (question 2). SceneMixer makes segments of 15 or 30 seconds and joins them into an episode; a series is as long as the script. 15 or 30 second segments.

## 2. Do the characters stay the same?

This is the question that separates tools most. Test it: generate three shots with the same two people and check the faces, the clothes and the room in shots 2 and 3. SceneMixer reads the script once into a cast, locations and props that are fixed before any video is made, and every shot uses them. Why a whole series stays consistent.

## 3. Is the sound made with the picture?

Some tools make silent clips and leave voices to a second tool. Check whether the lines you write are spoken in the video, by which voice, and in which languages. In SceneMixer dialogue is spoken inside the video itself, one voice per character, in any of 15 languages, and each episode is scored for you. Dialogue and sound in one generation.

## 4. What formats and quality do you get?

Look for vertical and landscape, and the actual resolution of the file you download, not the best resolution on the pricing page. SceneMixer makes 9:16 or 16:9, from 480P up to 1080P, with a 2K tier on some models. Which quality tier.

## 5. Can you fix one shot?

Every tool makes a bad shot sometimes. What matters is the cost of fixing it: does it mean generating the whole video again? In SceneMixer you edit the shot’s text, which is free, and regenerate that one segment; the earlier take is kept. Fixing one shot without redoing everything.

## 6. What do you pay, and when do you find out?

Check the unit (per second, per clip, per credit), whether the price is shown before you press the button, and what a failed attempt costs. SceneMixer charges per second, shows the price on the button, and refunds a segment that fails. On the default tier video is 10 credits a second, and a 60-second episode is about 660 credits before retakes. Where credits go in one episode.

## 7. What can you do with the file?

Check whether you can download the original-quality file, whether a badge or watermark is added and who can remove it, and what a shared link carries. In SceneMixer a free account’s videos carry a small badge, members can turn it off, and a shared link always carries it. AI video without a watermark.

## A ten-minute test for any tool

- Write a three-shot scene with two named characters in one room, and one line of dialogue.
- Generate it. Open shots 2 and 3 and compare them with shot 1: same faces, same clothes, same room?
- Is the line spoken in the video, in the voice and language you wanted?
- Change one line in shot 2 and make only that shot again. Did the others stay as they were, and what did it cost?
- Download the file. What size, what shape, what mark in the corner?

## Red flags

- A price you only learn after you press the button.
- “Unlimited” with no mention of what limits the speed or the quality. Why unlimited AI video plans disappeared.
- Samples that are all single clips of a few seconds, with no finished episode.
- No way to see the cast or the script before the video is made.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with one mark each and a voice; the stairwell and the balcony as locations; the letter as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Question | Why it matters | SceneMixer |
| 1. Length of a finished video | Clips are short; a long video is clips joined | Segments of 15 or 30 seconds joined into episodes; a series is as long as the script |
| 2. Same characters | The hardest thing to get right | One cast, locations and props fixed before video is made |
| 3. Sound with the picture | Silent clips need a second tool | Spoken dialogue inside the video, 15 languages, scored music |
| 4. Formats and quality | What you can post where | 9:16 or 16:9; 480P to 1080P, 2K on some models |
| 5. Fixing one shot | The cost of every miss | Edit the text (free), regenerate one segment |
| 6. Price and when you learn it | Surprise bills | Per second, shown on the button, refunded on failure |
| 7. The file | What you can do with it | Original-quality download; badge off for members; shared links carry it |

## FAQ

### How do I choose an AI video generator?

Ask seven questions: length of a finished video, whether characters stay the same, sound, formats and quality, fixing one shot, price and when you learn it, and what you can do with the file.

### Which AI video generator is best?

That depends on what you are making. This page does not rank tools; it gives you the questions and a ten-minute test to compare any of them, ours included.

### How do I test whether characters stay the same?

Generate three shots with the same two named characters in one room and compare shots 2 and 3 against shot 1: faces, clothes and room.

### Why does the price unit matter?

Per second, per clip and per credit are not comparable until you know what a finished minute costs, including failed attempts and retakes.

### Do AI video generators add a watermark?

Many add a mark on free plans. In SceneMixer a small AI badge is on by default for free accounts, members can turn it off and shared links always carry it.

### Can I fix one shot without regenerating the whole video?

In SceneMixer, yes: edit the shot's text and regenerate that one segment. Check this in any tool you try.

### What is a good first project to test with?

A one-minute scene with two named characters, one room and one line of dialogue. It exposes consistency, sound and price in ten minutes.

### Does SceneMixer pass all seven?

The table above says what it does on each. The ten-minute test is the way to check that for yourself.

Updated November 3, 2026


---

# AI Video Maker Free: What a Free Account Lets You Take Away

URL: https://scenemixer.com/guides/ai-video-maker-free/
What a free AI video maker lets you keep, file by file: the badge on it, who can switch it off, and how the free preview differs from a finished clip.

# AI Video Maker Free: What You Can Take Away Without Paying

Quick answer: A free AI video maker is worth what you can take out of it. On SceneMixer a new account can leave with four kinds of thing without paying: setup images, the storyboard text of the first episode, one batch of three cover posters, and a 5-second preview you can watch and share but not download. A downloadable video file comes only from spending the 50 welcome credits on a short segment, and it carries the small SceneMixer AI badge. This page goes through each export, what is on it, and what buying credits does and does not change.

## The export inventory, file by file

Most free pages describe an allowance. This one describes the thing you walk away with.

| What | Free? | How it leaves the site | Mark on it |
| Character, location and prop images | 15 per account | Image files, from the Assets page: every entry has Download | None. The badge is for video, and for covers |
| Storyboard of episode 1 | First drama only | Text you read, edit and copy; it is not a video | None |
| Cover posters | One batch of three, on the first drama | Image files in portrait 9:16 and landscape 16:9; Download offers this size or all sizes | The badge, on a free account |
| The preview clip | Once per account, up to 5 seconds | Plays in the editor and can be shared; it has no download button | Shared copies carry the badge |
| A short segment you pay for | Paid from the welcome credits | A full-quality video file, from Download under the player | The badge, in the picture |

The allowances are counted per account, not per project, and a used one stays used: deleting an image and making it again does not give it back. What SceneMixer gives you for free has the conditions.

## The one video file the welcome credits can buy

A video is the one step that is not free, and the sum is short enough to do by hand. A new SceneMixer account starts with 50 credits. Reading a short script is charged at 4 credits for every 3,333 English characters, so a short script costs one unit. Once the read is paid, the sheets and the first storyboard come out of the free allowance. That leaves the video: the shortest segment the editor allows is 4 seconds, and the default tier charges 10 credits a second. So the clip costs four times the per-second price. Subtract the read and that clip from 50, and what remains is the whole of your margin: a second segment does not fit.

To make it that short, open the segment in the editor and bring its shots down until the total reads 4 seconds; the price tag on the card updates before anything is charged. Give a spoken shot enough time to say its line. Do this on the cheapest tier, which delivers at 480P; the price on the button is the price you pay. How to change a segment's duration.

The free preview is a separate thing and does not touch your credits. It is one shot of your own episode, preferring a shot with dialogue, made once per account. Use it first to see your cast move, then spend the credits on the shot you want to keep. The free preview.

## What a free file looks like

The badge. It is a small rounded outline with the words SceneMixer AI in white, in the bottom-right corner. It is drawn into the pixels at delivery, so it stays through a download, a re-upload and re-encoding; it is not an overlay the player adds. Its height is about an eighth of the frame's shorter side, which makes it roughly two fifths of the width on a vertical frame and under a quarter of the width on a landscape one. A line of AI-generated-content metadata is written into the file in the same pass.

Two files per clip. The player on the site streams a lighter copy at the same resolution. Download gives the original, which is the file you want to keep and the one the episode is joined from. Both carry the same badge. Download: the original versus the play version.

Where it lives. Everything a free account generates is stored at the site's expense: generated images and videos never count against your storage, and only your own uploads do (and uploads are a paid-plan feature). Deleting is final: the item leaves your account at once and the files are cleaned up about a week later. Download the original of anything you cannot afford to lose. Where your files live.

## The free preview versus a finished clip

SceneMixer’s preview and a paid segment are both real generations, but they are different things. Do not judge your whole episode by the preview, and do not expect to keep it.

| | Free preview | A segment you pay for |
| Length | One shot, up to 5 seconds | The seconds you set, from 4 up to the tier's cap |
| Which shot | Chosen for you: a shot with dialogue when the episode has one, otherwise the opening shot | Any segment of the storyboard |
| Tier | The lowest-resolution tier | Your choice, with the price on the button |
| Download | No download button; it plays in the editor and can be shared | Download under the player gives the full-quality original |
| How often | Once per account, not once per project | As often as you have credits |

The preview is the cheap way to see your own cast move before you commit a credit; the finished clip is what you take away. The free preview explains how the shot is chosen.

## What paying changes, and what it does not

Buying SceneMixer credits and joining a plan are different things, and only one of them touches the badge.

| You do this | More seconds? | Badge on new clips? |
| Buy a credit pack (from $1.49) | Yes, at the same price per second | Still on. A pack is not a membership |
| Join a plan (from $7.99 a month) | Yes, from the monthly allowance | Off by default for clips generated afterwards; you can switch it back on |

The badge decision is made when a clip is generated, from whether your account has an active membership at that moment. The switch sits beside the model picker in the editor and in the account settings, and it only affects clips generated after you change it. Clips you already have keep what they were made with, and the original and the lighter play copy of one clip always match; to get a badge-free version of an old clip on a plan, regenerate it, which is a new charge. A share link always shows the badge, whatever your plan, and so does a shared copy of the preview. Uploading your own images and voice samples is also a plan feature, and once you have bought credits or a plan at least once the Subtitles switch appears next to Music in the editor, as the subtitles guide describes. The AI badge switch and AI video without a watermark go through who gets a clean file.

## A short list to run before you call a maker free

Ask the same five questions of any free video maker. What file do I actually get? Is there a mark on it, and where? How many seconds can I make before it asks for money? Where is the file kept, and for how long? Can I share it without anything being added? A page that answers all five in plain words is telling you what free means, even when the answers are small. If you want the hidden-cost angle on the same question, AI video generator free: seven hidden costs takes it from the other side.

## FAQ

### Can I download a video for free?

Only by spending your SceneMixer welcome credits on a short segment: its Download gives a full-quality file with the badge. The free preview plays and can be shared, but has no download button.

### How many seconds of video can the welcome credits make?

One short segment. SceneMixer’s shortest segment is 4 seconds, which costs four times the per-second price of the tier, plus the small charge for reading your script. A second segment does not fit.

### Is the free video watermarked?

A free account's videos carry a small SceneMixer AI badge in the bottom-right corner, drawn into the picture, plus a line of AI-generated-content metadata in the file. It is not a full-frame watermark.

### Does buying credits remove the badge?

No. SceneMixer switches the badge off for an active membership, not for buying credits. Members' new clips are badge-free by default.

### Is the free preview the same as a finished clip?

No. The SceneMixer preview is one shot of up to 5 seconds, picked for you, on the lowest-resolution tier, once per account. A finished clip is the seconds and tier you choose.

### Do free images and storyboards carry a badge?

SceneMixer setup images and storyboard text do not. Cover posters do on a free account, and so does every shared copy of a video.

### Do I need a card to start?

No. SceneMixer sign-up uses Google, Apple or an emailed code and asks for no card.

### Can I use a free video commercially?

The terms treat output made with a free allowance the same as any other output; the exact wording, with its conditions, is in section 06. This page does not give legal advice.

### Do free exports take up my storage?

No. Only files you upload count against storage. Everything SceneMixer generates for you is stored at its expense.

### Can I get a longer video for free?

No. SceneMixer charges video from the first second apart from the one preview. A longer clip is more seconds at the per-second price.

Updated November 2, 2026


---

# AI Video Production Software: From Script to Finished Episode

URL: https://scenemixer.com/guides/ai-video-production-software/
What a video production needs, stage by stage, and how SceneMixer covers each: script, cast, locations, storyboard, shoot, join. With two worked budgets.

# AI Video Production Software

Quick answer: A production, even a small one, goes through the same stages: develop the script, cast the roles, find the locations, board the shots, shoot, assemble. SceneMixer does each of those in one project and shows what each costs: the script is read into a cast, locations and props; a storyboard is written for each episode; each segment is shot on the tier you choose; and the segments are joined into an episode. It replaces the crew and the camera, not the writer and the director, which is you.

Try it: take a page of a script you already have and read it into a project. The cast, the locations and the first storyboard show you the whole pipeline before you spend on video.

## The stages of a production, and where each one happens

| Stage | On a set | In SceneMixer | What it costs |
| Develop | A script, breakdown, schedule | Paste a script, or write one on AI Create; the read breaks it into episodes | 4 credits per 1,000 characters to read |
| Cast | Casting, wardrobe, make-up | Characters with a description, a picture and a voice, approved once | First 15 images free |
| Locations and props | Scouting, art department | Locations and props, each with a picture, fixed for the whole series | Images |
| Board | Storyboard, shot list | A storyboard per episode, shot by shot, with seconds; edited as text | First drama’s episode 1 free; editing free |
| Shoot | Camera, crew, takes | Each segment is generated; every take is kept | 10 credits a second on the default tier |
| Assemble | Edit, sound, score | Segments joined; music scored; subtitles on | 1 credit a second |
| Deliver | Master and delivery | Original-quality download, or a share link | Included |

## Budget a production before you start

One 60-second episode. On SceneMixer’s default tier, video plus the join is about 660 credits, about $4.11 at the lowest credit price, before retakes.

A series from a novel. A 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 at the lowest credit price.

Retakes are the part you control: editing text is free, and only the regeneration of the segment you change is charged. Where credits go in one episode · Spend fewer credits · AI video cost calculator.

## Who it suits

| If you are… | Use it for |
| A writer with a script and no crew | Seeing the script as episodes with a consistent cast, before anyone is hired |
| A creator running a channel | A series with the same faces every episode, in a vertical or landscape shape. Making a multi-episode series |
| A small business with product photos | An ad of 15, 30 or 60 seconds. An ad from one photo |
| A team with footage | A remake with a new cast and look. AI video remake |

## What stays with you

SceneMixer does not decide what the story is or whether a take is right. You write the brief or the script, approve the cast, read the storyboard, and pick the take that goes in. A generation takes minutes, so plan a day for the first episode and less for the next. For frame-accurate cutting, grading or titles, download the original-quality file and finish it in an editor made for that. Composite and share an episode · AI video editor.

## One project, five levels

A SceneMixer project holds a cast and a list of episodes; an episode holds segments; a segment holds every take you made. Almost every confusing moment in the editor is an edit landing on a different level from the one you meant: change a character in the project cast and they change everywhere; change them inside an episode and only that episode follows. What is inside a project.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with one mark each and a voice; the stairwell and the balcony as locations; the letter as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Output | Options |
| Shape | 9:16 vertical or 16:9 landscape |
| Quality | 480P to 1080P, and 2K on some models |
| Segment length | 15 or 30 seconds |
| Dialogue | 15 languages, one voice per character |
| Input | A script, a novel or a one-line idea; photos for an ad; a clip for a remake |

## FAQ

### What is AI video production software?

Software that does the stages of a video production, from script to finished video, with generated pictures and voices instead of a crew and a camera.

### What does SceneMixer replace and what does it not?

It replaces the crew, the camera and the actors. It does not replace the writer or the director: you write the brief or script, approve the cast and choose the take.

### How much does one episode cost?

A 60-second episode on SceneMixer’s default tier is about 660 credits including the join, before retakes.

### Can it handle a whole novel?

Yes. SceneMixer reads long books in two passes into episodes; a 10-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits.

### Can I change a character after the series has started?

Yes. Edit the character in the SceneMixer project cast and every episode not yet generated follows; edit inside an episode and only that episode changes.

### How long does a production take?

Each segment SceneMixer generates takes minutes. Plan around that, and use Generate all once the first segment is right.

### Can I finish the video in another editor?

Yes. Join the segments and download the original-quality file.

### Is there a free trial?

A new SceneMixer account has signup credits, free setup images, a free first storyboard and one free short preview.

Updated November 3, 2026


---

# AI Video Prompts: 15 Copyable Examples and the Formula Behind Them

URL: https://scenemixer.com/guides/ai-video-prompts/
Fifteen AI video prompts you can copy: cinematic shots, a character image, an ad line and dialogue, with the five-part formula and weak-versus-strong rewrites.

# AI video prompts

Quick answer: A good AI video prompt is five things in order: the framing, the camera, who or what is in the picture, the one thing that happens, and what we hear, with the seconds it needs. Plain sentences beat strings of tags, and what is in the picture beats what should not be. Below are fifteen prompts to copy, grouped by job (shots, characters, places, ads, dialogue), each with the reason it is written that way. They are model-neutral; in SceneMixer they go straight into the storyboard as shot lines.

Start with this one and change a noun at a time: [Shot 1] 4s: Medium shot, handheld, following @Mara as she crosses a rain-wet market at dusk, lanterns lit behind her. She stops at a stall and picks up a small brass key. Ambient: rain on awnings, distant voices.

## The formula

Write each prompt as one shot, in the order a viewer notices things:

| Part | What goes in it | Example |
| Framing | One framing: wide, medium, close-up | Medium shot |
| Camera | Still, handheld, tracking, a quick push-in: one movement | handheld, following her |
| Who and where | The person, the place and where things are | @Mara in a rain-wet market at dusk, lanterns behind her |
| The one thing that happens | One action, not three | She picks up a small brass key |
| Sound | Speech and ambience, positively worded | Ambient: rain on awnings |
| Seconds | Enough time for what happens | 4s |
Everything else is the look, and the look is set once for the whole project by the style presets, not repeated in every prompt. How to write one shot · Choosing the look.

## Five cinematic shot prompts

| Job | Prompt |
| Opening | Wide shot, high angle, moving forward over a harbour town at dawn; boats below, mist on the water. Ambient: gulls and a distant engine. |
| Tension | Close-up on @Jun's hands as they hover over the red button, fixed camera, a bead of sweat sliding down his wrist. Ambient: a faint electrical hum. |
| Chase | Handheld, following @Mara as she cuts between market stalls, the crowd blurring past, wet stone underfoot. Ambient: shouting, running footsteps. |
| Reveal | Pulling back from @Elena at the window to show the empty ballroom behind her, chairs stacked, dust in a single beam of light. |
| Quiet moment | Medium close-up, fixed camera. @Teo reads the letter by the stove, lips moving slightly; the kettle begins to whistle. No dialogue. |
Notice what is missing: no cinematic, 8K or masterpiece. Those words add nothing the model can place in the frame, push the useful words down the line, and are sometimes painted literally. Why tags do nothing · Cinematic camera moves, in plain words.

## Three image prompts for a character

A character sheet is a full-length casting photo on a plain white background, so write the person, not the mood: face first, then build, then one outfit. Give the face one anchor nobody else has.

| Character | Prompt |
| A detective | A woman in her late forties, long face, heavy straight brows, deep-set grey eyes, a small scar through the left eyebrow, grey streak at the temple; tall and spare; a charcoal wool coat. |
| A young baker | A man of about twenty-two, round face, wide-set brown eyes, freckles across the nose, a chipped front tooth, short curly black hair; solid build, slightly stooped; a flour-dusted apron over a green shirt. |
| A fantasy elder | An old woman, narrow face, hooded dark eyes, deep lines at the mouth, long white hair in a single braid, a mole under the right eye; small and upright; a patched blue travelling cloak. |
Words like beautiful or charismatic describe nothing that can be drawn and pull toward a stock face. Describing a face · AI character sheet generator.

## Two place prompts

| Place | Prompt |
| An interior | A narrow kitchen in a flat above a shop: a gas stove on the left wall, a small table with two chairs under the window, a door to the stairs at the far end; late afternoon light from the window. |
| An exterior | A market square at dusk, wet cobbles reflecting string lights; stalls with striped awnings along the left side, a stone fountain in the centre, a church tower at the far end. |
Say where the door is, which side the window is on and where the light comes from; a moving camera exposes vague rooms. Your scene image becomes the set.

## Two ad prompts

An ad takes a photo and one line rather than a shot list: what it is and who it is for.

| Product | One line |
| A bottle | A recycled-glass water bottle with a bamboo lid, for people who walk to work and hate buying plastic. |
| A shop | Our family bakery on the corner of Elm Street, open since 1987; for people who want bread that is still warm; walk-ins welcome. |
When the line and the photo disagree, the line wins. Writing the one-line brief · AI ad video generator.

## Two dialogue prompts

| Moment | Prompt |
| A threat | Close-up on @Brook, jaw set. @Brook (low, through his teeth): You had no right. (voice: Brook) Ambient: a clock ticking in the hall. |
| A confession | Medium shot, fixed camera, @Lina on the balcony at night. @Lina (quiet, not turning): I already knew. (voice: Lina) Ambient: far traffic, a door closing. |
About three English words a second: a short line needs about four seconds with its reaction. Writing dialogue in the storyboard.

## Weak prompts and their rewrites

| Weak | Why it fails | Stronger |
| A cinematic 8K video of a girl in a city | Tags the model cannot place; no framing, action or place | Medium shot, @Lina walking through a crowded night market, neon signs reflected in the wet ground. Ambient: sizzling woks. |
| No crowd, no rain, no text | Every noun named is a candidate for the picture | An empty street, dry pavement, a plain wall behind her. |
| She runs, then stops, then turns and cries while the camera zooms out | Four actions and two movements in one shot | Split into three shots of two or three seconds, one action and one move each. |
Say what is there, not what is not.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, American Small Town; Costume Style Soft Everyday, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Paste the three character prompts into the appearance of each character card and the two place prompts into the scenes; the shot prompts go into the storyboard. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Prompt for | Write | Avoid |
| A shot | Framing, one camera move, who and where, one action, sound, seconds | Tags, stacked actions |
| A character | Face, anchor, build, one outfit | Beautiful, charismatic, neutral expression |
| A place | Where the door, window and light are | No furniture, no people |
| An ad | A photo plus what it is and who it is for | Best, number one, guaranteed |
| A line | Name (tone): the line, then who voices it | More words than the shot has seconds |

## FAQ

### What is a good AI video prompt?

One shot: framing, one camera move, who and where, one action, sound and seconds, in plain sentences.

### Do I need magic words like cinematic or 8K?

No. They give the model nothing to place, push useful words down, and sometimes get painted literally. Pick the look with the style preset.

### How long should a prompt be?

A shot is one or two sentences; a character description is a short paragraph; an ad is one line plus a photo.

### Should I list what I do not want?

No. Naming something invites it. Say what is there instead: an empty street, not no crowd.

### Can I reuse these prompts in any tool?

They are model-neutral and written as ordinary sentences, so they carry over; only the @Name tags and the voice marker are SceneMixer's.

### Where do I paste a prompt in SceneMixer?

Shot prompts go into the storyboard text; character and place prompts go into the descriptions in Step 2; an ad takes a photo and one line.

### How do I write a character image prompt?

Face first, with one unmistakable anchor, then build, then one outfit. Leave the lighting and finish to the sheet.

### Does a longer prompt make a better video?

Not by itself. Extra detail helps only if the model can draw it; extra actions hurt.

### Is it free to try these?

Writing and editing the storyboard is free; new SceneMixer accounts start with 50 credits and the first 15 setup images are free.

Updated November 3, 2026


---

# AI Video Tools: What Is Built In, and What You Still Bring

URL: https://scenemixer.com/guides/ai-video-tools/
A finished video needs more than a generator: script, cast, music, captions, cover, upload, label. What is built in here, what you bring, and what tool the rest needs.

# AI Video Tools: The Stack Around a Generator

Quick answer: A generator makes clips. A finished video also needs a story, a cast, sound, captions, a cover, a file in the right shape, an upload and a label, and no single AI video tool covers all of them. This page walks the chain from idea to published video in thirteen stages and says, for each, whether it is built into SceneMixer, what you still bring, and what kind of tool the rest calls for. The pattern is short: story, cast, shots, clips with voices, a score, subtitles and a cover are built in; titles, licensed music, thumbnails for a given platform, the upload and the platform’s own AI declaration are yours.

Try it: write the thirteen stages below on a sheet and tick the ones you already have a tool for. Whatever is unticked is what to line up before your first episode is done.

## From idea to shots: all built in

1. Idea and script. Built in. SceneMixer’s AI Create turns one or two sentences into a script, or you paste your own. Bring nothing, except a story worth telling. 2. Cast and look. Built in. Characters, locations and props are drawn from the script, with state pictures where a look changes. Bring your own photo only if you want a specific face, and that is a paid-plan feature. 3. Shots. Built in. The storyboard splits each episode into numbered shots with seconds, which you can read and edit for free. Four ways to start · What is inside a project.

4. Clips with voices and ambience. Built in. Each segment is generated with its speech, lip movement and ambient sound inside it, at 10 credits a second on the default tier. Bring a voice recording only if you want your own, and only one you have the right to use. Dialogue and sound in one generation.

## Sound and words on screen: half in, half out

5. Music. Built in, with one switch. Each episode is scored for you and mixed under the voices; there is no track picker. If you want your own licensed track, turn Music off, download the version without it and lay yours under the file in an editor. Bring music you have the right to use. There is no separate sound-effects library; ambience arrives with each clip. The built-in score.

6. Subtitles. Built in for SceneMixer accounts that have bought credits or a plan at least once. A switch shows each line as it is spoken and offers a burned-in download, and an exported episode can be downloaded with subtitles or as an SRT file you can upload to a platform. No credits are charged. 6b. Titles and styled on-screen text. Not built in. The site asks for a clean picture, with dialogue carried by sound, so a title card or a lower third is added in an editor after download. Add subtitles · Text in the picture.

## Cover, review and the file

7. Cover. Built in. Three poster options drawn from your leads’ sheets with the title painted in, in portrait and landscape, with episode covers and a share-card picture. Your first drama gets one free batch; ad projects have no cover art. For a platform thumbnail you may still need to crop, or add a short title, in an image editor, because each platform has its own shape and rules. Make a promo cover poster.

8. Private review. Partly. A share link is free, but it is public to anyone who holds it, cannot be password-protected, has a daily play limit and always carries the badge. For controlled review, download the file and send it through a channel you control. Share links.

9. Joined episode file. Built in. SceneMixer’s Composite Episode joins finished segments at 1 credit a second, and the download is the full-quality original; the player and share pages use a lighter copy. Original versus play version.

## Beyond one episode: language, editing, publishing

10. A second language. Built in as a second SceneMixer project: the dialogue language is fixed once the script is read, so you start a new project from the same script, choose the new language and bring the cast across with Pick existing. If viewers only need to follow the story, subtitles on the existing episodes are the cheaper route. A second language version.

11. Cutting with other footage, colour grading, branding. Not built in. There is no timeline and no mixing with your own footage. Fixes to one clip are inside (trim, one of eight picture changes, a re-voice, an enhance); anything that combines files is a conventional editor’s job. 12. Upload and label. Yours. Make the project 9:16 from the start for a vertical feed, upload the downloaded file natively, and turn on the platform’s own “made with AI” setting. The badge in the corner and that setting are separate things, and platforms can limit unlabelled realistic AI content. Posting guide · AI short drama compliance.

13. Backup. Yours. Generated files stay in your account until you delete them, and deleted items are cleaned up about seven days later. Uploaded reference material is kept 30 days by default unless it is in an active project. Download the original of anything you cannot lose. The seven-day window.

## What to ask of any AI video tool stack

Can I export the original file? If the only copy is a streaming one, it is not yours to edit. What stays on the file? Ask whether a mark or a metadata line is added, and whether you can switch it. Can I swap the music? A tool that mixes in a score you cannot turn off blocks your licensed track. Are captions a file or a picture? An SRT you can upload is more flexible than burned-in words only. Is sharing private? Most share pages are public; plan review accordingly. Where does the work stop? The stage where a tool says nothing is usually the stage it does not do.

## A minimum stack for one person

SceneMixer for stages 1 to 10; a simple video editor you already know for titles, your own music and the odd join; an image editor for thumbnails; and the platform itself for the upload and the AI declaration. That is four kinds of tool, and only the editor is optional if you are happy with the built-in score and no titles. Start there, and add a tool when you can name the stage that needs it.

| Stage | Built in? | You bring or do elsewhere |
| 1 Idea and script | Yes | A story |
| 2 Cast and look | Yes | Your own photo (paid plan), if you want a specific face |
| 3 Shots | Yes | Nothing |
| 4 Clips with sound | Yes | Nothing; your own voice recording is optional |
| 5 Music | Yes, on or off | Licensed music to lay in an editor |
| 6 Subtitles and titles | Subtitles yes; titles no | Titles, styled text |
| 7 Cover | Yes | Cropping or lettering for a platform |
| 8 Private review | Partly | A private channel for the file |
| 9 Joined file | Yes | Nothing |
| 10 Second language | As a new project | Nothing |
| 11 Mixing, grading, branding | No | An editor |
| 12 Upload and label | No | The platform, and its AI setting |
| 13 Backup | No | Your own copy of the originals |

## FAQ

### What AI video tools do I need besides a generator?

A way to write the story, a cast, sound, captions, a cover and an upload. SceneMixer covers the story, cast, shots, clips, score, subtitles and cover; you bring titles, licensed music, platform thumbnails and the upload.

### Can I use my own music?

Yes. Turn SceneMixer’s Music switch off, download the version without it and lay your track under the file in an editor. Use music you have the right to use.

### Does it make captions?

SceneMixer builds in subtitles for accounts that have bought credits or a plan once, with a burned-in download and an SRT file. Styled titles and on-screen text are added in an editor.

### Can I make a thumbnail?

SceneMixer can make cover posters in portrait and landscape from your lead characters. Each platform has its own thumbnail shape and rules, so you may crop or add a title elsewhere.

### Can I edit my own footage into the video?

No. There is no timeline and no mixing with your own clips. A remake reads your clip and films it again with new characters. Combining files is done in a regular editor after you download.

### How do I keep a review private?

SceneMixer share links are public to anyone with the link and cannot be password-protected. Download the file and send it through a channel you control.

### Do I need a tool for the AI label?

No tool, but a step: turn on the platform’s own AI setting when you upload. The badge in the corner is separate and does not replace it.

### Do I need to install software?

No. SceneMixer runs in the browser. The only other tools are the ones for stages it does not cover.

Updated November 2, 2026


---

# AI Video Tools for Creators: Where Your Material Goes

URL: https://scenemixer.com/guides/ai-video-tools-for-creators/
Scripts, photos, voice samples, source clips, finished videos and share links: where each goes in an AI video tool, how long it stays, who sees it and how to remove it.

# AI Video Tools for Creators: Where Each Piece of Your Material Goes

Quick answer: Before a creator trusts a tool with material, the useful question is not what it can make but where each piece of what you give it ends up. Here is that map for SceneMixer, material by material: your script, images and voice samples you upload, a clip you give it to remake, the videos and images it makes, share links and your account details. For each: where it goes, how long it stays, who can see it and how to remove it. Then the points the terms of service and the privacy policy make, in their own order, repeated rather than interpreted.

## Seven kinds of material, seven fates

Read the table across: the same question, asked of each kind of thing you hand over or get back.

| Material | Where it goes | How long it stays | Who can see it | How to remove it |
| Script or novel text | Kept in your project (the Novels tab of the asset library); sent to a model provider for each task, to fulfil that request | Until you delete the project | You; support in a read-only view if you report a problem | Delete the project |
| Images you upload (a paid-plan feature; JPG or PNG up to 20 MB) | Stored in your account; a copy is placed at an address the model service can read, possibly in another region | 30 days by default, unless kept in an active project | You | Delete the sheet, or the project that holds it |
| Voice sample (a paid-plan feature) | Trimmed to a short reference and attached to a character | While the character keeps it | You | Clear it on the character, or delete the project |
| A clip you give it to remake | Read once into text and a few still frames; never handed to the video model | Queued for deletion when the read completes, gone about a week later | No one else; it is not shown on any page | It removes itself |
| Images and videos it makes for you | Stored at the site's expense, in your account | Until you delete them | You | Delete the sheet, clip, episode or project |
| A share link | A public page that plays one clip | No expiry date; stops when the clip, episode or project is deleted | Anyone who has the link | Delete what it points to |
| Account details | Held by the service | Life of the account, then up to 6 months after deletion | You | Close the account in the app, or ask for deletion by email |

None of this is hidden in the interface, and most of it is in the guides: where your files live, the asset library and its quota and what happens to a clip you upload for a remake.

## What the terms and the privacy policy say, in their own order

This section repeats points from the two pages; it does not interpret them, and it is not advice. For anything that matters to you, read the pages themselves: terms of service and privacy policy.

What is collected. The privacy policy lists account details (an email, a nickname, and a password stored only as a salted hash), usage (the prompts and scripts you submit, generation records, the model invoked, timestamps, IP and device identifiers), payment references from payment processors (the site never receives a card number), and the images, video and audio you upload as generation references.

How it is used. To provide the service, to process content through third-party model providers (the policy says your prompts and reference material are sent to the provider for a task solely to fulfil your request), for billing and fraud prevention, and for security. The policy states that personal information is not sold.

Retention. Account data lasts the life of the account plus up to 6 months after deletion. Uploaded reference material is kept 30 days by default, unless kept in an active project. Deletion requests sent by email are honoured within 30 days unless retention is legally required.

What the terms say about your content. Section 06 of the terms covers your uploads, generated results, the licence you grant the service and public use. Among its points: you keep rights in your uploads and state that you may use them for generation; you grant the service a licence solely to run it (hosting, processing, forwarding to model providers, showing results back to you), to run content-safety review and to improve quality on a de-identified basis; unless you separately consent or the law requires it, uploads and outputs are not used to train models, and the upstream providers process data under their own terms; and content you publish publicly, submit as a sample or enter into a campaign is treated as licensed for showcase. The same section has a paragraph on generated results and commercial use, with its own conditions and exclusions. Read that paragraph in full instead of relying on anyone's summary, including this one.

## Who can see a project

Nothing you make on SceneMixer is public until you make a link. A project is visible to you. Only the project's owner can create a share link for a clip or a finished episode, and the link is then readable by anyone who holds it, with or without an account; the page can be found by search engines, so treat a link as public once it exists. Share links: who can see them, whether they expire has the daily play limit and how a link is taken back.

When you report a problem, support may open your account in a read-only view that can read projects, storyboards, clips, credit history and orders and cannot start generations, move credits, change settings or delete anything. Every request in that mode is logged with the staff member's identity. What support can see and do. The finished dramas listed as Featured Projects are samples opened read-only; they are not other users' projects.

## References that show real people

Three SceneMixer behaviours matter if your material includes photos or recordings of real people. The first time you upload, a window asks you to confirm that you have the right to use the material and that every real person whose face or voice appears in it has given consent. Reference images are also checked by the video provider when a segment is generated, and a generation is refused when a face in them looks like a real person; the segment is marked failed and its credits come back on their own. That message and the fix are in their own guide. And for a remake, the read does not check who owns the clip; that is left to you. The acceptable-use section of the terms (section 04) covers likeness of real people in more detail.

## Deleting: what is immediate and what is slow

Deleting a project, episode, sheet, clip or upload removes it from your SceneMixer account at once: it leaves the lists, shared links to it stop working, and storage it used is released. The files behind it are queued and removed about seven days later. That week is not a recycle bin you can open; treat deleting as final. A sheet that another project still uses is kept for that project, because references are checked before a file is removed. Closing the account erases the personal details on it and clears any remaining credits without refund; the account row is kept because orders and payments have to stay auditable. Deleted projects and files: the seven-day window and deleting your account have the steps.

## Four things to do before you upload

Ask if the step needs the upload at all. Most SceneMixer steps run from words alone; uploads are a paid-plan feature and every one counts against your storage. Check where it will sit. Material outside an active project is kept 30 days by default; a copy is also placed where the model service can read it. Keep your own copy. Download originals of anything you cannot afford to lose, since deleting here is one-way. Know who is in the picture. If a real person's face or voice is in it, make sure you have their say before you press Confirm and upload.

## FAQ

### Are my scripts and videos public?

No. A SceneMixer project is visible to you, and to others only through a share link that you create. Anyone holding a link can open that one clip.

### Are my uploads used to train models?

The terms say that, unless you separately consent or the law requires it, uploads and outputs are not used to train models, and that upstream providers process data under their own terms. See section 06 and the privacy policy.

### How long are my uploaded photos kept?

30 days by default, unless they are kept in an active project, according to the privacy policy. You can delete them earlier.

### What happens to the files when I delete a project?

The project leaves your SceneMixer account at once and its files are removed about seven days later. There is no restore on the site.

### Can support see my account?

When you report a problem, SceneMixer support may open your account in a read-only view. It cannot start generations, change settings, move credits or delete anything, and every request in that mode is logged.

### Where is my material processed?

Prompts and reference material are sent to the model provider for a task, and references handed to a model service are copied to storage it can read, which may be in a different region from your account's data. Servers for the account itself are in Singapore.

### What does the terms page say about commercial use of results?

Section 06 has a paragraph on generated results and commercial use, with its own conditions. It is better read in full than summarised.

### Is a share link private?

No. A SceneMixer share link is a public page for anyone who has the link, and it has no expiry date. It stops working when the clip, episode or project it points to is deleted.

Updated November 2, 2026


---

# AI Videos: Three Kinds You Can Make, With Real Length, Shape and Cost

URL: https://scenemixer.com/guides/ai-videos/
Three kinds of AI video on one site: story episodes, 15 to 60 second ads and remakes of your own clip. Real length, shape, quality and cost for each, and how to choose.

# AI Videos: The Three Kinds You Can Make, and What Each Costs

Quick answer: “AI videos” covers very different objects, and what you can make depends on what you hold. Here there are three kinds. A story episode comes from a script, a one-line idea or a novel and runs as long as the script does. An ad comes from up to four photos and is 15, 30 or 60 seconds. A remake comes from a clip you shot, read shot by shot and filmed again with a new cast. They share one pipeline but differ in input, length and the way the price is worked out. This page gives each kind’s real specs and one costed example, so you can see what a given amount of money buys before you start. For scale: a 60-second story episode on the default tier is about 660 credits, $4.11 at the lowest credit price.

Try it: write down where the video will be watched (a vertical feed, a product page, a channel) and what you already own (words, photos, a clip). Those two answers pick the kind below almost every time.

## Kind 1: a story episode

You bring: a script or a novel (.txt, .docx, .md or pasted), or one line of idea that SceneMixer’s AI Create tab writes into a script. You get: a cast with consistent faces, locations and props, a storyboard for each episode, and video with spoken dialogue, ambience and a score. The read splits the script into episodes of a minute or two each, so a longer script means more episodes, not longer ones.

Shape and length. Landscape 16:9 or portrait 9:16, chosen once in Step 1. One generation is a segment of up to 15 seconds, or up to 30 seconds on the long-take tiers; the segments are joined into the episode, and an episode can hold up to 60 of them. A segment may be much shorter than its ceiling: two short shots can run only a few seconds. Quality is the tier you pick per segment, from 480P up to 2K. Dialogue can be in any of 15 languages, one voice per character.

Cost shape. Reading the script is paid once, at 4 credits per 3,333 English characters. Each episode’s storyboard is 1 credit per 100 counted characters. Video is 10 credits a second on the default tier and the join is 1 credit a second. Worked examples at the default tier: a 60-second episode is about 660 credits ($4.11); a 90-second episode about 990; a series of ten 90-second episodes from a 30,000-character novel about 10,026 credits, $62. All before retakes. Where credits go in one episode · AI movies: how long they run.

## Kind 2: an ad of 15, 30 or 60 seconds

You bring: one to four photos of the same product, dish, shop or place, and optionally one line saying what it is and who it is for. You get: an ad with a script, a cast, sets, sound, music written for it and an end card carrying your brand.

Four pickers decide it in SceneMixer. Ad vibe (Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life) changes the story and the end-card colours but not the price. Length is 15, 30 or 60 seconds, which is one, two or four 15-second segments, each shot on its own and joined with the end card last. Format is landscape or portrait. Quality is 480P, 720P or 1080P, and the menu prints the all-in price beside each one. Length and quality move the price; vibe and format do not.

Cost shape. The one price on the start button is a cap, not a bill. Reading the photo is free, cleaning it is one sheet, the script is a writing charge, the breakdown is a read, sheets are per image, shots are per second at the tier, and the final cut is per second. Each stage charges its own actual price, so the total lands at or below the number you saw. An ad is never the free first project, and it cannot draw on the free setup images. The ad’s one price is a cap · Ad vibe, length, format and quality.

## Kind 3: a remake of a clip you shot

You bring: one MP4 or MOV, up to 5 minutes at the time of writing and within the megabyte limit printed on the upload box, and optionally one line naming who is who. You get: the clip read shot by shot into a cast, sets, props and a finished storyboard, then filmed again with your own characters in the look and language you choose.

What carries over and what does not. The blocking, the order of shots and the dialogue structure are read from your clip. The faces are not: the source is never handed to the video model, and the people in the result are your cast’s. Reading a clip does not give you a licence to its story, so use material you have the rights to.

Cost shape. The read is priced by the clip’s length alone, in minutes rounded up; the number of cuts never enters the price. Everything after the read (sheets, storyboard, video, join) is priced as in a story project. How to remake a video with AI · Source video limits, read time and price.

## The three side by side

| | Story episode | Ad | Remake |
| You bring | Script, novel or one-line idea | One to four photos | One MP4 or MOV |
| Length | As long as the script; segments of 15 or 30 seconds | 15, 30 or 60 seconds | Follows the clip; you edit the storyboard |
| Shape | 16:9 or 9:16 | Landscape or portrait | 16:9 or 9:16 |
| Quality | 480P to 2K per segment | 480P, 720P or 1080P | 480P to 2K per segment |
| Price shape | Read once, storyboard per episode, video per second, join per second | One capped price; stages charge their own | Read by minutes, then as a story |
| Free first project | Yes: up to 15 setup images and the first storyboard | No | No |

All three end in the same SceneMixer editor: cast, sets, props, episodes and segments. You can fix one shot by editing its text and regenerating only that segment. Four ways to start a project.

## Choose by where it will be watched

| Where it will be watched | Start with | Why |
| A vertical feed (TikTok, Shorts, Reels) | A 9:16 story episode of about a minute | Open on a hook and end on a turn; the shape cannot change after Step 1. Posting guide |
| A product or shop page | A 15 or 30 second ad | One idea, your brand on the end card, a price you can see first. Ad from one photo |
| A channel with a returning cast | A multi-episode story in 16:9 | One cast carries across every episode. Making a series |
| A scene you already shot | A remake | Keeps your blocking, replaces the people. Recast a remake |

## What none of the three is

SceneMixer builds these for stories, ads and remakes with people in them. They are not screen recordings, slide narration, or a single still brought to life on its own, and they do not keep a real person’s face from a photo or use famous characters and brand mascots. A generation takes minutes, not seconds, and two identical requests can give two different clips, which is why you read the storyboard before you pay for video.

| Kind | Starts from | Typical first spend |
| Story episode | Words | The read, then setup images (15 free), then one segment |
| Ad | Photos | The capped price on the start button; stages charge less if they can |
| Remake | A clip | The read, priced by the clip’s minutes |

## FAQ

### What kinds of AI video can I make here?

SceneMixer makes three: story episodes from a script, idea or novel; 15, 30 or 60 second ads from up to four photos; and remakes of a clip you shot, filmed again with your own cast.

### How long can an AI video be?

A single SceneMixer generation is up to 15 seconds, or 30 seconds on the long-take tiers. An episode joins segments, and a series is as long as the script. Ads are 15, 30 or 60 seconds.

### How much does an AI video cost?

At SceneMixer’s default tier a 60-second story episode is about 660 credits, $4.11 at the lowest credit price, before retakes. An ad shows one capped price before you start.

### Which kind is cheapest to try?

A story project, because the first drama gets free setup images and a free first storyboard, and each account has one free preview of up to 5 seconds. Ads and remakes do not.

### Can I make a vertical video?

Yes. Choose 9:16 Portrait in SceneMixer’s Step 1 of a story or remake, or Portrait format in an ad. The shape cannot change after you continue.

### Does the video have sound?

Yes. SceneMixer speaks dialogue inside each clip, one voice per character, with ambience, and each episode gets a score you can switch off.

### Can I turn one photo into a moving picture?

Not on its own. SceneMixer turns a photo into an ad, or the face or the room that a story follows. A single still brought to life is not what this makes.

### Can I use my own footage?

You can remake it: the clip is read and filmed again with new characters. The source clip is not given to the video model and the faces are not kept.

Updated November 2, 2026


---

# AI Videos for YouTube: Format, Length, Cover and Label

URL: https://scenemixer.com/guides/ai-videos-youtube/
Making AI videos for YouTube: landscape 16:9, how long one episode can be, the quality tier, a cover, captions, the AI badge and how to upload. What is set where.

# AI Videos for YouTube: A Specification Sheet, Setting by Setting

Quick answer: An AI video for YouTube is a 16:9 file you upload yourself, and almost everything that decides how it turns out is set before the first clip is generated. In SceneMixer that means landscape in step 1 (the default), a type that sets how long an episode may be, a quality tier chosen for the final pass, and a cover and captions made after the episode is joined. There is no upload button: you download the original-quality file and upload it in YouTube Studio. This page goes through each setting, what it costs and what is on the finished file, including the small badge and the separate disclosure step YouTube asks for.

## The frame: 16:9 is the default, and it locks

A new SceneMixer project starts as 16:9 Landscape, marked Recommended and described as film and large screen. The frame applies to every video in the project and is fixed once you continue past step 1; to make the same story in the other shape, start a new project with the same script and bring the cast across from the asset library. Setup images are the exception: character sheets are always portrait and location and prop images are always widescreen, whatever the project is. Choosing 16:9 costs the same as 9:16, because video is priced by the second and not by the frame. Vertical or landscape.

If your channel also posts vertical Shorts, make the Short as its own 9:16 project from the same script rather than cropping the wide one. Faceless YouTube channel with an AI drama series and an AI anime series for YouTube show how a channel uses both shapes.

## Length: what one episode can be

In SceneMixer, length comes from two places. The type you choose in step 1 sets the longest an episode may run, and the segments inside it set the grain.

| Setting | Value | Notes |
| Short Drama | Episodes up to 240 seconds | Fast, hook-driven, dense dialogue |
| TV Series | Episodes up to 300 seconds | Room for a scene to build; the choice for ensembles and several threads |
| Film | No episode cap | Slow, atmospheric; still generated segment by segment and joined |
| One segment | 4 seconds up to 15 or 30 | Standard 15s by default; Long take 30s only on tiers that accept 30 seconds |
| Segments in an episode | Up to 60 | On paper that is around half an hour of finished video; most episodes are far shorter |

A longer cut is more segments, joined: a second costs the same whether it is the first or the four-hundredth, and joining adds 1 credit a second of the finished length. One minute of finished video on the default tier comes to 660 credits, about $4.11 at the lowest credit price, before any retake. The joined episode is made of hard cuts; the joins are where continuity has to be planned in the script. Short Drama, TV Series or Film and how to make an AI video longer than one minute go further.

## Picture quality: draft low, finish once

In SceneMixer, resolution is a property of the tier, not a separate setting, and the picker shows each tier's resolution next to its price per second. The cheapest tier delivers at 480P; the steps above it reach 720P and 1080P, and the top tier is 2K. For a watch-page video, the usual habit is to draft every segment on the cheapest tier, fix the text and the sheets, then generate the keepers once on the tier you want to publish. When every segment of an episode comes from the same tier, joining copies the streams and nothing is re-compressed; mix tiers in one episode and the join re-encodes onto a single canvas, which costs a generation of quality. So re-shoot all of an episode's keepers at the final tier, not some of them. Draft low, finish high and What comes out.

Download gives the original-quality file. The site's own player streams a lighter copy at the same resolution, so judge sharpness on the downloaded file, not in the browser. Download: the original versus the play version.

## Cover, captions and music

Cover. In step 2, SceneMixer’s Cover art tab paints three poster options from your lead characters, with the title drawn into the picture; both a landscape 16:9 and a portrait 9:16 version come with the one you pick, and the landscape one is the one to download for a watch page. The first set of three is free on your first drama; each extra size (3:4, 1:1, 4:3) is priced as one image. Keep the title to 12 characters or fewer and check its spelling on each option, because the image model draws it. The site does not make thumbnails as such; the poster is a starting point for one. Promo cover posters.

Captions. On accounts that have bought credits or a plan at least once, exporting a whole episode with Add subtitles ticked gives a burned-in version and a plain-text SRT file with times; the words are your script's lines, so invented names come out as you spelled them. The SRT is a standard file you can upload with the video. Subtitles are free, with a daily limit per account. Add subtitles to your AI drama.

Music. Every episode is scored automatically and the download carries the score. A Music switch turns it off and gives the dialogue version, which is the one to take into an editor if you are laying your own track. Background music in an AI short drama.

## Two different labels

Two things get called a label, and a YouTube upload can involve both. The first is on the file: a small SceneMixer AI badge in the bottom-right corner of the picture, drawn into the pixels, and a line of AI-generated-content metadata written into the file. A free account's videos always carry it; members' new clips come without the badge by default, and the switch beside the model picker applies to clips generated afterwards. A share link always shows it. The AI badge switch.

The second is YouTube's own disclosure step in the upload flow, which is the uploader's to complete and which the badge does not replace. This site's answer page on labels, read on October 14, 2026, summarises YouTube's rule this way: creators disclose when AI meaningfully alters or generates photorealistic content; photorealistic uploads can get a label in the player, stylised or animated ones a note in the description, and disclosing does not limit a video's audience or its eligibility to earn. Rules change, so check YouTube's current page before you upload. This page is not legal advice. Do AI-generated videos need a label or a watermark? has the sources, and can you monetize AI videos on YouTube and TikTok? covers the earnings question.

## What the tool does not do for a YouTube upload

It does not upload for you: the share menu lists social sites for sending a link, and YouTube is not one of them. It does not write a title, description, tags or chapters. It is not a timeline editor, so an intro, an end screen or a cold open with a title card are added in a video editor after you download the file. A SceneMixer link is fine for sending a preview to a collaborator, but a video uploaded to YouTube itself is the one that plays there.

## FAQ

### What aspect ratio should an AI video for YouTube be?

16:9 Landscape, which is the default in SceneMixer’s step 1. It is fixed once you continue, so choose it before the script is read. For Shorts, make a separate 9:16 project.

### How long can an AI video for YouTube be here?

SceneMixer caps an episode by type, 240 seconds for Short Drama and 300 for TV Series, and Film has no cap. An episode holds up to 60 segments of up to 30 seconds each on the long-take tiers.

### Can I upload to YouTube straight from SceneMixer?

No. Download the original-quality file and upload it yourself. The share menu is for sending a link.

### How do I get a thumbnail?

Make cover art in SceneMixer’s step 2 and download the landscape version; it is a poster with your title drawn in, meant as a starting point for a thumbnail.

### Can I upload captions with the video?

Where your account has the Subtitles switch (accounts that have bought credits or a plan once), a whole-episode export includes a plain-text SRT file that YouTube accepts alongside the video.

### Does the video have a watermark?

A small SceneMixer AI badge in the bottom-right corner, drawn into the picture, plus a metadata line in the file. Members' new clips are badge-free by default and can switch it back on.

### Does the badge replace YouTube's AI disclosure step?

No, they are separate. The badge is on the file; the disclosure is a step in YouTube's upload flow that the uploader completes. Check YouTube's current rules; this site's labels answer summarises them as of October 14, 2026.

### What does a long video cost?

Seconds times the tier's per-second price, plus 1 credit a second of the finished length for joining. A minute on the default tier is 660 credits before retakes.

Updated November 2, 2026


---

# AI Visuals: What It Can Change in a Finished AI Video Clip

URL: https://scenemixer.com/guides/ai-visuals-what-it-can-change-in-a-finished-clip/
AI visuals re-renders a finished segment with one thing changed: clothes colour, time of day, weather, background people, lamps, a scar, an expression or the actor.

# AI Visuals: What It Can Change in a Finished Clip, and What It Cannot

Quick answer: AI visuals is the first tab of Edit this clip. You pick one change from a fixed menu of eight — clothes colour, time of day, weather, background people, lights, a scar, an expression, or swapping one character for another — and the clip is made again from itself with only that changed. It cannot change the framing, the camera or where people stand. It costs 2× the per-second price of the tier the clip was made on, for the length of the clip, is refunded if it fails, and saves a new version.

Cheapest way to see what it does: a short clip made on Wan 3.0 · 480P, and a Time of day change to Night — 20 credits per second of clip.

## The eight changes

- Clothes color — red, blue, black, white, green, yellow, purple or gray, for the character you pick under Who. Clothes, scars, expressions.
- Time of day — Daytime, Night or Dusk. Weather — Rain, Clear sky or Snow. Day to night, rain.
- Background people — Clear them out, or Add a crowd. Lights — Light them up (lamps, candles, lanterns in the scene). Background people and lamps.
- Scar — Add or Remove, on the character you pick. Expression — Terrified, Calm, Smiling, Angry, Sad or Crying.
- Swap person — pick Who and Swap with: another character of this project that has a character sheet. Swap a character.

SceneMixer has no free-text box, on purpose: these are the changes that come out reliably. A change the menu does not offer is a job for the storyboard and a regeneration.

## What stays the same

The clip is made again with your clip as the starting point, so how it was shot carries over: the framing, the camera, where each person stands, the movements, the timing and the spoken lines. Under the button the panel says it plainly: Can’t change framing, camera or where people stand. It also cannot remove a duplicated person, clean up garbled text in the picture, or stop a crowd that walks in half-way. For those, edit the shot and regenerate. When a shot comes out wrong.

Like any generation, the result can differ from your clip in small details. Watch it before you composite; if you prefer the old one, switch back in Version History.

## How to run one

- Select the segment and open Edit this clip → AI visuals.
- If the clip is longer than 15 seconds, choose First half or Second half first. Why.
- Pick the change, then its option. For clothes, scars, expressions and swaps, pick Who; characters seen in this segment are listed first, and if only one is in it, they are picked for you.
- Press Edit into a new version; SceneMixer shows the price on the button. The panel shows AI is changing the visuals… until it is done, usually a few minutes.

## Price, tier by tier

AI visuals is charged at 2× the per-second price of the tier the clip was made on, for every second of the clip (rounded up to a whole second; at least two seconds). The reason for the multiple: the edit reads your clip in full and produces a new one, and both are paid for. Clips made on the two MiniMax H3 tiers are edited on Wan 3.0 · 480P and priced there. There are no free edits; the credits are held when you press and refunded if the edit fails. All fix prices.

## Where the result goes

A new version labelled AI edit in Version History. It becomes active when it lands if the clip you edited is still the one showing. A SceneMixer segment keeps one AI-edited version: a second AI visuals edit replaces the first. The generated clip it was made from stays. How versions are kept.

## When it is not available

Not while the segment or the episode’s storyboard is generating, and not while another AI visuals or Redub change to the same clip is running. A few clips made by an earlier version of the site must be regenerated first. An edit blocked by content review is refunded and counts toward the same pause as a blocked generation. Every message.

| Tier the clip was made on | Generating, credits per second | AI visuals, credits per second of clip |
| Wan 3.0 · 480P | 10 | 20 |
| Wan 3.0 · 720P | 19 | 38 |
| Wan 3.0 · 1080P | 38 | 76 |
| Seedance 2.5 · 480P | 19 | 38 |
| Seedance 2.5 Lite | 38 | 76 |
| Seedance 2.5 Pro | 75 | 150 |
| Seedance 2.0 · 480P | 13 | 26 |
| Seedance 2.0 Fast | 25 | 50 |
| Seedance 2.0 | 63 | 126 |
| MiniMax H3 | 25 | Edited on Wan 3.0 · 480P, 20 |
| MiniMax H3 Fast | 16 | Edited on Wan 3.0 · 480P, 20 |

## FAQ

### Can I type what I want changed?

No. AI visuals on SceneMixer only offers its eight changes. To change something else, edit the storyboard and regenerate the segment, or edit the text with AI edit and regenerate.

### Does it keep the dialogue?

The lines and the timing on SceneMixer follow your clip. To change what is said, use Redub.

### Why does it cost twice the generation price?

SceneMixer processes your existing clip and produces a new one of the same length; both halves of that work are charged.

### Can I do two changes at once?

No, one change per edit on SceneMixer. Run a second edit on the result; each is charged, and the newer AI-edited version replaces the older one.

### Is it cheaper to regenerate?

Yes: on the same SceneMixer tier, regenerating costs half as much. But it rolls the dice on the whole clip again; AI visuals is worth it when you want to keep everything else exactly as it is.

Updated November 3, 2026


---

# AI YouTube Channel Ideas: 16 Faceless Formats and How to Make Each

URL: https://scenemixer.com/guides/ai-youtube-channel-ideas/
Sixteen AI YouTube channel ideas with the format, length, cast size and what repeats every episode, plus how to cost a weekly schedule and where side-hustle ideas stop.

# AI YouTube channel ideas

Quick answer: The AI channel ideas that work as a series share one trait: something repeats every episode, so the cast, the places and the look are drawn once and reused. Below are sixteen formats, each with the project settings, the length, the size of the cast and a guide that goes into the detail. The ideas are formats, not forecasts: this page makes no promise about views or income, and money questions are answered where the facts are. Pick by what you can keep up weekly, because a channel is mostly a schedule.

Brief for AI Create to test an idea: A weekly 90-second series: a night-shift baker in a small town fixes one stranger's problem each episode with bread; one bakery, one street, three recurring characters; each episode ends on a new customer at the door.

## How to choose

- What repeats? A person, a place, a rule or a question. If you cannot say it in a sentence, the series has no engine.
- How big is the cast? Every character is a drawing and a voice. Two or three recurring characters and one or two places is the cheap end; the first 15 setup images on an account are free.
- How long is an episode? Video is priced by the second: 10 credits a second on the default tier plus 1 credit a second to join. A 60-second episode is four 15-second segments.
- Can you keep it up? A segment usually takes ten to twenty minutes to come back, and you can close the page. A weekly episode is a day of work, not a week.
Run a faceless YouTube channel · Making a multi-episode series · Why a series stays consistent.

## Sixteen formats

| Idea | Settings | Cast and places | What repeats | Guide |
| Reddit-style confessions | Short Drama, 9:16, 60–120 s | 2–3 people, 1–2 places | A new story every episode, the same narrator | Reddit story |
| Fairy tales and fables | Short Drama, 16:9, 90–180 s | 2–4 characters | A moral, a different tale | Fairy tales |
| Horror shorts | Film, 16:9 or 9:16, 60–120 s | 1–2 people, 1 place | A rule that breaks at the end | Horror short |
| Kids' cartoons | Short Drama, 16:9, 90–180 s | 3 characters, 2 places | A lesson and a song-like ending | Kids channel |
| Anime series | TV Series, 16:9, up to 300 s | 4–5 characters | A cliffhanger each episode | Anime for YouTube |
| Vertical micro-drama | Short Drama, 9:16, 60–240 s | 2–3 leads | A hook in the first seconds, a hook at the end | Vertical dramas |
| History episodes | Short Drama, 16:9 | 2–4 figures, 1–2 places | One event per episode | History drama |
| Documentary-style stories | Film, 16:9 | A narrator and 1–3 people | One question per episode | Documentary style |
| Language-learning scenes | Short Drama, 9:16, 60 s | 2 people, 1 place | One situation, one phrase set | Language learning |
| Talking animals | Short Drama, 9:16, 30–60 s | 1–2 animals | A joke with the same two voices | Talking animals |
| Sitcom and sketch | Short Drama, 16:9, 60–120 s | 3–4 people, 1 set | A running gag | Sitcom or skit |
| Mockumentary | Short Drama, 16:9 | An office, 3 people | Talking-head asides | Mockumentary |
| Game lore and trailers | Film or Short Drama, 16:9 | A world, 2–3 heroes | One region or faction per episode | Game trailer |
| Tabletop campaign recaps | TV Series, 16:9 | The party, the dungeon | One session per episode | Tabletop campaign |
| Bedtime stories | Short Drama, 16:9 or 9:16, 120–240 s | 1–2 characters, gentle places | The same narrator and a calm ending | Bedtime stories |
| Fictional crime stories | Short Drama, 16:9 | A detective, 1–2 places | A case per episode; invented people only | True-crime style |
Real people, celebrity faces and known brands are not used as characters; every figure here is invented. Content rules.

## Cost a weekly schedule before you start

Seconds are the unit. A weekly 90-second episode is 90 seconds of video, six segments; four of them in a month are 360 seconds, plus the same seconds again at the joining price, plus any retakes. Draft on SceneMixer’s default tier, fix the storyboard where it is wrong, and only then regenerate the keepers on a sharper tier. What a micro-drama costs · Draft cheap, finish sharp · AI video cost calculator.

| Episode length | Segments of 15 s | Video seconds |
| 60 seconds | 4 | 60 |
| 90 seconds | 6 | 90 |
| 120 seconds | 8 | 120 |
| 240 seconds (the Short Drama maximum) | 16 | 240 |

## Video ideas that are not a channel

Two other uses make videos without needing an audience of your own: a greeting or a story for one person (a birthday, a memorial, a tribute), and an ad for a local business, made from the owner's photos and one line. A birthday story video · Video for a small business.

## Where the side-hustle talk stops

There is no figure here for views or earnings, and none should be trusted from a page that sells the tool. What can be said is how to think about it: the cost of an episode is knowable in advance, the audience is not; client work (an ad for a shop) has a price you can quote, and SceneMixer quotes the video to you as a ceiling before it starts. Platform rules on AI content and on monetisation are the platforms' own and change; read them where they are published. Can you make money with AI videos? · Monetising on YouTube and TikTok · How much to charge for a project · Short drama and AI rules in 2026.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style American Small Town, Pastoral Healing; Costume Style Soft Everyday, Smart Casual; Visual Style Soft Glow, Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the three recurring characters and the one bakery and street; they are reused in every episode. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| If you want | Start with | Length |
| The cheapest series | One place, two characters, a repeating joke or lesson | 30 to 60 seconds |
| A feed-first channel | A vertical micro-drama with a hook and a cliffhanger | 60 to 120 seconds |
| A watch-time channel | A landscape series with a narrator | 120 to 300 seconds |
| Income without an audience | Ads for local businesses from their photos | 15 to 60 seconds |

## FAQ

### What are good AI YouTube channel ideas?

Series where something repeats: a confession narrator, a fable, a horror rule, a kids' lesson, an anime cliffhanger, a sitcom running gag. Sixteen are listed above with settings and guides.

### Can I run a faceless YouTube channel with AI video?

Yes. The cast on SceneMixer is drawn once and reused, so no one has to appear on camera. The faceless channel guide covers the routine.

### How long should an episode be?

Sixty to ninety seconds for a feed-first series, longer for a landscape channel: a Short Drama episode on SceneMixer runs up to 240 seconds and a TV Series episode up to 300.

### How much does a weekly episode cost?

Video is priced by the second: 10 credits a second on the default tier plus 1 credit a second to join. Multiply by the seconds you plan, and add retakes.

### Is there a cheap way to start?

Yes: one place, two characters, 30 to 60 seconds. The first 15 setup images on SceneMixer are free and the first storyboard on your first drama is free.

### Can it make money?

This page makes no income claim. The cost of an episode can be known in advance; views and earnings cannot. See the answer pages on making money and on monetisation.

### Can I make videos for clients?

Yes, for example ads on SceneMixer for local businesses from their photos and one line. You need the right to use the photos.

### Can I use real people or celebrities?

No. SceneMixer does not use real faces, celebrities and known brands as characters, and uploading a real person's photo needs their consent.

### Which format is easiest to keep consistent?

Two or three recurring characters in one or two places, because SceneMixer draws each once and reuses it in every shot.

Updated November 3, 2026


---

# Also Called: Finding the Right Character by Nickname or Old Name

URL: https://scenemixer.com/guides/also-called-nicknames-and-aliases/
Where the other names on a character, scene or item card come from, what they change in the storyboard, and how to fix one that lands on the wrong person.

# Also Called: Nicknames, Old Names, and How the Storyboard Knows Who Is Who

Quick answer: Every character, scene and prop card can carry a short list of the other names your script uses for the same thing — a nickname, a short form, an address like “Cousin” or “Mom”, or an old name left behind by a rename. In step 2 it shows as the grey line under the card’s name: Also called: Danny, Dan. It is not decoration. It is how the storyboard still knows that “Danny” in your script is Daniel Cole — his card, his sheet, his voice. The list is read-only: it comes from the script read and from renames, and there is nothing to type into.

In step 2, click a character, scene or prop card: the grey line under the name lists the other names your script calls it.

## What the list is

Cards are matched by name: the speaker of a line, the in-frame list, the place in a segment header. Scripts rarely say a name the same way every time — halfway through, Daniel Cole is “Danny”, and the Riverside Boardroom is just “the boardroom”. So each card can keep those names: Daniel Cole with Also called: Danny, Dan, the scene Riverside Boardroom with Also called: the boardroom, the prop Grandmother’s Jade Pendant with Also called: the pendant.

For a character it can also be an address only they answer to — Cousin, Mom — or, for a creature, its species word. Words that could point at anyone are deliberately not collected: the stranger, the man, the princess. A name that fits half the cast cannot be resolved to one of them.

## Where the names come from

There are exactly two sources.

- The script read. When SceneMixer reads your script, each person, place and thing gets the other names your own text uses for it — nicknames, short forms, addresses, and for scenes and props the alternative names or translations the text uses. If the text only ever writes the full name, the card’s list stays empty and the line is simply absent. A card keeps up to eight entries.
- Renames. Rename a card in step 2 and the old name moves into the list, so anything that still says it — your script, a storyboard written earlier, notes you keep — keeps pointing at the same card. Two cases add nothing: a rename that only changes case, spacing or punctuation (the name is unchanged once normalized), and an old name that is already another card’s name (adding it would make two cards answer to one name).

Reading the script again does not reset the list — new names are merged in, old entries stay.

## Where you see it

- Open the project in SceneMixer and go to step 2, Characters & Scenes.
- Click a character, scene or prop card. Its detail opens: the picture, the name, the badges.
- Right under the name is the grey line — Also called: Danny, Dan — written the way your script writes them.
- The Previous / Next arrows at the top move between cards of the same kind.

A card without the line has no other names in your script. That is normal, not an error.

## Can you edit or add to it?

Not directly: there is no field to type into, and an entry cannot be deleted by hand. The two sources above are the levers:

| You want | Do this |
| A nickname the card should answer to | Write it in the script and read the script again; the name merges into the list |
| An old name to keep working after a change of mind | Rename the card: the previous name becomes an entry by itself |
| An entry gone | Not possible on the card; keep the word for this character only in the script, or rename the card so the word is no longer part of its name |

One thing that adds nothing: writing the nickname into the card’s description. The description is written for the picture; matching reads the name and the list.

## What it changes in the storyboard and the clip

- The cast list SceneMixer hands to the storyboard carries each card’s other names next to the main one, so a nickname in the script can still be tied to the right character while the shots are written.
- When the storyboard is built, a name in a character slot — the speaker of a line, the in-frame list — is looked up against the cast: the full name first, and only if nothing matches, the other names. A match counts only when exactly one character answers to that name.
- Matched, the line belongs to that character: it is voiced, rendered and counted in the picture as them — their voice, their reference sheet.
- Not matched, the line is treated as coming from someone outside the cast: it does not get the character’s voice, and the character himself is kept lips-closed in the shot — the clip behaves as if a stranger were speaking while he stands there.
- For a prop, being mentioned under any of its names counts as used in this episode — that is what keeps the prop in front of the storyboard when the shots are written.
- Only characters are matched by their other names. Scenes and props are matched by their full names: their entries travel to the storyboard model as a hint (so it reads “the boardroom” as the Riverside Boardroom while writing), but a scene will not be pulled in from a nickname. Write places and things the way their cards are named.
- None of it shows on screen: the clip says the words written in the shot. The list only decides who says them and who stands in the picture. Who Is in Frame · How Character Voices Work.

## Names the system finds on its own

Besides the stored list, some call-forms are recognized on the fly by SceneMixer and never appear in the dialog: the first name of a full name (Rafael Vidal → Rafael), a nickname in quotes inside the name (Tavian ‘Lil Tay Tay’ St. Clair III → Tay Tay), the tail after a dash (Bully Boy 1 - Arda → Arda), and the usual shortenings of a bracketed name (Ahriman (the Deceiver) → Ahriman or Ahriman the Deceiver). They follow from the card’s name; you do not write them anywhere. Same rule as the list: a match only when one character can take it.

## How to write so the right person is matched

- Let each character appear once in full, then call them the way the story does. SceneMixer’s read collects exactly what you write.
- Give every nickname one owner. If two characters share a first name or a nickname, write full names for both — with two possible answers the system matches neither rather than guess.
- Prefer an address only they answer to (Cousin) over a role word anyone could answer (the guard): the first is collected, the second is not.
- After a rename you do not have to chase the text: this project’s storyboards are rewritten to the new name, and the old name stays recognized. Renaming and unique names.
- Spelling one person two ways (a translation, an invented short form)? Settle on one: use the card’s name in the script, or rename the card and let the old form ride in the list.

## When a name goes wrong

| What you see | What it comes from | What to do |
| The line does not use the character’s voice; he stands there lips closed | The name was not matched to any card: it is not on the list and not one of the forms found on its own | Write the card’s full name in that line, or rename the card to the name the script uses; read the script again |
| Two characters are both called “Sam”, and neither is matched | One name, two possible owners | Write full names for both in the script |
| A plain word was taken as a character | A one-owner match from a form derived from a name — an English card like “Doorman Ray” also answers to “Doorman” | Avoid the bare word for other people, or rename the card so the word is no longer part of its name |
| A card has no grey line although the script uses a nickname | The read collected nothing for that card | Check the nickname is written in the script itself, not only in your notes; read again |
| An unwanted entry cannot be removed | The list is maintained for you and is read-only | Keep the word for this character only, or rename the card; there is no third lever |

## FAQ

### Why do some cards show the line and others not?

The line on SceneMixer appears only when the script calls that person, place or prop something besides its full name — or after a rename. A card you added by hand starts without one.

### Can I type a name into the list myself?

No; the list is read-only in SceneMixer. Add a name by writing it in the script and reading the script again, or by renaming the card — the previous name becomes an entry by itself.

### I renamed a character. Do my existing storyboards still work?

Yes. SceneMixer rewrites references in this project’s storyboards to the new name, and the old name stays on the card as an entry — anything that still says it keeps matching.

### Why is “Rafael” accepted when the card only says “Rafael Vidal”?

Shortenings derived from the name itself — the first name, a nickname in quotes, the tail after a dash, the usual forms of a bracketed name — are recognized by SceneMixer on the fly and are not listed in the dialog.

### Do the other names appear in the video?

No. The clip on SceneMixer speaks the words written in the shot; the list only decides who a line belongs to and who stands in the picture.

### Two characters are both called Sam. Which one answers?

Neither: with two possible owners SceneMixer does not guess. Write full names in the script.

### Do scenes and props use their other names the same way?

SceneMixer shows them on the card and passes them to the storyboard model as a hint, but resolving a name to a card is done for characters only. For a prop, being mentioned under any of its names also marks it as used in that episode.

### How many names can a card keep?

Up to eight on SceneMixer. Reading the script again merges new names in rather than replacing the list.

### Are the names translated in other display languages?

SceneMixer protects them as names: when text is translated for display, the speaker is kept as written, old name or nickname included.

Updated November 3, 2026


---

# Amber Notes After You Save a Storyboard Edit: What They Mean

URL: https://scenemixer.com/guides/amber-notes-after-you-save-a-storyboard-edit/
After a hand edit, an amber note may appear above the storyboard: a name written without @, or a 'do not show someone' sentence. What each means and how to fix it.

# The Amber Notes After You Save a Storyboard Edit: A Name Without @, and “Do Not Show” Sentences

Quick answer: After you save a hand edit, the editor reads the parts you changed and may show an amber note above the storyboard. It can say two things. Written as a plain name without an @ tag: a person in the shot is typed, not linked, so their sheet will not be used and the model may draw someone else — insert them with @. This reads like a “do not show…” instruction about people: a person named in a negative sentence tends to end up in the frame — describe who is in the frame instead. The note never changes your text, never blocks the save and costs nothing. Close it with ×.

Edit a shot, type a character’s name without @, save, and watch for the amber note above the text.

## Where the note appears

Above the storyboard text of the segment you just saved, on a computer and on a phone. It appears only after a hand edit; segments as the storyboard produced them do not get one. It is a suggestion: the save already went through, and nothing is regenerated. Close it with × once you have read it.

## A name without @

The full message, with the name it found first: Jules: written as a plain name without an @ tag. If this person is in the frame in this shot, insert them with @ (the @ button above the editor); without @ no reference image is attached and the model may draw someone else.

Fix it if the person is on screen in that shot: delete the name, type @ and pick them. Leave it if they are only talked about, or heard off screen. Shots with no linked characters.

## A “do not show” sentence

The full message: This reads like a “do not show…” instruction about people. Video models do not follow negative sentences: a person named in “do not…” tends to end up in the frame. Describe who is in the frame instead. The sentence it means is quoted under it.

Rewrite it as what the camera sees. “Do not show Jules” becomes “Only @Mara in frame, alone at the window”. “No one else in the room” becomes “the empty room behind her”. Say what is there, not what is not.

## When no note appears

No note means none of the two was found in your changes, or the check was skipped because you saved many edits in a short time. It is not a guarantee that every name is linked. Before generating, SceneMixer still asks if some shots have no linked characters.

| The note says | Means | Do |
| …written as a plain name without an @ tag | A person in the shot is typed, not linked | Insert them with @ if they are on screen |
| This reads like a “do not show…” instruction | A negative sentence names a person | Rewrite as who is in the frame |
| (no note) | Nothing found in your changes | Nothing to do; generate when ready |

## FAQ

### Does the note change my text?

No. It only points things out. SceneMixer saves your text exactly as you wrote it.

### Does it cost credits?

No.

### Can I ignore it?

Yes. Close it with ×. It is advice, and some cases, like a person who is only talked about, are fine as they are.

### Why do negative sentences backfire?

Video models treat every person named as someone to draw. Naming who should not appear puts them on the list.

Updated November 3, 2026


---

# Are AI-Generated Movies Allowed on YouTube? What YouTube Says

URL: https://scenemixer.com/guides/are-ai-generated-movies-allowed-on-youtube/
What YouTube's own help pages say about AI-generated videos: the disclosure rule, its examples, labels and the monetization policy, quoted with links and a check date.

# Are AI-generated movies allowed on YouTube?

Quick answer: This page quotes YouTube and decides nothing. YouTube's disclosure page says: “We require creators to disclose AI-generated or altered content that seems realistic.” It lists what needs disclosing and what does not, says disclosing alone won't limit a video's audience, and keeps its Community Guidelines and monetization policies in force for all content. Below: the wording, where the setting is, and what a SceneMixer file carries into YouTube Studio. All YouTube pages checked November 2, 2026; YouTube edits them, so open the links before you publish.

Sources: Disclosing use of GenAI content · YouTube channel monetization policies.

## What YouTube asks creators to disclose

YouTube's page opens by saying it encourages creators' “innovative and responsible use” of editing and generation tools, and that viewers want to know if what they watch is real. It requires disclosure of realistic AI-generated or altered content, and lists four cases:

- content that makes a real person appear to say or do something they didn't;
- content that alters footage of a real event or place;
- content that generates a realistic scene that didn't actually occur;
- music that is the main focus of the video.
Source: YouTube Help, Disclosing use of GenAI content (checked November 2, 2026).

## What YouTube says does not need disclosure

The same page says non-realistic content made with AI, and minor edits to realistic content, don't need it. Its own examples include “Someone riding a unicorn through a fantastical world” and an AI-generated or altered animation inside a fully animated video. It also lists production assistance — generative AI used for a video outline, script, thumbnail, title or infographic — as not needing disclosure. The list is marked as not exhaustive.

Because the line YouTube draws is realistic versus not, the look you pick in step 1 is one of the facts that matter. SceneMixer's looks run from realistic ones (Digital HD, Shot on Film) to drawn ones (Painterly Anime, Watercolor Storybook, Family 3D Animation, Claymation). We do not tell you which side a given video falls on; that is yours to decide against YouTube's page. Every visual style.

## Where the setting is, and what viewers see

In YouTube Studio, upload, then the Attributes section, then AI use: choose Yes if the content meets the disclosure requirements. YouTube says that for photorealistic or music content a label may appear in the video player, and that for non-photorealistic or animated content it may appear in the expanded description. YouTube may also apply a label itself, for content made with its own GenAI tools, content carrying C2PA metadata, or content its systems detect as AI-generated or altered.

On the cost of disclosing, the page says: “Disclosing AI content alone won't limit a video's audience” or affect eligibility to earn money. On not disclosing, it says creators who consistently choose not to may have a label applied by YouTube, or face penalties up to removal of content and suspension from the YouTube Partner Program. And it says Community Guidelines apply to all content, including AI-generated or altered content.

## Monetization is a separate document

YouTube's channel monetization policies say that if you're making money on YouTube, your content should be original and authentic, and should “not be mass-produced, generic, repetitive, or manipulative.” A July 15, 2025 note says the “repetitious content” policy was renamed “inauthentic content”. Under generic or repetitive content, YouTube lists as allowed “a series following a set of characters across episodes” where each video has a distinct storyline, and “using AI to visualize a unique character and narrative you invented”. It lists as not allowed AI-generated content made with “generic or unoriginal templates giving the impression of mass production”. Source: YouTube Help, YouTube channel monetization policies (checked November 2, 2026). Do AI movies make money? · Can you monetize AI videos on YouTube and TikTok?.

## What a SceneMixer file brings to the upload

- Download is the full-quality file. Upload that, not the lighter copy the site streams. Original and play versions.
- The corner badge. Free accounts' videos carry a small SceneMixer AIGC badge in the bottom-right corner; on a paid membership it is off by default and you can switch it back on; on a site with a labelling requirement it cannot be turned off. The AI badge switch.
- Shared links always carry it. A video shared through a SceneMixer link keeps the badge, members included; your own download follows your setting. Why shared videos carry the badge.
- It is not YouTube's setting. The badge is a mark in the picture; AI use is a choice you make in YouTube Studio. They are separate steps.
The Terms say: “If your output is delivered without the visible label, you are responsible for any AI disclosure required where you publish it, and you must follow the AI-content rules of the platforms you publish on.” (Terms, section 06.) How to post: Posting to TikTok, YouTube Shorts and Reels.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Campus Youth; Costume Style Streetwear, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the cast once, then choose the look knowing YouTube's line between realistic and not. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Question | What YouTube's page says (short) | Where |
| Do I have to disclose? | Realistic AI-generated or altered content: yes | Disclosing use of GenAI content |
| Does an animation need it? | Non-realistic content, per its examples: no | Disclosing use of GenAI content |
| Where do I set it? | YouTube Studio, upload, Attributes, AI use | Disclosing use of GenAI content |
| Does disclosing cut my reach or earnings? | “Disclosing AI content alone won't limit a video's audience” | Disclosing use of GenAI content |
| Can YouTube label it for me? | Yes: its tools, C2PA metadata, its own detection | Disclosing use of GenAI content |
| Can I earn from AI videos? | Original, not mass-produced or templated; see its examples | Channel monetization policies |

## FAQ

### Are AI-generated videos banned on YouTube?

YouTube's disclosure page does not ban them: it asks for disclosure of realistic AI content and applies its Community Guidelines to all content. Read the page for the exact wording.

### Do I have to label my AI movie?

YouTube requires disclosure of realistic AI-generated or altered content in the cases it lists, and says non-realistic content and minor edits don't need it. Which side your film is on is yours to decide against its page.

### Does labelling hurt my reach or earnings?

YouTube's page says disclosing AI content alone won't limit a video's audience or its eligibility to earn money.

### Does YouTube add the label by itself?

It says it may, for content made with its GenAI tools, content carrying C2PA metadata and content its systems detect as AI-generated or altered.

### Does the SceneMixer badge replace YouTube's AI use setting?

YouTube does not say so. The badge is a mark in the picture; the AI use setting is a step in YouTube Studio.

### Can I post an AI series on a faceless channel?

YouTube's monetization policies list a series following a set of characters, each video with a distinct storyline, as allowed, and templated mass production as not. A faceless channel.

### What about YouTube Shorts?

YouTube's monetization page says the Shorts monetization policies also apply if you monetize Shorts. Posting to Shorts.

### Is this a ruling on my video?

No. It quotes YouTube's pages with a check date; they change, and they are the authority.

Updated November 2, 2026


---

# Ask SceneMixer: The AI Help Window, Its Limits, and Official Replies

URL: https://scenemixer.com/guides/ask-scenemixer-ai-help-and-official-replies/
The Ask window answers questions about using SceneMixer from the site's guides, with links. What it answers, its limits, and how unanswered questions get a reply.

# Ask SceneMixer: What the Help Window Answers, and When a Person Replies

Quick answer: Ask SceneMixer is the help window behind the Ask AI capsule in the bottom-right corner, behind the speech-bubble icon in the top bar, and behind the Ask AI what to do link that appears next to an error message. It answers questions about using the site — what something costs, what a step does, why a button is grey — from these guides, and links the guide it used. When a question is about your own account — where your credits went, why a segment failed, an order — it can also look that up, read-only. It only knows SceneMixer and says so when it does not know. If it cannot help, Still not solved? → Send feedback hands the conversation to a person; and questions it could not answer may get an Official Reply under the bell.

Ask the way you would ask a colleague: 'How many credits does one episode take?' or 'Why is Generate greyed out?'

## Where to find it

Once you are signed in, the help window is one tap away in four places. They all open the same window, so use whichever is nearest to you.

| Where you see it | What it looks like | When it is there |
| The capsule in the corner | A white pill with two purple sparkles and the words Ask AI, bottom-right | On every page you are signed in to, except the sign-in and sign-up pages and the episode editor — the same on a computer and on a phone |
| The top-bar icon | A speech bubble, next to the bell | On the top bar of a page. On a phone it is not shown; inside the episode editor it is the entry to use |
| Next to an error | A small Ask AI what to do link | Under a message the editor could not finish: a clip that failed to generate, or a composite that failed |
| The account menu / side menu | A row labelled Ask AI | From the round avatar in the top bar on a computer, or from the side menu on a phone |

### The capsule in the corner

The capsule is the everyday entry. It is a white pill with two purple sparkles and the words Ask AI, in the bottom-right corner, and it looks the same on a computer and on a phone. On pages that carry a bar along the bottom — the project wizard and the ad progress page — it sits above that bar instead of covering it, so it never sits on top of your Next button. While the window is open the capsule steps aside: on a computer the window opens directly above where the capsule was, and on a phone the window takes over the screen.

### The greeting, once

The first time you meet it on a device, a couple of seconds after the capsule appears, the words change from Ask AI to a greeting — Hi! Got a question? Ask me — and the capsule grows a little to fit. It stays that way for about nine seconds and then shrinks back to Ask AI; the sparkles keep twinkling the whole time, a little faster while your pointer is on it. It is a greeting, not a popup: nothing is waiting for you, nothing is blocked, and you can carry on working.

It shows once per account, per device — the memory lives in that browser on that computer, not on your account, so signing into a second account on the same machine gives that account its own greeting, and using a different computer gives you one there too. If your browser blocks site data, the greeting simply never appears; everything else about the window works the same. It also holds back whenever something more important is on the screen: the first-run tour, any dialog (top-up, plan, share reward, a confirmation), or a progress pill parked in the same corner. Once those clear, the greeting appears if this device has not shown it to you yet. It never interrupts anything.

### Opening and closing the window

On a computer the window is a card in the corner rather than a full-screen takeover, and the page behind it stays usable — you can follow a guide it links to and come back to the conversation. Close it with the small × in its header or with the Esc key; clicking the page behind does not close it, which is deliberate so that a stray click cannot throw away a half-typed question. On a phone it opens as a full-screen sheet with the same header and the same ×.

### Next to an error message

The moment you are most likely to want to ask something is right after something fails, so the editor puts an entry next to the failure. Under the message that explains why a clip would not generate, or why a composite failed, there is a small Ask AI what to do link. Click it and the window opens with the question already written for you: I got this message: “…”, what does it mean, and what should I do?

It fills the sentence in and stops there. Nothing is sent for you and none of your question allowance is spent — you can add what you were actually doing, or delete the sentence and ask it your own way, then press Send. What it carries across is the same explanation you can read on screen, in plain words, rather than any raw internal text. The link shows in your own projects; in a project you can only watch, it is left out.

### When the menu says Help instead of Ask AI

The entries exist for signed-in accounts. If you are not signed in there is no help entry at all — not in the corner, not in the menu. If the help window is temporarily unavailable, the corner capsule does not appear at all, and neither do the Ask AI what to do links next to error messages. What stays is the top-bar bubble and the row in the account menu (the side menu on a phone); both are labelled Help while this lasts, and they open the feedback form instead of the question window, so the name always tells you what pressing it will do. One more case: when a reply to your feedback is waiting, the top-bar icon carries a red number; pressing it then opens your feedback history, which is where the reply is, rather than the question window.

## What it answers well

- What things cost, how credits and plans work — SceneMixer reads prices live from the site, never guesses them.
- What each step, button and status means.
- What to do about a message you just saw.
- Where something is.
- Your own account: where your credits went, why a segment failed, whether credits came back, your orders and plan — it reads your own records to answer, read-only. What Ask AI can check in your account.

Each answer lists the guides it came from under From the guides, with Read the guide. Common questions are one tap at the start. If your interface is in another language, you get the answer and the guide titles in that language.

## What it will not do

- Answer anything outside SceneMixer: I only know SceneMixer — how it works, what it costs, what to do next.
- Guess: when the guides do not cover something it says Our guides don’t spell this one out, and I’d rather not guess.
- Change anything in your account, orders or projects. It can look them up but not act: a refund, a cancellation or a fix you do yourself or ask for through feedback. What support can see.

Please do not type personal or sensitive details into it.

## Limits

A question can be up to 500 characters (That question is too long otherwise). A conversation holds a few rounds; after that it says That’s the end of this conversation. Start over to keep asking. — press Start over. There is an hourly and a daily limit per account: You’re asking a bit fast. Try again in N min. It is free to use.

## When it does not solve it

After an answer that did not help, or after two answers, the window shows Still not solved? → Send feedback. The feedback form opens with your questions and the answers attached, so a person starts where the AI stopped. Replies arrive under the bell and in your feedback history. How to get help.

## Official replies to questions it could not answer

Questions the AI could not answer are also read by the SceneMixer team, and some get a reply even if you did not send feedback. It arrives under the bell as Your question got a reply. Open it to see Your question quoted and the Official Reply below, in your interface language, with Still not solved? Tell us more to go straight to a person. The quote of your question disappears after 30 days, when the help conversation itself is deleted. What else can arrive under the bell, and how read and unread work, is in The Notification Bell and the Announcements Page.

| You see | It means |
| From the guides · Read the guide | The sources of the answer |
| Our guides don't spell this one out… | Not covered; send feedback |
| I only know SceneMixer… | Off-topic question |
| You're asking a bit fast… | Hourly or daily limit |
| That's the end of this conversation… | Press Start over |
| Your question got a reply (bell) | A person answered an unanswered question |

## FAQ

### Is it free?

Yes.

### Can it look at my project?

Yes, read-only. Since October 2026 it can look up your own projects, recent generations, credit history and orders when your question needs it, and never anyone else’s. It cannot change anything; for a refund or a fix, send feedback.

### Is the conversation kept?

SceneMixer keeps it for 30 days, so a person can follow up; then it is deleted.

### Can I ask in my own language?

Yes. The answer comes in your interface language.

### Where exactly is the Ask AI capsule?

Bottom-right corner of the page, a white pill with two purple sparkles and the words Ask AI, the same on a computer and on a phone. It is on every page you are signed in to except the sign-in and sign-up pages and the episode editor.

### I do not see it in the episode editor. Where is it there?

Use the speech-bubble icon in the top bar, next to the bell. The editor fills the whole width of the bottom of the screen with the clip timeline and its tabs, so the corner capsule would sit on top of them.

### What is the greeting that changes the words on the capsule?

A one-time welcome, about two and a half seconds after the capsule appears, that changes the words to Hi! Got a question? Ask me for about nine seconds before they shrink back. It shows once per account per device, holds back while a tour or a dialog is on screen, and never blocks anything.

### What does Ask AI what to do do when I click it?

It opens the question window with the error message already written into the question, ready for you to send or to rewrite. It does not send anything by itself and does not use up any of your questions.

### Why does the menu say Help instead of Ask AI?

It means the question window is not available right now. The top-bar bubble and the menu row are still there, renamed Help, and they open the feedback form instead, so you are never left without a way in; the corner capsule is not shown while this lasts. Signed-out visitors do not get an entry at all.

### Does closing the window lose what I typed?

Close it with the small cross in its header or with the Esc key on a computer; clicking the page behind does not close it. Starting over with Start over clears the conversation on purpose.

Updated November 3, 2026


---

# Background People in a Scene: The One Line That Puts Them There

URL: https://scenemixer.com/guides/background-people-in-a-scene/
Whether a scene has people in the background is one line in the storyboard. Where it lives, how to write it, how to add or remove it, and what happened to crowd images.

# Background People in a Scene: The Line That Puts Them There, and How to Keep Them Consistent

Quick answer: Whether a shot has people in the background is a stated fact, not a mood: it lives in one clause at the end of the segment header’s set dressing line — background people: who, where, doing what. The storyboard writes it when the script says a public place is busy; you can add, edit or delete it. With the clause, the video gets a crowd; without it, only the named characters are asked for, though a busy public place may still show a few passers-by. The clause applies to every shot of the segment. Separate crowd images are no longer offered: the line is all you write, and it costs nothing.

Newer storyboards. In storyboards written or regenerated since late October 2026 there is no Set dressing line, so there is no Background people: clause to edit: whether a crowd is there is decided from your shots, so describe the crowd (or its absence) in the shot. Older storyboards, ads and remakes work as described below. The segment header, field by field.

Open a segment set in a street or a hall and read the set-dressing line: the clause after the semicolon is the crowd. Remove it and the price tag does not change; the shot does.

## The one line that decides it

The segment header has a set dressing line describing what is on the set. When background people are present it ends with a clause: ; background people: a dozen guests at the tables, eating and talking. That clause is the only place the video is told about extras. Shot text that mentions “the crowd” without the clause produces an empty room with a line about a crowd. All header fields: The Segment Header, Field by Field.

## How to write it

Who, where, doing what, in one breath: background people: six commuters on the platform, waiting, looking at phones. Keep them in the depth of the frame and away from the named characters’ positions; do not name any of the cast among them. Write the same clause on every segment set in the same place while the crowd is there, and drop it on the segment where the room empties; a crowd that flickers between segments is almost always a missing clause.

## Keeping the crowd consistent

The extras are generated fresh in each segment from the clause, so their number and clothes can drift between segments. Keep the clause word for word the same in every segment of the scene, and make it specific: eight wedding guests in dark suits and pastel dresses, seated at round tables drifts less than some guests.

## What happened to crowd images

Earlier, a scene that declared a crowd could get a separate crowd image, generated on request and priced like a sheet. Since late September 2026 SceneMixer's editor no longer offers one and Generate all in Step 2 no longer includes it; the clause is carried into every shot instead. Crowd images you already generated are kept in Step 2, and one you picked by hand for a segment is still used there. Costume Changes and Scene States.

## What it does not do

Extras stay extras: no lines, no close-ups, no names. A person who speaks or is looked at is a character and belongs in Step 2. A crowd cannot be placed at a specific spot by text alone — “a guard behind the counter” is a character, not a crowd. Leads, Supporting Cast and Voice-Only Characters.

## Cost

The clause itself is free; it changes what the video shows, not what it costs. Regenerating segments after adding or removing the clause is a normal regeneration.

| Segment header | What the video shows |
| Clause present, the same in every segment | Background people as described, close from segment to segment |
| Clause present, worded differently per segment | Background people that change between segments |
| Clause absent | Named characters asked for; a public place may still show a few people |

## FAQ

### Why does one segment have a crowd and the next not?

The clause is missing on the second. Copy it over.

### Can the extras include a specific character?

No. Anyone specific is a character in Step 2 and is named in the shot.

### Where is the Generate crowd image button?

It is gone. Background people now come from the clause alone, in every shot of the segment. Crowd images made earlier are kept.

### Does the clause cost credits?

No. It is text; SceneMixer prices the segment the same with or without it.

### Can I put a crowd in an interior?

Yes, if the story says the room is busy: a bar, a hall, a classroom. Write the clause the same way.

Updated November 3, 2026


---

# Become a SceneMixer Affiliate Partner and Earn Cash Commission

URL: https://scenemixer.com/guides/become-an-affiliate-partner-cash-commission/
Creators with a channel, blog or community can apply to earn cash instead of credits on people they refer. What the form asks, how review works, and the rules.

# Become an Affiliate Partner: Apply for Cash Commission

Quick answer: Everyone earns 20% in credits on referred payments. Creators with an audience can apply to earn cash instead: 15% to 25% of each payment, by audience size. Open Affiliate program (in the footer of every page), select Apply, and fill in where you promote, how many followers you have, who your audience is and a payout method. Every application is reviewed by hand and the result appears on the same page; you keep earning credits meanwhile. Once approved you get a partner dashboard with your link, your code, a cash wallet and a promo code for your audience.

Have the links to your channel or community ready before you open the form: the first field asks for them, one per line. You can fill the form in before you create an account; it stays on the page while you sign up.

## Who it is for

Creators with a channel, a blog, a newsletter or a community who will bring more than a few sign-ups. If you only want to pass your link to friends, you do not need to apply: Invite friends in the top navigation works for every account from the start. How the affiliate program works.

## What the form asks

- Where you promote — links to your channel, blog or community, one per line.
- Followers or subscribers — pick the tier that matches your biggest platform. Your rate follows this tier, and we check it against your channels.
- Who your audience is and your country or region.
- How you will promote (optional) — tutorial videos, reviews, community posts and so on.
- Payout method — PayPal or Wise, with the email on that account.
- Partner rules — tick I have read and agree, then Submit.

| Followers or subscribers | Cash commission |
| Under 10,000 | 15% of each payment |
| 10,000 to 100,000 | 20% of each payment |
| Over 100,000 | 25% of each payment |

## The partner rules

- Commission is based on what the referred person actually pays, and is taken back if they refund or charge back.
- A commission becomes withdrawable 30 days after the payment; SceneMixer pays out once a month from $50.
- No self-referrals, no search ads on our brand name, no spam and no misleading claims.
- Tell your audience that you earn a commission.
- If the rules are broken, commissions may be voided and partner status suspended.

## Review and its results

- Application under review — checked by a person; the result will show on this page. You keep earning credits in the meantime.
- Your partner application was not approved — you stay on credit commission and can Apply again later.
- Approved — the Affiliate program page becomes your partner dashboard (also under Partner dashboard in your account menu), with your rate, your cash wallet and the Paid in choice.
- Partner status suspended — after a rule was broken; commissions in question may be voided.

## After approval: a promo code for your audience

On your partner dashboard, select Create promo code under Promo code for your audience. It uses the same name as your referral code and gives your viewers 300 bonus credits on a SceneMixer credit top-up of $1.49 or more, once per person. Your viewers type the code when they create an account, which binds them to you, and again into Promo code (optional) when they buy credits. You can create one.

## After approval: cash or credits

Under Paid in, pick Cash at your rate or Credits 20%. The choice applies to new orders only; commissions already recorded keep their kind. The window is the same either way: payments made in the 12 months after a person signs up count. How cash payouts work.

| | Everyone | Approved partners |
| Paid in | Credits | Cash or credits, your choice |
| Rate | 20% | 15%–25% cash by audience size, or 20% credits |
| When | Credits as soon as the payment clears | Cash withdrawable after 30 days |
| How to join | Nothing to do | Apply on the Affiliate program page |

## FAQ

### Is there a minimum audience?

No fixed number. Pick the tier that matches your biggest platform and say honestly how you reach people; SceneMixer judges applications by hand, and the tier sets your cash rate.

### Can I switch back to credits?

Yes, under Paid in, for new orders.

### Can I change my payout method later?

Yes, on the cash commission card of your partner dashboard. A payout already in progress is not affected.

### Do the people I already referred count after approval?

Yes. Everyone bound to your link stays yours; their new orders pay you in whatever you choose under Paid in.

Updated November 3, 2026


---

# Best Resolution for YouTube Shorts: 9:16, Pixels and Aspect Ratio

URL: https://scenemixer.com/guides/best-resolution-for-youtube-shorts/
What YouTube's own help pages say about Shorts (square or vertical, up to 3 minutes), the 9:16 pixel sizes that follow, an aspect-ratio table and which tier to render.

# Best resolution for YouTube Shorts

Quick answer: Make it 9:16 and render at the sharpest tier you are happy to pay for. YouTube's help pages (read November 2, 2026) say a Short is up to three minutes with a square or vertical frame; the resolution page lists 16:9 sizes and says the player adapts to other shapes, and neither page publishes a Shorts-specific pixel size. Turn the 1920×1080 on its side and you have 1080×1920, the vertical equivalent; the short side is the number that matters. In SceneMixer the quality is a tier (480P, 720P or 1080P), and you upload the downloaded original.

Check what you have: open the downloaded file's properties and read its width and height. For a vertical video the width is the short side, and it should match the tier you paid for.

## What YouTube's pages say (read November 2, 2026)

| Topic | The page says | Official page |
| What counts as a Short | Up to three minutes, with a square or vertical aspect ratio. | Upload YouTube Shorts |
| Resolutions | For the default 16:9 aspect ratio, encode at sizes from 426×240 up to 7680×4320, including 1280×720 and 1920×1080. The player adapts to other shapes; vertical 9:16 videos on computer browsers may get padding. | Video resolution & aspect ratios |
| Padding | Avoid adding padding or black bars directly to your video; it interferes with the player adapting to the video's size. | Video resolution & aspect ratios |
| Encoding | MP4 container, H.264 video, AAC-LC or Opus audio; upload at the frame rate it was recorded. | Recommended upload encoding settings |
That is everything the official pages commit to. Third-party lists give a single “Shorts size”, usually 1080×1920; it is the sensible vertical match for a 1920×1080 master, but it is not a rule on the pages above, so this page calls it arithmetic, not a requirement.

## 9:16 and 16:9: the arithmetic

| Short side | 9:16 (vertical) | 16:9 (landscape) |
| 480 | 480×854 | 854×480 |
| 720 | 720×1280 | 1280×720 |
| 1080 | 1080×1920 | 1920×1080 |
| 1440 | 1440×2560 | 2560×1440 |
The rule behind the table: for 9:16 the height is the width times 16 over 9; for 16:9 the width is the height times 16 over 9. Square (1:1) is the same on both sides; 4:5 is 1.25 times as tall as wide. SceneMixer makes 16:9 and 9:16, not 1:1 or 4:5, so a square Short would have to be cropped from a 9:16 master in another tool. The sizes in the first two columns match YouTube's own list when read sideways (1280×720, 1920×1080 and 854×480 are on it).

## Which SceneMixer tier for a Short

Resolution is the tier, not a separate setting, and the picker shows each tier's resolution beside its price. The default is deliberately a low one:

| Tier | Use it for | Per second |
| Wan 3.0 · 480P | Drafts: does the shot read at all? | 10 credits |
| Wan 3.0 · 720P | Finals for phones | 19 credits |
| Wan 3.0 · 1080P | Finals you want sharp on a big screen too | 38 credits |
A Short is watched on a phone, which is why 720 is the usual finishing tier for phones and 1080 is for the cases where the same file will also play on a bigger screen. Draft every segment at the cheap tier, fix the storyboard, then regenerate only the keepers at the finishing tier. Which tier for what · Draft cheap, finish sharp · 480, 720 or 1080.

## Three habits that protect the sharpness

- Finish every segment on the same tier. When all the segments come from one tier, joining them copies the streams; mix tiers and SceneMixer’s join re-encodes everything onto one canvas, which costs a generation of quality. Compositing and sharing.
- Upload the download, not the player. The player streams a lighter copy; Download is the full-quality original. A clip that looks soft in the browser may be fine on disk. Original or play version.
- Do not add bars. A 16:9 project placed on a vertical canvas has black bars baked in, which YouTube's page tells you to avoid. Choose 9:16 in Step 1 instead; it cannot be switched later, so start a new project with the same script if you need the other shape; the cast can come across from the asset library.

## 9:16 or 16:9?

| Where it will be watched | Choose | Why |
| YouTube Shorts, TikTok, Reels | 9:16 | Square or vertical for Shorts; vertical for the other feeds |
| A YouTube video on a computer or TV | 16:9 | The standard shape on a computer |
| Both | Two projects | The frame is fixed after Step 1; same script, same cast |
Vertical or landscape · The editor in vertical and landscape · Vertical video AI generator.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the characters at their normal size; the character sheets are always 9:16 whatever the project is. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Question | Answer | Source |
| Longest Short | 3 minutes | YouTube help, November 2, 2026 |
| Shape of a Short | Square or vertical | YouTube help, November 2, 2026 |
| Vertical Full-HD size by arithmetic | 1080×1920 | Not a published rule |
| Tier to draft on | Wan 3.0 · 480P | SceneMixer picker |
| Tier to finish on for phones | Wan 3.0 · 720P | SceneMixer picker |

## FAQ

### What is the best resolution for YouTube Shorts?

Vertical 9:16, at the highest quality you will pay for. YouTube's pages do not publish a Shorts-specific pixel size; 1080×1920 is the vertical equivalent of its 1920×1080 entry.

### What aspect ratio is a YouTube Short?

Square or vertical, up to three minutes long, according to YouTube's help pages read on November 2, 2026. SceneMixer makes 9:16 vertical.

### Can I upload a 16:9 video as a Short?

YouTube's page says a Short has a square or vertical frame, so a 16:9 landscape video is not categorised as one.

### What is 9:16 in pixels?

Width times 16 over 9 gives the height: 1080 wide is 1920 tall, 720 wide is 1280 tall, 480 wide is 854 tall.

### Does SceneMixer make square videos?

No. Projects are 16:9 or 9:16. A square version would be cropped from a 9:16 file in another tool.

### Which SceneMixer tier should I use for Shorts?

Draft on the default tier, then finish on the 720P tier for phones or the 1080P tier for a sharper file.

### Why does my video look soft in the browser?

The player streams a lighter copy; the download is the full-quality original.

### Can I change a project from 16:9 to 9:16?

No. The frame is fixed after Step 1; start a new project with the same script and bring the cast across from the asset library.

### Does a higher resolution cost more?

Yes, by the second: the 480P tier on SceneMixer is 10 credits a second, 720P is 19 and 1080P is 38.

Updated November 3, 2026


---

# Blocked by Moderation: What to Change, and the Pause | SceneMixer

URL: https://scenemixer.com/guides/blocked-by-moderation/
Two messages mean a clip was refused: content safety, or an IP check. What each looks at, what to rewrite, what happens to credits, and why repeat blocks pause you.

# Blocked by Moderation: What to Change

Quick answer: Two failure messages mean the clip was refused rather than lost. Blocked by content safety rules is about what the shot shows or says: gore, sexual content, sensitive figures, slogans. The provider’s IP check flagged … a known franchise or brand is about likeness: a character, logo or style that looks like someone else’s property. In both cases change the text and generate again; the message tells you what happened to the credits. Trigger blocks repeatedly in a short time and video generation is paused for a while, with a message saying so.

The compliance checklist lists what to keep out of a script in the first place.

## The two messages

Content safety. Blocked by content safety rules — edit the script or description and retry. The rules look at violence shown in detail, sexual content, real public figures, political slogans and the like — in the shot text, in a character or scene description, or in the resulting picture.

IP check. The provider’s IP check flagged the prompt, reference images or output as resembling a known franchise or brand. Change the visual style or character design and retry, or switch the video model in Settings. This is about resemblance: a character that looks like a famous one, a logo, a costume everyone recognises.

## What to change for a content block

Show the moment before or after, not the act: the hand reaching for the knife, the empty corridor after. Keep injuries described as state, not process. Cut real names of living public figures and real political slogans. If the block sits on a character or scene description rather than a shot, edit that description in step 2 and regenerate its image. Then regenerate the segment.

## What to change for an IP block

Rename and redesign what resembles someone else’s property: a character who looks like a franchise hero, a car with a visible marque, a uniform from a well-known series. Take brand names and logos out of descriptions. If the visual style itself is the problem — a style named after a studio — pick another preset in step 1. The message also offers switching the video model in Settings; different tiers apply different checks.

## When only the sound or the image is refused

Two shorter messages mean the check refused one part of a finished result rather than your text. The generated soundtrack was rejected by the provider’s content check. You weren’t charged — retrying usually passes; if it keeps failing, change the dialogue. The generated image was rejected by the provider’s content check. You weren’t charged — change the description or the visual style and retry. Retry the soundtrack case as it is first; for the image case, rewrite the description before retrying.

## The credits

A block is a refusal, not a delivery. The failure message says what happened to the credits for that attempt; in most cases they are returned automatically, and if they are on hold the message says so. Failure messages, one by one.

## The pause

If clips are blocked repeatedly within a day, video generation is paused for a while and generating shows Content moderation was triggered repeatedly in the last 24 hours; video generation is paused for 1 hour. It clears on its own. Use the time to fix the text rather than resubmitting the same shot. On SceneMixer, the pause covers video generation on your account, including ads.

## Ads

An ad whose segments were blocked cannot simply be retried: edit the ad script first, then retry. The ad progress page.

| Block | Looks at | Change |
| Content safety | What is shown or said: violence in detail, sexual content, real figures, slogans | The shot text or the description; show before/after, not the act |
| IP check | Resemblance to a known character, brand, logo or named style | Rename, redesign, remove logos, pick another style preset |
| Soundtrack rejected | The generated sound of a clip | Retry; change the dialogue if it repeats. Not charged |
| Image rejected | A generated sheet or picture | Change the description or visual style. Not charged |
| Repeated blocks | Several blocks within a day | Wait for the pause to clear; fix the text meanwhile |

## FAQ

### Was I charged for a blocked clip?

The failure message says. In most cases SceneMixer returns the credits for that attempt automatically.

### The shot is not violent. Why was it blocked?

Descriptions count too: a character or scene description with a sensitive phrase can trigger the block. Check the entries named in the segment.

### What counts as an IP problem?

Anything that resembles someone else's property: a famous character's look, a logo, a recognisable uniform or car, a style named after a studio.

### How long is the pause?

The message says. It clears on its own; do not resubmit the same shot to test it.

### Can I switch the model to get around a block?

The IP message offers switching the video model in Settings because tiers apply different checks. Content safety blocks should be fixed in the text.

Updated November 3, 2026


---

# “Building the World”: Why a Storyboard Waits Before It Starts

URL: https://scenemixer.com/guides/building-the-world-before-a-storyboard-starts/
Why Parse Storyboard can wait minutes before any text appears: the four steps (scene image, 360° panorama, measuring the set, placing the cast) and how to shorten it.

# “Building the world”: what the four steps on the episode card are

Quick answer: When you press Parse Storyboard on an episode, the storyboard is not written until the scenes it uses are ready. While it waits, the episode card on step 3 shows Building the world with four blocks, filling in order: Scene image, 360° panorama, Measuring the set, Placing the cast. It is usually two to three minutes, the longest it will wait is about eight, and waiting costs nothing extra. When the last block is done the card changes to the usual writing progress.

The other waits on this card: the wait messages · closing the page is fine: does generation continue?

## Why it waits at all

Before the storyboard is written, SceneMixer wants to know the places it will be set in: what the room looks like, how big it is, where the furniture is, and where each person would be standing or sitting. With that known, a shot that says over her shoulder, with the door behind him can be framed from the room instead of guessed, and people do not swap sides between cuts. That work happens once per scene, so the first storyboard in a new project waits longest and later ones mostly find the rooms already built. What a world model is.

## The four steps

| Block | What is happening | What helps |
| Scene image | The picture of the place is still being drawn, or is being looked at to work out what kind of space it is (indoors or out, an enclosed room or open ground). | Generate the scene images in step 2 before you parse; this block is then already done. |
| 360° panorama | The place is turned from one picture into a room you can look around. | Nothing; this is the longest block. |
| Measuring the set | The room’s size and its furniture are worked out so people can be put in it. | Nothing. |
| Placing the cast | Who starts where, so the first shot already has everyone in the right place. | Nothing; this is the shortest block. |

The blocks are different widths because the steps take different lengths of time. The line below them names the current step, shows the time so far and what is usual, for example Building the world · 360° panorama · 1:12 · usually about 2 minutes. If it passes the usual time the line changes to say it is taking longer than usual; it has not failed.

## When you see only “Preparing scenes”

Not every scene is turned into a 360° room — for example on an account that has not paid yet, or for some enclosed spaces. When none of the scenes in the episode needs one, the card shows a single bar: Preparing scenes and the time so far. The storyboard is still written against the scenes you have; it just has no room to measure. Looking around the set.

## How long, and what happens at the end

Over the last weeks the middle wait on SceneMixer has been around two to three minutes, with the slowest tenth near eight. Eight minutes is as long as the card waits. If a scene’s room is not ready by then, the storyboard is written anyway, without that room: shots set there still work, but the framing has less to go on. If a room could not be built at all, the site tries again the next time you press Parse Storyboard for an episode that uses it, after about an hour.

Nothing about the wait is charged. The price is the storyboard’s price on the button; if the storyboard fails, it is refunded as for any failure. Generate the storyboard for an episode.

## How to make it shorter

- Generate the scene images first. In step 2, open the Scenes tab and generate the images before you go to step 3. The first block is then finished before you start. Generate all.
- Parse the episode that uses the most scenes first; the rooms it builds are reused by every episode after it.
- Do not re-parse for nothing. Regenerating a storyboard reuses the rooms, but it throws the segments away. What you lose.
- Leave the page. The work is on the site’s side. Come back later and the card has moved on.

## When it looks stuck

A block that does not move for several minutes is normal during the panorama. If the card says Building the world for much longer than eight minutes, refresh once; if it is still there, the storyboard task has probably already moved on and the card has not caught up. How to get help has what to send us.

## FAQ

### Why does Parse Storyboard take minutes before any text appears?

The storyboard waits for the scenes it uses to be ready: scene image, 360° panorama, measuring the set, placing the cast. The card shows which step it is on.

### How long does Building the world take?

Usually two to three minutes on SceneMixer. The card waits for at most about eight minutes; after that the storyboard is written without a room that is still not ready.

### Am I charged for the wait?

No. You pay the storyboard price on the button, and SceneMixer refunds a failed storyboard as usual.

### What is Preparing scenes?

The single-bar version of the wait, shown when the scenes in the episode are not being turned into 360° rooms, for example on an account that has not paid.

### Can I close the page?

Yes. The work continues on the site's side.

### Can I make it faster?

Generate the scene images in step 2 first, and parse the episode with the most scenes before the others.

### Is it slower for the first episode?

Usually, because the rooms are built for the first time. Later episodes reuse them.

### What if it never finishes?

After about eight minutes the storyboard is written anyway. If the card is still on Building the world much later, refresh; if it stays, contact support with the project and episode numbers.

Updated November 3, 2026


---

# Bullet Time AI: What You Can Write, and What Cannot Be Done

URL: https://scenemixer.com/guides/bullet-time-ai/
Bullet time in AI video is a look you write, not a mode: a hit, a slowed arc around the subject, a release. How to write it, plus slow motion, time-lapse and one-take.

# Bullet time AI

Quick answer: There is no bullet-time effect to switch on. Real bullet time needs a ring of cameras around a frozen subject; an AI video has one virtual camera, so what you can get is the look: time slowed almost to a stop while the camera arcs around the subject. Write it as three short shots instead of one clever line: the hit at normal speed, the slowed arc around the subject with one movement, then the release. The same approach covers slow motion, time-lapse and one-take, with one clear limit: the editor does not retime a clip, so any change of speed has to be in the shot text and the cuts.

Three shots to try: [Shot 1] 2s: Medium shot. @Kade swings and the punch lands on @Brook's jaw. Ambient: a dull crack. [Shot 2] 3s: Low angle, the camera arcs slowly around @Brook in mid-air, slow motion, rain drops hanging around him. No dialogue. [Shot 3] 2s: Wide shot. @Brook hits the wet pavement; the rain falls at normal speed again.

## What bullet time is, and why a model cannot switch it on

Bullet time is the look made famous by action films in the late 1990s: a moment hangs almost motionless while the camera sweeps around it. On a real set it is done with a ring of still cameras firing in sequence around the subject, so the “camera move” is really a row of different cameras at one instant. An AI video has no ring, no array and no frozen 3D capture to fly through. It has one virtual camera and the words of the shot.

So bullet time becomes a look you describe: slowed time, a camera arc, things hanging in the air. That is achievable as a shot; the geometry of a true rig is not, and nothing here promises a full 360 degrees or a matched freeze. How camera moves are written.

## Write it as three shots

The storyboard on SceneMixer takes one framing and one movement per shot, and a move such as “push in while orbiting” reads as two behaviours asked to be one. Bullet time has three beats with different jobs, so give each its own shot:

| Beat | Shot | Notes |
| The hit | [Shot 1] 2s: Medium shot. @Kade swings and the punch lands on @Brook's jaw. Ambient: a dull crack. | Normal speed. A blow is one of the moments the built-in impact shake is made for (it needs at least 2 seconds). |
| The hang | [Shot 2] 3s: Low angle, the camera arcs slowly around @Brook in mid-air, slow motion, rain drops hanging around him. No dialogue. | One movement (an arc). Write what hangs in the air so the slowed time has something to show. |
| The release | [Shot 3] 2s: Wide shot. @Brook hits the wet pavement; the rain falls at normal speed again. | A different framing, so the cut is read as a cut. |
Seven seconds, three cuts. The cut back to normal speed is how a speed ramp is made here. The eight built-in moves · Writing camera movement.

## Slow motion

Write what slows down and why we are watching it: slow motion, a glass tumbling toward the floor, the water rising out of it in a slow ribbon. Put the words slow motion in the shot and give it two to four seconds. Keep speech out of it: dialogue needs about three words a second in English, so a slowed shot is written as No dialogue. and the line goes in the shot before or after. Shot seconds and timing.

What you will not find on SceneMixer is a speed control. The editor trims, regenerates, re-voices and enhances a clip; it does not change a clip's speed, so a clip cannot be slowed after it is made. If a shot came out too fast, the fix is the shot text and a regeneration. Edit one shot and regenerate.

## Time-lapse

A time-lapse is also a look: time-lapse, clouds racing over the rooftops, shadows sweeping across the square. Write the change you want to see. The limit is the rule that a segment stays in one place and one timeline: a shot that jumps to another day or hour is stopped before it is generated, with no charge. For a night that falls over the same square, use two segments in the same place and set the header's Time to Dusk for one and Night for the other, or take a finished clip and change its time of day with the AI visuals edit. One place, one timeline · The segment header · Change day to night in a clip.

## One take

An unbroken shot on SceneMixer is limited by the segment: up to 15 seconds on the 15-second tiers and up to 30 on the 30-second ones. Inside one segment there are no seams; beyond it, each segment opens where the last one ended, so write the last shot of a segment to land on a position you would be happy to continue from, and keep the camera's heading and the subject's travel the same across the join. The plan for a one-take is a single sentence of path (where the camera enters, what it follows, where it stops), and then one shot per beat as a timing ruler rather than a cut. Planning a one-take · How the next segment opens · 15 or 30 seconds.

## What you cannot get, and how to check

- A true freeze with a matched 360-degree geometry.
- A speed ramp inside one clip, or slowing a clip after it is made.
- A set interval for a time-lapse, or a day-to-night change inside a single shot.
- A guarantee of how any tier draws it: none of the tier guides here publishes a test of bullet-time looks, so this page describes how to write the shots, not how they render.
The check is cheap: render the three shots as one segment on the default tier at 10 credits a second, watch it, and change the words of the shot that missed. Draft cheap, finish sharp.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Streetwear, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the two fighters and the one wet street they fight on; the rain is written into the shot, not drawn. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Effect | Write | Seconds | Can the editor do it after? |
| Bullet-time look | Hit, then a slowed camera arc, then release: three shots | 2 + 3 + 2 | No; rewrite the shot and regenerate |
| Slow motion | slow motion plus what hangs or tumbles; No dialogue. | 2–4 | No speed control |
| Time-lapse | time-lapse plus the change you want to see | 3–5 | Time of day can be changed by an AI visuals edit |
| One take | One path sentence, one shot per beat, same heading across the join | Up to the segment | Regenerate the segment that missed |

## FAQ

### Can AI video do bullet time?

As a look, yes: a slowed moment with the camera arcing around the subject. It is written into three short shots. It is not a rig, so there is no true freeze and no guaranteed 360 degrees.

### Is there a bullet-time button?

No. SceneMixer has eight built-in camera moves and none is bullet time; the impact shake and the arc you write cover the beats around it.

### How do I write slow motion?

Put slow motion in the shot, say what hangs or tumbles, give it two to four seconds and mark it No dialogue.

### Can I slow a clip down after it is made?

No. SceneMixer’s editor trims, regenerates, re-voices and enhances; it has no speed control. Rewrite the shot and regenerate.

### How do I make a time-lapse?

Write the change you want to see in one place; for day to night, use two segments on SceneMixer with different times in the header, or change the time of day of a finished clip.

### Can I do a one-take shot?

Inside one SceneMixer segment, yes, up to 15 or 30 seconds depending on the tier; beyond that the next segment opens where the last one ended.

### Why not write the arc and the slow motion in one line?

A move that changes mid-line reads as two moves. One movement per shot and a cut between beats is what SceneMixer’s storyboard handles best.

### Does it work in vertical?

Yes. Choose 9:16 Portrait in Step 1.

### What does it cost?

Video on SceneMixer is 10 credits a second on the default tier, plus 1 credit a second to join. A seven-second test is seven seconds.

Updated November 3, 2026


---

# How to Buy Credits on SceneMixer, Step by Step | SceneMixer

URL: https://scenemixer.com/guides/buy-credits-step-by-step/
Where the Buy Credits button is, how to pick a pack, what the summary lines mean (package, bonus, discount, due today), how to pay, and when the credits land.

# Buying Credits, Step by Step

Quick answer: Click your credit balance at the top right, or Buy Credits, and the purchase window opens on the credits tab. Pick a pack — larger packs carry a bonus of up to 15% extra, and packs start from $1.49. The summary shows Package, Bonus, any Credit discount from a paid plan or a promo code, and Due today. Pay by card or wallet, with PayPal, or with a local method such as Alipay or WeChat Pay on the Airwallex tab; the form opens in the same window. When the payment is received, the window says so, shows your new balance, and the credits are usable at once.

What you get for free first; buy only when the free allowance is used up.

## Open the window

The balance at the top right of every page is a button. Click it and the panel shows Available Credits, Frozen (credits held by tasks still running), Credit History and Buy Credits. The purchase window also opens by itself when a button you press costs more than you have — with a line at the top saying how many credits short you are. When a button is greyed out.

## Pick a pack

Packs are laid out as cards; the first is selected by default and one carries a Recommended tag. Each card shows the price, the credits, and the bonus on top. On SceneMixer, bigger packs carry a bigger bonus, up to 15%. Every card has its own buy button, so choosing and buying is one click.

## Read the summary

Package: the pack’s base credits. Bonus: the extra from the pack, and, if it applies, a First top-up bonus or a limited-time bonus with a countdown. Credit discount: the percentage off that a paid plan gives on packs. Promo discount: from a code typed into Promo code (optional). Due today: what you pay. You’ll receive: base plus every bonus. Discounts stack but are capped. What a paid plan changes.

## Pay

The window has a row of tabs, each shown by its logos. Card, Apple Pay or Google Pay: the card form loads inside the window; Apple Pay and Google Pay appear when your device supports them. PayPal: sign in and pay in the PayPal window — you stay on this site. The third tab, processed by Airwallex, shows the local methods available where you are, such as Alipay or WeChat Pay: press Continue to payment, pick a method, and you can pay in your local currency (any conversion fee is charged by Airwallex). Local methods. Prices are in USD. If the form does not load, the window says so after a few seconds with a retry that reopens the same order. If the payment does not go through.

## When the credits land

Card payments land the moment the bank confirms; the window shows Payment received, the credits added, any bonus included, and your new balance, with Start creating to go on. PayPal and Airwallex occasionally take longer to clear; the window says so and the credits arrive automatically when it completes. Every top-up is a line in Credit History. Reading the history.

## Unpaid orders

Close the window before paying and the order waits under Pending Orders on the credits page for a short while; reopening it resumes the same order at the same price rather than creating a new one. Unpaid orders expire on their own and never charge anything.

| Line | Means |
| Package | The pack’s base credits |
| Bonus | Extra credits from the pack size, plus any first-top-up or limited-time bonus |
| Credit discount | Percentage off from a paid plan |
| Promo discount | Percentage off from a promo code |
| Due today | What you pay |
| You’ll receive | Base plus every bonus |

## FAQ

### Where do I buy credits?

Click your balance at the top right, then Buy Credits. The window also opens by itself when a button costs more than you have.

### How do I pay?

Card or wallet in the window, PayPal, or a local method such as Alipay or WeChat Pay on the Airwallex tab. Prices on SceneMixer are in USD; on the Airwallex tab you can pay in your local currency.

### When do the credits arrive?

Card payments land at once. PayPal and Airwallex usually at once, occasionally after the payment clears; the credits arrive automatically.

### Do bigger packs give more?

Yes. Larger packs carry a bonus of up to the percentage shown on the cards.

### Can I use a promo code?

Type it into Promo code (optional) in the summary. Its discount stacks with a plan's discount on SceneMixer, up to a cap.

Updated November 3, 2026


---

# Camera Angles and Moves for AI Video: The Eight Built-In Story Moves

URL: https://scenemixer.com/guides/camera-angles-for-ai-video/
Camera angles for AI video, plus the eight special moves built into SceneMixer: dolly zoom, eye macro, falling follow, overhead spin, roll, tilt, shake, overhead track.

# Camera angles and moves for AI video

Quick answer: Most camera work in AI video is written in words in each shot: wide, medium, close-up, push-in, pan, track. On top of that, SceneMixer has eight special camera moves it applies at matching story moments: the dolly zoom for a shock, a push-in to the pupil for a jolt of recognition, a follow down for a dropped object, an overhead spin for someone alone on the floor, a barrel roll for a spin out of control, a 90-degree tilt for a collapse, an impact shake for a blow, and an overhead track for someone running. Each has a minimum shot length.

Write one shot with an ordinary move and one with a special move, and compare: Medium shot, slow push-in on @Mara at the door versus Close-up on @Mara; the camera pushes in to an extreme close-up of one eye, the pupil reflecting the letter in her hands.

## Everyday framing and moves

| Write | It gives | Good for |
| Wide | The place and where people stand | Opening a scene |
| Medium | Two or three people and what they do | Talk and action |
| Close-up | A face or a hand | A line, a reaction |
| Over the shoulder | One person's view of another | Dialogue |
| Push-in / pull-out | Moves toward or away from the subject | Tension, release |
| Pan / track | Follows across or along | Movement through a place |

How to write one shot · How to write camera movement.

## The eight special moves

| Move | The story moment | Min. shot |
| Dolly zoom | Shock, realising the truth, freezing | 3.5 s |
| Push to the pupil (eye macro) | Seeing a screen, a photo or something that jolts | 3.5 s |
| Falling follow | Something slips from a hand and drops | 2.5 s |
| Overhead spin and rise | Someone lying on the ground, helpless | 3 s |
| Barrel roll while pushing in | Drunk, dizzy, stumbling, losing control | 3 s |
| 90-degree tilt as they fall | Fainting or collapsing (not the first shot) | 3 s |
| Impact shake | A slap, a line that lands, something heavy falling | 2 s |
| Overhead track | Walking away alone, running, being chased | 2.5 s |

## Use special moves sparingly

These moves are punctuation. One or two per episode lands; one in every shot reads as a tic. Use the dolly zoom for a revelation (dolly zoom in AI video), the impact shake for a blow, and the overhead track for a chase. The built-in moves need a single clear subject in frame and the minimum shot length above.

## Where this fits a production

The camera is written in the storyboard, which you can edit shot by shot for free. What you can change · AI fight scene generator · An action movie.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Neon Nights; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the one person the special moves are about; a single subject is what the moves need. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Move | Camera | Needs |
| Dolly zoom | Back while the lens zooms in | One subject, 3.5 s |
| Eye macro | Push-in to one eye | One face, 3.5 s |
| Falling follow | Down with a dropped object | 2.5 s |
| Overhead spin | Straight down, rotating and rising | 3 s |
| Barrel roll | Push-in while rolling a full turn | 3 s |
| 90-degree tilt | Frame turns as they fall | 3 s, not the first shot |
| Impact shake | One jolt, brief blur | 2 s |
| Overhead track | Straight down, moving with them | 2.5 s |

## FAQ

### What camera angles can I use in AI video?

Write wide, medium, close-up, over the shoulder, push-in, pan or track into each shot in plain words.

### Does SceneMixer have special camera moves?

Yes, eight: dolly zoom, push to the pupil, falling follow, overhead spin, barrel roll, 90-degree tilt, impact shake and overhead track.

### How are the special moves chosen?

The storyboard on SceneMixer applies them at story moments they suit, such as a shock, a drop or a blow.

### Is there a minimum shot length?

Yes, SceneMixer's minimum is 2 to 3.5 seconds depending on the move.

### Can I use them in every shot?

You can, but they work best once or twice per episode.

### Can I edit the camera?

Yes. The camera is part of the shot text in the storyboard, and editing it on SceneMixer is free.

### Does it work in vertical?

Yes. Choose 9:16 Portrait in Step 1.

Updated November 3, 2026


---

# Can AI Create Short Stories? Write One, Then Watch It

URL: https://scenemixer.com/guides/can-ai-create-short-stories/
Can AI create short stories? Yes: AI Create writes one from a brief as episodes, scenes and dialogue. What carries over from prose to video, and how to bring your own.

# Can AI create short stories?

Quick answer: Yes, with one thing to know. SceneMixer's AI Create writes a short story from a one- or two-sentence brief, but what it writes is a script — episodes, scenes and dialogue — because the next step is turning it into a video. It is not a prose editor. If you want a story to watch, that is the right shape; if you want a story to read, it is a first draft to rework. Either way you can paste in a story of your own, in prose, and have it read into a cast and a storyboard.

Brief for AI Create: A short story: a seamstress hems a wedding dress for a bride who never comes; on the third day a stranger arrives to collect it for someone else; quiet, a little sad, and it ends with the dress leaving the shop.

## What AI Create writes

Type the brief on the AI Create tab, choose a Length and press Start Writing. Mini (1,000) is one scene, Brief (2,500) is one or two short episodes, Short (5,000) is a handful. For a story of a few minutes, Mini or Brief is plenty. The script comes back in the language you use the site in, and every draft stays in Writing History. Writing the one-line brief · Length and cost.

## A short story is not yet a script

Prose can say things a camera cannot. Here is how the usual tricks of a short story land on screen:

| In the short story | On screen |
| What a character thinks | A line said aloud, a look, or an object they handle; or a narrator's voice-over |
| A paragraph describing the town | One place, drawn once; the place is a location the whole cast shares |
| ‘Three days later’ | A new segment; say it in the shot, and the place can change state (night, rain) |
| A metaphor | A prop that stays the same object in every shot |
| The narrator's irony | Voice-over lines: heard, while the people in frame keep their mouths closed |
How voices work · Props and locations · Lines, voice-over and captions.

## Three ways to get a story in

- Write it with SceneMixer’s AI Create. Brief in, script out.
- Paste your own. Prose is read as it is: it needs at least a paragraph (roughly 330 characters of English), and a link is not accepted, only the text. What you can feed in.
- Bring one made elsewhere. Paste it or upload a .txt, .docx or .md file, up to 300,000 characters of English per project.
AI story maker · AI story video generator · Turning a book into a movie.

## Make the draft yours

A first draft written from a brief tends to be the average of its genre; what makes it yours is the specific thing you add. After the first draft: change one name to a name you chose, replace one generic detail with a particular one (the dress's hem, not ‘a dress’), and cut any line that explains what the scene already shows. There is no setting that learns your style; the tone word in the brief and your edits carry it. Text edits are free.

## Who owns the story

That is a question of law as well as of the product, so it has its own page with the official wording and the site's terms. Can AI movies be copyrighted?.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday; Visual Style Soft Glow, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the seamstress, the stranger and the shop; a short story with one place and two people is the easiest to keep steady. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| You want | Route | Result |
| A story to watch | AI Create, then read, then render | A video of the story with a cast that holds |
| A story to read | AI Create as a first draft, then rework the text yourself | A script you rewrite into prose elsewhere |
| To film a story you wrote | Paste the prose | A cast, places and a storyboard to edit |
| Just to try | Mini length | One scene, the lowest writing price |

## FAQ

### Can AI write a short story?

Yes. In SceneMixer, AI Create writes it from a brief, as a script of episodes, scenes and dialogue.

### Can I get the story as plain prose?

SceneMixer’s AI Create writes scripts, not essays; treat the draft as a first draft and rework it. If you want to watch the story, the script is the right form.

### How long can a story be?

SceneMixer's Length runs from one scene (Mini, 1,000) to a full series (Long, 20,000). For a short story of a few minutes, Mini or Brief.

### Can I write in my own style?

There is no style-learning setting on SceneMixer. The tone in the brief and your own edits carry the voice.

### Can I paste a story I wrote?

Yes. Paste it or upload a .txt, .docx or .md file; it needs at least a paragraph, and SceneMixer accepts up to 300,000 characters of English per project.

### Can I turn the story into a video?

Yes. That is what the next steps in SceneMixer do: read it into a cast, read the storyboard, render the segments.

### What does it cost?

Each Length option shows its price before you press Start Writing; a writing that fails is refunded. New SceneMixer accounts start with 50 credits.

### Who owns a story AI wrote for me?

The site's terms say you own the results to the extent permitted by law, with no guarantee they are copyrightable. The details, with sources.

Updated November 3, 2026


---

# Can AI Make Movies? What It Does Today and Where It Stops

URL: https://scenemixer.com/guides/can-ai-make-movies/
Can AI make movies, direct them or make a TV show? Yes for short films and series: what it holds (cast, rooms, voices, score), what it can't, and the sizes in segments.

# Can AI make movies?

Quick answer: Yes. AI can make something you would recognise as a movie: a story told in scenes, with a cast that stays the same, speech with moving lips, the sound of the place and a score. What it cannot do is make ninety minutes in one go: video comes in segments of up to 15 or 30 seconds that are chained and joined, so length is a question of how many segments you pay for and review. You stay the director — you choose the type, the look, the cast, the camera language and what every shot says. Short films and series are practical now; a feature is a budget question, not a limit in the software.

Test it on one scene: A mother and her teenage son at a kitchen table the night before he leaves for university; one room, two people, ten lines, and nobody says ‘I'll miss you’.

## What making a movie means here

- Script. Written with AI or pasted from your own.
- Read. The script becomes episodes, a cast, places and props.
- Draw once. Every character, place and prop gets a reference image; every shot is made from them.
- Storyboard. Numbered shots with seconds, camera language, dialogue and ambient sound, which you can read and rewrite before anything is rendered.
- Render and join. Segments are generated, then put together; the episode is scored.
How to make a short movie with AI · What is inside a project · AI movie maker.

## What the software holds for you

| Thing | How it is held |
| Faces and clothes | Drawn once on a character sheet; every shot is generated from it, with costume changes as named states |
| Rooms | A scene image is drawn once; where the account and the space allow it, it is turned into a 360° room, measured, and the cast is placed before the storyboard is written |
| Voices | Each character has a voice; lines are spoken by the person whose lips move |
| Sound | Ambience written into each shot, and a score chosen for the episode and mixed under the voices |
| Order | The episode is a timeline of segments you can retake one at a time |
Why a series stays consistent · Building the world · The built-in score.

## What it does not do

- One unbroken take longer than a segment. The longest a segment can be on SceneMixer is 30 seconds, and only on tiers that go that far; a longer shot is two segments.
- A performance. It holds a cast and a room, not an actor's choices. You write the acting into the shot, as a tone beside the line: (low, steady).
- Readable on-screen text. Signs, menus and interface text are not drawn reliably; titles go in your edit.
- Real people and franchises. Recognisable public figures and franchise characters are refused by the IP check.
- Your own song or recording. It takes text, up to four ad photos, or a video, and no audio. What you can feed in.

## Sizes: from a trailer to a feature

| What you make | Seconds | Segments (15 s / 30 s) | Type |
| A trailer | 60 | 4 / 2 | Film, or a one-episode project |
| A short drama episode | up to 240 | 16 / 8 | Short Drama |
| A TV series episode | up to 300 | 20 / 10 | TV Series |
| A 10-minute short | 600 | 40 / 20 | Film, read in sequences |
| A 90-minute feature | 5,400 | 360 / 180 | Film, read in sequences |
Video is 10 credits a second on SceneMixer's default tier, and joining an episode is 1 credit a second, so a film costs its seconds times that rate plus the retakes you choose to make: ten minutes is 600 of those seconds, a feature is 5,400. Can AI make a full-length movie yet? · How hard is it?.

## Can it direct? Can it make a TV show?

Direct. The storyboard drafts numbered shots from your script, and you can rewrite any of them in plain words: the shot size, the move, who is in frame, how long it lasts. Eight special camera moves (a dolly zoom for a shock, a push in to an eye when someone sees something that shakes them, a 90° tilt as someone falls) are applied at matching story moments. The eight special moves.

A TV show. Set the type to TV Series: episodes of up to 300 seconds with the same cast, made in order. How to make a TV series with AI · AI TV series creator.

## A sensible first project

Start with one scene, then one episode, then the series. New accounts start with 50 credits; the first 15 setup images, the storyboard for episode 1 of your first drama and one 5-second preview are free. Video is charged from the first second, so read the storyboard before you render. What you get for free · Make a movie with AI for free.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the mother and the son and the one kitchen; a scene with two people and one room is the cheapest honest test. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Question | Short answer |
| Can AI make a movie? | Yes: short films and series today, in segments joined into one video |
| Can AI direct a movie? | You direct; the storyboard drafts the shots and you rewrite them |
| Can AI make a TV show? | Yes: type TV Series, episodes up to 300 seconds with the same cast |
| Can AI make a full-length film? | The software does not stop at 90 minutes; the cost and the review time do |
| Can it use a real actor's face? | No; recognisable people are refused |
| Can it cut to my song? | No; it takes no audio. Lay your track under the finished video in an editor |

## FAQ

### Can AI make a movie from a script?

Yes. Paste or write the script in SceneMixer, read it into a cast and places, read the storyboard, then render the segments. Script to film.

### Can AI make a movie from a prompt?

Yes: one or two sentences in SceneMixer’s AI Create write the script first. AI movie from a prompt.

### Can AI make a full-length movie?

SceneMixer will read a feature script and generate it sequence by sequence; the limits are money and the time to review every shot.

### Will the characters look the same in every shot?

Each is drawn once on SceneMixer, and every shot is generated from that drawing, so faces and clothes hold.

### Does AI make the sound too?

Yes: voices, ambience and a score come with the video. You cannot upload your own audio into SceneMixer.

### Do I need filmmaking experience?

No, but reading the storyboard before you render is the habit that saves the most.

### Can I make a movie with AI for free?

Part of it: the account allowances on SceneMixer are free, video is not. What is free.

### Who owns what I make?

That is covered with sources on its own page. Can AI movies be copyrighted?.

Updated November 3, 2026


---

# Can AI Movies Be Copyrighted? What the US Copyright Office Says

URL: https://scenemixer.com/guides/can-ai-movies-be-copyrighted/
Can AI-generated movies be copyrighted? The US Copyright Office's own words (Jan 2025), links and a check date, plus what SceneMixer's Terms say about ownership.

# Can AI movies be copyrighted?

Quick answer: This page gives no legal advice and no verdict. It quotes the US Copyright Office (checked November 2, 2026), then the site's own Terms. In the Office's January 2025 release on AI outputs: “the outputs of generative AI can be protected by copyright only where a human author has determined sufficient expressive elements.” Whether a given film meets that, and what the rule is where you live, is for the copyright office or a lawyer in your country, not for this page.

Source first: the Office's release is at copyright.gov/newsnet/2025/1060.html and the full report at copyright.gov/ai.

## What the US Copyright Office says

The Office published Part 2 of its report on copyright and AI, on the copyrightability of outputs created using generative AI, on January 29, 2025. Its own release says:

“It concludes that the outputs of generative AI can be protected by copyright only where a human author has determined sufficient expressive elements. This can include situations where a human-authored work is perceptible in an AI output, or a human makes creative arrangements or modifications of the output, but not the mere provision of prompts.”

“The Office confirms that the use of AI to assist in the process of creation or the inclusion of AI-generated material in a larger human-generated work does not bar copyrightability.”

“It also finds that the case has not been made for changes to existing law to provide additional protection for AI-generated outputs.”

Source: U.S. Copyright Office, NewsNet Issue 1060, January 29, 2025; the report itself: Part 2: Copyrightability (PDF). The Office's AI page lists Part 1 (digital replicas, July 31, 2024), Part 2 (copyrightability, January 29, 2025) and Part 3 (generative AI training, a pre-publication version released May 9, 2025), and links its registration guidance for works containing AI-generated material. Checked November 2, 2026.

## The layers of a film made here, without a verdict

The Office's sentences name things that can matter: a human-authored work perceptible in the output, human creative arrangement or modification, and prompts alone. Here is which parts of a SceneMixer project are which, as plain description. We do not say where your film falls.

| Layer | Made by |
| Your brief, or the script you pasted, and every line you change | You (a draft from AI Create is machine-written until you change it) |
| The storyboard text: shot order, shot lengths, who is in frame, what is said | Drafted by the software; rewritten by you wherever you edit |
| Which segments you keep, retake, trim and in what order | You |
| The pixels and sound of each generated segment | The generation models |
| Your uploaded photos, voice recordings and videos, where you use them | You, subject to your rights in them |
The project keeps dated records you can look back at: Writing History for AI Create drafts, a text history for storyboards, and Version History for each segment. Text history · Segment versions.

## What SceneMixer's Terms say about ownership

From the Terms of Service, section 07 (checked November 2, 2026; read the live page, which governs):

“You own the generated results to the extent permitted by law, royalty-free, including commercial use (free allowances included), subject to compliance with these terms.”

“No guarantee that outputs are copyrightable or registrable. Similar inputs may yield similar or identical outputs for other users, so outputs are not exclusive.”

“You retain rights in your uploads and warrant you may use them for generation.”

The same section says that, subject to full payment and compliance, you may use generated results commercially “to the extent permitted by law and upstream model terms”. The Terms also say not to remove or alter implicit marks, watermarks or provenance metadata (sections 06 and 07); removing the visible label is a paid option where the law permits it. The AI badge switch.

## What this page does not cover

- Other countries. The Office's report is about United States law. We have not checked, and do not summarise, any other country's rule.
- Whether your film is registrable. Read the Office's registration guidance, linked from copyright.gov/ai, before you file anything.
- Other people's rights in what you upload. Uploads need your right to use them, and every real person whose face or voice appears has to have consented; recognisable public figures and franchise characters are refused. The IP check · What you can feed in.
- Platform rules for posting or earning. Those are separate documents. YouTube · Making money.

## Selling an AI-generated movie

The Terms quoted above are the site's side: results may be used commercially to the extent law and the upstream model terms permit. Whether a buyer, a festival or a platform accepts an AI-assisted film is that party's rule; ask for it in writing. Can you monetize AI videos on YouTube and TikTok? · What to charge for an AI video project.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Medieval Castle, Urban Opulence; Costume Style Smart Casual; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Write the script and the storyboard changes in your own words and keep them; they are the layers that are plainly yours. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Source | What it is | Checked |
| US Copyright Office, NewsNet 1060 | Release on Part 2 of the AI report (January 29, 2025) | November 2, 2026 |
| US Copyright Office, AI page | Parts 1–3 of the report and the registration guidance | November 2, 2026 |
| SceneMixer Terms, sections 06 and 07 | Marks on outputs; ownership of results and uploads | November 2, 2026 |
| The AI badge switch | How the visible mark works | November 2, 2026 |

## FAQ

### Can AI-generated movies be copyrighted?

The US Copyright Office's January 2025 release says outputs of generative AI can be protected only where a human author has determined sufficient expressive elements, and not by the mere provision of prompts. How that applies to a particular film is not decided on this page.

### Does SceneMixer say I own my video?

The Terms say you own the generated results to the extent permitted by law, including commercial use, subject to the Terms, and that there is no guarantee outputs are copyrightable or registrable.

### Can another user get the same output?

SceneMixer's Terms say similar inputs may yield similar or identical outputs for other users, so outputs are not exclusive.

### Can I register an AI-generated film?

The Copyright Office publishes registration guidance for works containing AI-generated material; its AI page links it. Read it before you file, and ask the Office or a lawyer about your case.

### Can I sell an AI-generated movie?

The Terms allow commercial use of results to the extent permitted by law and upstream model terms. Whoever you sell to may have its own rules.

### Does the AI badge affect any of this?

The SceneMixer badge is a visible mark on the video; the Terms say not to remove or alter implicit marks, watermarks or provenance metadata. This page does not say how a mark relates to authorship.

### Is this legal advice?

No. It quotes the Copyright Office and the site's Terms and leaves the judgement to you, the Office and, where it matters, a lawyer.

Updated November 3, 2026


---

# Can AI Turn a Book Into a Movie? What Works and What Does Not

URL: https://scenemixer.com/guides/can-ai-turn-a-book-into-a-movie/
Can AI turn a book into a movie? Yes for a book you wrote: it reads the text into a cast and sequences and renders a voiced film. Length, cost and limits.

# Can AI turn a book into a movie?

Quick answer: Yes, for a book you wrote or have the rights to. Upload the text (.txt, .docx or .md) and SceneMixer reads it into characters, places, props and a sequence-by-sequence outline, draws each once and renders the film with spoken dialogue. One project reads up to 300,000 characters of English, about 50,000 words; a longer book is split into volumes that share one cast. What it will not do is turn a whole book into a feature in one click: the film is made one sequence at a time and paid by the finished second.

Start small: paste chapter one into a new project, set the type to Film, and read the cast and the outline before you decide how much of the book to make.

## What AI can do with a book

- Read it. Characters, places, props and an outline by sequence or episode.
- Draw it. Each person and place drawn once in a look you choose.
- Film it. Each sequence storyboarded and rendered with dialogue, a narrator if you want one, and a score.
Book to movie AI · Book to video.

## What it cannot do

- Use a book you do not have rights to. SceneMixer refuses other people's characters and well-known film versions.
- Take a PDF or an e-book file. Paste or upload the text as .txt, .docx or .md.
- Show inner thoughts. Keep them as voice-over or turn the important ones into lines.
- Make 90 minutes for free. Video is paid by the finished second.

## Film, series or both

A short book or one story suits Film. A novel suits TV Series, one or two chapters per episode. Some authors make both: a trailer and the opening sequence as a film, then a series for the rest. Short drama, series or film · A book trailer · A web novel into a series.

## The honest budget

Reading a book is cheap (4 credits per 1,000 characters, English counted lighter). Making it is not: the video is 10 credits a second on the default tier plus 1 credit a second to join. A ten-episode series of 90-second episodes from a 30,000-character novel is about 10,026 credits, about $62 at the lowest credit price. Choosing script length and cost.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Medieval Castle, Pastoral Healing; Costume Style Soft Everyday, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Check the cast list the book produced: merge a nickname and its full name, remove a place you do not need. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Kind of book | Works as | Notes |
| Your own novel | A film, or a series of one or two chapters per episode | The usual case |
| A memoir or family history | A short film or a few episodes | Real people become invented faces unless you have their consent |
| A children's picture book | A short cartoon | A drawn look such as Watercolor Storybook |
| A public-domain classic | An adaptation in any style | Check the text is public domain where you publish |

## FAQ

### Can AI turn a book into a movie?

Yes, for a book you wrote or have the rights to: SceneMixer reads the text into a cast and sequences and renders a voiced film one sequence at a time.

### How long a book can I upload?

Up to 300,000 characters of English per SceneMixer project, about 50,000 words. Longer books are split into volumes with one cast.

### Which file types work?

.txt, .docx and .md, or pasted text. SceneMixer refuses PDFs and e-book files.

### Can I adapt someone else's book?

Only with the rights; SceneMixer refuses famous characters.

### How much does it cost?

Reading is 4 credits per 1,000 characters. Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Should I make a film or a series?

A short book suits Film; a novel usually suits a TV Series season.

### Is there a free try?

Chapter one, the cast, the outline, the first storyboard and a 5-second preview are free or small on SceneMixer.

Updated November 3, 2026


---

# Cancel Auto-Renew When You Can't Sign In: Ask Support

URL: https://scenemixer.com/guides/cancel-auto-renew-when-you-cannot-sign-in/
Locked out but a renewal is coming? Support can cancel the auto-renew for you. What to send, how long it takes, and how long your plan keeps working after.

# Cancel Your Plan's Auto-Renew When You Can't Sign In

Quick answer: If you can sign in, cancel it yourself: Membership → Cancel subscription, immediately. If you cannot — you no longer have the email, Google sign-in fails — write to support@scenemixer.com from the email address on the account and ask for the renewal to be cancelled. Since October 2026 support can cancel the auto-renew on your behalf, the same way the button does: your plan keeps its benefits and credits until the end of the period you paid for, and nothing further is charged. Nothing already paid is refunded by cancelling; a refund is a separate request.

Before writing, try signing in once more with the email code: a new code can be sent from the sign-in page, and it often solves the problem in a minute.

## First, try to sign in

Sign-in on scenemixer.com is with Google or with a code sent to your email; there is no password to forget. If one way fails, try the other with the same email address. Once you are in, open Membership and press Cancel subscription; it takes effect at once. Sign up and log in · Subscribe, upgrade or cancel a plan.

## When you really cannot get in: write to support

- Write to support@scenemixer.com, from the email address on the SceneMixer account if you still can. That is how support can tell the request is yours.
- Say plainly what you want: “Please cancel the automatic renewal of my membership.”
- Include what identifies the subscription: the account email, the plan, how you pay for it (card, PayPal or Airwallex), and the date of the last charge. If you write from another address, add a receipt or the last four digits of the card shown on the receipt — never the full card number.
- Send it well before the renewal date. Support replies on business days, typically within two business days; a request sent the night before a renewal may not be handled in time.
The reply tells you when it is done. You do not get an automatic confirmation email from the system, so keep the reply. How to get help.

## What happens after it is cancelled

| You pay with | What cancelling does | Your plan |
| Card or Airwallex subscription | The renewal is switched off at the end of the current period. Once you can sign in again, Resume on the membership page turns it back on with no new charge, as long as the period has not ended. | Keeps working until the end of the paid period |
| PayPal | The PayPal subscription is stopped straight away, so nothing more is collected. There is no Resume; subscribe again later if you want it back. | Keeps working until the end of the paid period |
- Credits already in your balance stay there; cancelling does not touch them.
- After the period ends, the account goes back to SceneMixer's free plan.

## A charge already happened

Cancelling stops future renewals. If a renewal went through before your request was handled and you want that money back, say so in the same message: a refund is decided under the refund rules, not by the cancellation. Refund rules · Chargebacks and disputes.

## Other ways to stop a renewal

- PayPal: you can also stop it from your own PayPal account, under automatic payments. The plan still runs to the end of the period.
- A renewal payment that failed is not a cancellation; see what the plan does next. Renewal payment failed.
- Deleting the account is a different, final step, also done by writing to support. Delete your account.

| Situation | What to do |
| You can sign in | Membership → Cancel subscription |
| You cannot sign in | Email support@scenemixer.com from the account address and ask for the renewal to be cancelled |
| Renewal is tomorrow | Write today; PayPal users can also stop it in PayPal |
| A renewal was charged and you want it back | Ask for a refund in the same message |

## FAQ

### Can support cancel my subscription for me?

Yes, since October 2026. Write to support@scenemixer.com from the email on the account and ask for the automatic renewal to be cancelled. It works the same way as SceneMixer’s button on the membership page.

### Will I lose my plan straight away?

No. Benefits and credits on SceneMixer stay until the end of the period you paid for; only the next renewal is stopped.

### Do I get a refund when it is cancelled?

Cancelling on SceneMixer does not refund what was already paid. If you want a recent renewal back, ask for a refund in the same message; it is decided under the refund rules.

### How long does support take?

Replies come on business days, typically within two business days. Write well before the renewal date.

### I paid with PayPal. Is there a faster way?

Yes. You can stop the subscription yourself in your PayPal account under automatic payments. Your SceneMixer plan still runs to the end of the period.

### Will I get a confirmation email?

Support's reply is the confirmation; SceneMixer does not send a separate email for a cancellation made on your behalf.

### Can I turn the renewal back on later?

For a card or Airwallex subscription, yes: sign in and press Resume on SceneMixer's membership page before the period ends. A PayPal subscription cannot be resumed; subscribe again instead.

### Does this apply to the Chinese site?

No. Memberships on SceneMixer (scenemixer.cn) are one-time purchases that do not renew automatically, so there is nothing to cancel.

Updated November 3, 2026


---

# Change a Character's Voice in a Finished AI Video Clip

URL: https://scenemixer.com/guides/change-a-characters-voice-in-a-finished-clip/
A character sounds wrong in a finished clip? Pick another voice from the library in Redub, listen first, keep the words, redub. Which voice is used by default.

# Change a Character's Voice in a Clip You Already Have

Quick answer: Open Edit this clip → Redub. Each line has a voice pill; click it, listen to voices in the library, pick one, and press Redub into a new version without changing the words. The character’s other lines that used the same voice switch with it. With Lip sync it costs 1 credit per second of clip. The voice you pick is used for this clip; the character’s own voice setting does not change.

Click the play button beside a voice pill before you redub: it plays that voice's sample, so you hear it before paying for it.

## Which voice each line starts with

If the character has a voice from the library assigned, that is the one on the pill. Most characters have only a written voice description and no assigned voice; for them, and for speakers who are not in the cast list, the site picks a default from the library that matches the gender, language and age of the character. It is only a starting point for this redub — if it does not sound right, change it.

## Pick and listen

- Select the segment and open Edit this clip → Redub.
- Click the voice pill on one of the character’s lines. The library opens, grouped into male, female and other voices, with a search box; voices in the clip’s language are listed first, others carry a language tag.
- Press play on a few voices to hear them, then click the one you want.
- Back in the panel, every line of that character that used the same voice now shows the new one. Press the play button on a row to hear it again.
- Leave the words alone and press Redub into a new version.

Lines by the same speaker that share a voice always change together, so a character keeps one voice through the clip. Lines by different speakers are independent. Every line needs a voice before you can submit.

## Only the voice, not the words

If you do not touch a line’s text, its words stay the same. Emptying a text box does not delete the line either — the original words come back when you click away — so changing voices never rewrites your storyboard. If you do edit words at the same time, those changes are written back into the storyboard. Change a line after the video is made.

## Voices you uploaded yourself

Redub only offers SceneMixer’s voice library. A voice you uploaded for a character is used when segments are generated, but it cannot be used to voice a redub; that character starts with a library voice chosen by the rule above. If the uploaded voice matters, regenerate the segment instead. How character voices work.

## For the whole series

A voice picked in Redub affects this clip only. To make a character sound different everywhere, change the voice on the character itself, then regenerate or redub the segments you want in the new voice. New segments generated after the change use the new voice.

| You want to | Do this |
| Hear a voice before paying | Press play beside the voice pill or in the library |
| Change one character's voice in this clip | Pick a new voice on one of their lines; matching lines follow |
| Give two characters different voices | Pick a voice on each character's own lines |
| Keep the words exactly | Leave the text boxes as they are |
| Change the voice for every future segment | Change the voice on the character |

## FAQ

### Why did two lines change when I picked one voice?

Lines by the same speaker that used the same voice change together, so one character does not end up with two voices in one clip.

### Why is the voice pill orange?

That line has no voice yet. Pick one; every line needs a voice before you can redub.

### Does this change the character's voice in other clips?

No. It changes this clip only.

### Can I clone a real person's voice?

No. Redub uses the site's voice library.

Updated November 3, 2026


---

# Change a Line of Dialogue After the AI Video Is Made | SceneMixer

URL: https://scenemixer.com/guides/change-a-line-after-the-video-is-made/
Rewrite what a character says in a finished clip without regenerating: Redub lists the lines, you edit them and choose Lip sync or Re-perform. Price and limits.

# Change a Line of Dialogue After the Video Is Made

Quick answer: Open Edit this clip → Redub. Every spoken line of the segment is listed with its speaker and voice; change the words you want, choose Lip sync (the picture stays and the mouths follow the new audio) or Re-perform (the clip is acted again with the new lines), and press Redub into a new version. Lip sync costs 1 credit per second of clip; Re-perform costs the same as AI visuals. The changed lines are written back into the storyboard too.

Good first use: a line with one wrong word. Fix the word, keep Lip sync, and compare the new version with the original in Version History.

## What the panel shows

One row per line of dialogue in the segment, in order. Each row has the speaker’s name, a voice pill (the voice that line will use), a play button to listen to that voice, and a text box with the line itself. Only lines written as dialogue in the storyboard are listed; if the segment has none, the tab says This segment has no dialogue.

## Change what is said

- Select the segment and open Edit this clip → Redub.
- Edit the text of the lines you want to change. Leave the others as they are.
- Check the voice pill on each row; click it to pick a different voice from the library. Changing a voice.
- Choose Lip sync or Re-perform; each shows its own price.
- Press Redub into a new version. The panel shows Redubbing… until it is done.

If you empty a text box, the line is not deleted: when you click away the original words come back, and that line is spoken as before, only in the voice you picked. Lines cannot contain brackets or line breaks.

## Lip sync or Re-perform, in one paragraph

Lip sync keeps every frame and replaces the sound: all the lines of the segment are voiced again, one after another from the start of the clip with short pauses, and the mouths are moved to match. The clip’s own soundtrack — its ambience, and any music the take carried — is not kept; the episode’s score is mixed back under the new version on its own, usually within minutes. SceneMixer charges 1 credit per second of the whole clip. Re-perform plays the clip again from itself with the new lines, so timing and delivery fit the words, but the picture can change a little; it costs 2× the per-second price of the clip’s tier and is only offered for clips of 15 seconds or less. How to choose.

## Keep it short enough to fit

The new lines have to fit in the clip. Before anything is charged, the site estimates how long they take to say at a natural pace, plus the pauses; if that is longer than the clip, it stops with Too much dialogue for this clip’s length. Shorten the lines. A clip cannot grow longer to make room — for a longer speech, edit the storyboard, give the shot more seconds and regenerate. Shot seconds and timing.

## The storyboard is updated too

Changed words are written back into this segment’s storyboard text, only those words, and the panel confirms: The new lines were also written into the storyboard. So if you regenerate the segment later, it says the new line. Voices you pick here are used for this redub; to change a character’s voice everywhere, change it on the character. How character voices work.

## Where the result goes

A new version labelled Redubbed in Version History, active straight away if the clip you started from is still the one showing. A segment keeps one redubbed version; the next redub replaces it. The generated clip stays. In an ad project, whether the music comes back depends on how the ad was made. In an ad scored the new way, the ad’s piece is laid over the whole ad when it is composited, and the redubbed clip is scored like the rest, so the music returns by itself — composite again to carry the new version into the final cut. In an ad made before the new scoring, each clip was generated with the video model’s own music, and lip sync replaces the clip’s soundtrack, music included, so that clip comes back without music — re-perform it or regenerate the segment instead. If a redub does not go through, the credits come back; in the rare case its result cannot be confirmed, they are held for review. How versions are kept.

| | Lip sync | Re-perform |
| Picture | Unchanged, mouths re-synced | Played again, can change a little |
| Sound | All lines re-voiced from the start; ambience not kept | New performance with the new lines |
| Clip length | Any | Up to 15 seconds |
| Price | 1 per second of clip | 2× the tier’s per-second price × length |

## FAQ

### Can I add a new line or delete one?

No. Redub changes the words and voices of the lines that are there. To add or remove lines, edit the storyboard and regenerate.

### Why are lines I did not touch re-voiced too?

Lip sync replaces the whole soundtrack of the clip, so every line is voiced again in the voices shown; unchanged lines keep their words.

### Does the music disappear when I redub?

Not for long. Lip sync replaces the clip’s own soundtrack — its ambience, and any music the take carried — and the episode’s score is mixed back under the new version automatically, usually within minutes. If the clip’s own ambience matters, regenerate the shot.

### Can I use a voice I uploaded myself?

Not here. Redub offers the voices in SceneMixer's voice library.

### Does it change the next segment?

No. Only this clip, plus the words in this segment's storyboard.

Updated November 3, 2026


---

# Change Clothes Colour, a Scar or an Expression in an AI Video Clip

URL: https://scenemixer.com/guides/change-clothes-color-scar-or-expression-in-a-clip/
Recolour a character's outfit, add or remove a wound, or change a face to terrified, calm, smiling, angry, sad or crying, in a clip you already have. How it works.

# Change a Character's Clothes Colour, Scar or Expression in a Finished Clip

Quick answer: Open Edit this clip → AI visuals, pick Clothes color, Scar or Expression, choose the option, and pick the character under Who. Clothes keep their cut and fabric and only change colour; Scar → Add gives a fresh cut on the face, Remove takes every wound off; an expression holds for the whole clip while the body moves as before. It costs 2× the per-second price of the clip’s tier for the length of the clip, and changes this clip only — not the character sheet.

Pick the character from the Who row: people who appear in this segment are listed first, and if only one does, they are already selected.

## Clothes color

Eight colours: red, blue, black, white, green, yellow, purple and gray. The garment keeps its cut, fabric and fit; only the colour changes, on the character you picked. Use it when one scene needs a different look — a red dress for the party, a black coat for the funeral — or when a clip came out with the wrong colour.

If the outfit should be different in the whole series, change the character instead: edit the sheet (or add a costume state) and regenerate the segments that use it. AI-edit a character sheet and Costume changes and states.

## Scar

Add puts a fresh cut on the character’s face with a thin trail of blood; the rest of the face stays as it was. You do not choose where it goes. Remove makes the skin clean everywhere: no wounds, no blood and no bruises on the face or body — not only the scar you had in mind. Wounds are a sensitive subject for content review; SceneMixer refunds a blocked edit. Blocked by moderation.

## Expression

Terrified, Calm, Smiling, Angry, Sad or Crying. The face holds that expression through the whole clip and the body moves exactly as before, so a character who shouted still shouts, now with a different face. Expressions are the subtlest of the eight changes: they read best in close and medium shots with the face toward the camera, and can be hard to see in a wide shot. If the acting itself is wrong, change the line or the action in the storyboard and regenerate.

## Step by step

- Select the segment and open Edit this clip → AI visuals.
- For a clip over 15 seconds, choose which half.
- Click the change, then the option (a colour, Add or Remove, an expression).
- Under Who, click the character.
- Press Edit into a new version. The new version is labelled AI edit in the history, with a note such as “Mara · clothes → red”.

## One person, one change per edit

Each edit changes one thing on one character. To recolour two people, or to recolour and add a scar, run a second edit on the result. Each run is charged, and the newer AI-edited version replaces the older one in the history. Everything AI visuals can change.

| Change | Options | What it keeps |
| Clothes color | Red, blue, black, white, green, yellow, purple, gray | Cut, fabric, fit, everything else |
| Scar → Add | — | The rest of the face |
| Scar → Remove | — | Removes every wound, blood and bruise |
| Expression | Terrified, calm, smiling, angry, sad, crying | The body movement and the lines |

## FAQ

### Does this change the character sheet?

No. Only this clip changes; other segments and later regenerations use the sheet as it is.

### Can I choose where the scar goes?

No. Add gives a fresh cut on the face; for a specific wound, describe it in the storyboard or the sheet and regenerate.

### Can I pick a colour that is not in the list?

No. For another colour, change the character's clothing and regenerate.

### Why are Clothes color, Scar and Expression greyed out?

The project has no characters. These three changes need a character to apply to.

Updated November 3, 2026


---

# Change Day to Night or Add Rain to a Finished AI Video Clip

URL: https://scenemixer.com/guides/change-day-to-night-or-add-rain-in-a-finished-clip/
Turn a finished segment from day to night or dusk, or make it rain, snow or clear up, without regenerating it. What each option looks like, the price, and continuity.

# Change Day to Night, or Add Rain or Snow, in a Clip You Already Have

Quick answer: Open Edit this clip → AI visuals, choose Time of day (Daytime, Night, Dusk) or Weather (Rain, Clear sky, Snow), and press Edit into a new version. The clip is made again with the same shot, people, movement and lines, and only the light or the weather changed. It costs 2× the per-second price of the clip’s tier for the length of the clip. Your storyboard and the other segments of the scene do not change.

A 10-second clip on Wan 3.0 · 480P changed to Night costs 10 × 20 credits; on Wan 3.0 · 720P, 10 × 38.

## What each option aims for

- Night — a deep blue-black sky, the scene lit only by its own lamps and lights, deep shadows, a quiet mood. A scene with no lamps in it gets very dark; see the tip below.
- Dusk — low warm orange light from one side, long soft shadows, a sky going from amber to violet.
- Daytime — bright neutral daylight, a clear sky, short soft shadows.
- Rain — visible rain, wet glossy ground with reflections, damp hair and clothes, grey overcast light.
- Snow — snowflakes drifting through the frame, a thin layer settling on surfaces, cold pale light.
- Clear sky — dry ground, a bright sky, crisp daylight. Use it to take rain or snow out.

## Step by step

- Select the segment; under the player open Edit this clip → AI visuals.
- For a clip longer than 15 seconds, choose First half or Second half.
- Click Time of day or Weather, then the option.
- Check the price on Edit into a new version and press it. It usually takes a few minutes.
- Watch the new version. If the old one was better, switch back in Version History.

## Night and lamps together

Night lights the scene only from the light sources already in it. An interior with lamps, a street with lanterns or a room with a lit screen turns into a convincing night; an open field with nothing in it turns into a dark frame. If the scene has lamps or candles that are off in your clip, you can run Lights → Light them up on the night version as a second edit. Lights and background people.

## One change per edit

Time of day and weather are two separate changes. For a rainy night, run Time of day → Night first and Weather → Rain on the result. Each run is charged, and the second AI-edited version replaces the first in the history (the generated clip stays). How versions are kept.

## Keep the scene consistent

An AI visuals edit changes this clip only. The storyboard still says what it said, the location image is untouched, and the other segments set in the same place keep their own light. If a whole scene should be at night, change it for the scene rather than clip by clip, so every segment and any later regeneration agree. In a newer storyboard, press Change on the segment header’s Time · Weather row and pick the time or weather: it applies to all the segments of that scene, is free, and takes effect the next time they are generated. In an older storyboard, change the Time or Weather line in each of those segments’ storyboards and regenerate them. Use AI visuals when one or two clips are off. The segment header fields.

| Option | What it aims for | Pairs well with |
| Night | A blue-black sky; light only from the scene's own lamps | Lights → Light them up |
| Dusk | Low warm light from one side, long soft shadows | Exteriors and rooms with windows |
| Daytime | Bright neutral daylight, short soft shadows | Turning a night clip into day |
| Rain | Visible rain, wet glossy ground, damp hair and clothes | Exteriors |
| Snow | Falling flakes, a thin layer on surfaces, cold light | Exteriors |
| Clear sky | Dry ground, a bright sky | Taking rain or snow out |

## FAQ

### Will the next segment match?

Only if you change it too. SceneMixer's AI visuals edits one clip; for a whole scene, change the Time · Weather on the segment header and regenerate.

### Can I make it night and rainy at once?

Not in one edit. Run Night, then Rain on the result; each is charged.

### Does the dialogue change?

No. The lines, the movement and the timing follow your clip.

### Can I choose an exact time, like 5 pm?

No. SceneMixer offers Daytime, Night and Dusk. For a finer choice, set the Time on the segment header (Dawn, Morning, Noon, Afternoon, Daytime, Dusk, Night or Late night) and regenerate; older storyboards have a Time line to edit instead.

Updated November 3, 2026


---

# Change Region, Dialogue Language or Type After Parsing: What to Do

URL: https://scenemixer.com/guides/change-region-dialogue-language-or-type-after-parsing/
Which project settings can you still change after setup? Step 1 settings are fixed once the script is read. What still changes, and when to start a new project.

# Changing the Target Region, the Dialogue Language or the Content Type After the Script Is Read

Quick answer: Target region, dialogue language and content type are chosen in Step 1 before the script is read, and they shape what the read writes: the region relocates names, faces and places, the dialogue language decides what language the lines are written in, the type sets the pacing. Once the read has finished, Step 1 shows the summary and these settings can no longer be changed in that project. Two ways forward: edit what the read produced — character and scene descriptions in Step 2, lines in the segments — or start a new project with the same script and choose different settings before it is read.

Open Step 1 of a parsed project: the settings form has been replaced by the summary of what was read.

## Where the three settings live

In the Step 1 form you fill in before the script is read: target region (where the story is set, from a list of regions, or follow the script), dialogue language (the language the cast speaks), and content type (short drama, series, film). When you continue, the script is read with them and they are saved with the project. Choosing each one: Target Region, Dialogue Language, Short Drama, Series or Film.

## Why they cannot be changed afterwards

All three do their work during the read: by the time it finishes, their effect is written into the cast, scene and prop descriptions and into the episode outlines. After that, Step 1 shows the summary instead of the form, and the project keeps the settings it was read with. The Re-parse button in Step 3 rewrites one episode’s storyboard from its outline; it does not read the script again and does not change these settings. Checking the Summary, and What Re-parse Means.

## Changing the region of an existing cast

Edit each character’s description in Step 2 (face, build, clothing) and the scene descriptions, rename characters if you want local names, then regenerate those sheets. Storyboards generated afterwards read the descriptions you edited. Write a Description That Works.

## Changing only the dialogue language

Edit the lines in the segments; the video speaks what the segment says. Regenerating an episode’s storyboard does not change the language: it copies the outline’s lines word for word. For a whole project in another language, start a new project with the same script. Remakes are different: their dialogue language is fixed when the clip is read.

## Starting again with the same script

Create a new project, paste the same script, choose the settings you want — including follow the script in the region list, which keeps names and places as written — and continue. The new project’s read is charged like any read; the old project stays as it was. If a read fails or you stop it, Step 1 goes back to the form and you can change the settings before trying again.

| What you want | In the same project | In a new project with the same script |
| Another region | Edit descriptions and names in Step 2, regenerate sheets | Choose the region before the read |
| Another dialogue language | Edit the lines in the segments | Choose the language before the read |
| Another content type | Not possible; the episodes were cut for the old type | Choose the type before the read |
| Another visual style or aspect ratio | Not possible | Choose it before the read |

## FAQ

### Can I change the region after the script is read?

Not in that project. Edit the descriptions and names in Step 2, or start a new project with the same script.

### Is there a button to read the whole script again?

No. The Re-parse button in Step 3 rewrites one episode’s storyboard from its outline; it does not read the script again.

### Can I change the dialogue language on a remake?

No. A remake’s dialogue language is decided when the clip is read.

### Does a new project with the same script cost a new read?

Yes. SceneMixer charges each project’s read by script length.

### What if the read failed or I stopped it?

Step 1 goes back to the form; change the settings and continue again.

Updated November 3, 2026


---

# Change a Segment's Duration: 15 or 30 Seconds, Which Tiers Allow It

URL: https://scenemixer.com/guides/change-segment-duration-15-or-30-seconds/
A segment's length is the sum of its shots. How to change the seconds, the two limits on the total, and what a longer segment costs.

# How to Change a Segment’s Duration, and Which Tiers Accept 30 Seconds

Quick answer: A segment is one generation, and its length is the sum of its shots’ seconds. Change a shot’s seconds in its header line (or through the @ menu) and the total follows. Two limits apply. The episode’s segment length — Standard 15s or Long take 30s, chosen in Step 3 when you generate its storyboard — caps what the editor lets you set. The video tier caps what it will generate: the Wan 3.0 and Seedance 2.5 tiers accept up to 30 seconds, the Seedance 2.0 and MiniMax H3 tiers up to 15. You pay per second generated, so a 30-second segment costs twice a 15-second one on the same tier.

Open a segment, change one shot from 4 seconds to 6, and watch the segment total and the price tag update before you generate.

## Where the seconds live

Each shot line starts with its number and its seconds. Edit the number, save, and the segment’s total and its price tag on the timeline card change immediately; nothing is generated until you press the button. The @ menu offers the same edit with a preset list. The segment total stays between 4 seconds and the episode’s segment length; the editor shows the total in red when the shots add up to more. It does not check whether a line fits its seconds, so give a spoken shot the time it takes to say. Details of the timing rules: Shot Seconds and Timing.

## The tier’s cap

One segment is one request to the video model, and each model accepts a maximum length per request: Wan 3.0 (480P, 720P, 1080P) and Seedance 2.5 (480P, Lite, Pro) take up to 30 seconds; Seedance 2.0 (480P, Fast, standard) and MiniMax H3 (H3, H3 Fast) take up to 15. Switch tiers in the model selector in the editor’s top bar and the cap switches with it. A segment longer than the tier’s cap is refused when you press generate, with a message naming the cap; it is not cut, and nothing is charged.

## The episode’s segment length

Each episode’s storyboard is written to a segment length you pick in Step 3 before generating it: Standard 15s (the default) or Long take 30s. It shapes how shots are grouped into segments, and it is also the ceiling for editing: in a 15-second episode a segment stops at 15 seconds whatever you type. To get 30-second segments, generate the episode’s storyboard in Long take 30s and generate them on a Wan 3.0 or Seedance 2.5 tier. What the two modes change in practice: 15-Second or 30-Second Segments.

## Making a segment longer without making it worse

Longer is not automatically better: one 30-second generation has to hold the same faces and the same room for twice as long, and a mistake in second 25 costs the whole take. Add seconds to shots that need them (a spoken line, a slow reveal), not to every shot; if you want more story, add a segment rather than stretching one — Insert or Delete a Segment. Leave a couple of seconds under the cap on the 30-second tiers.

## What it costs

Price is per second of the tier: SceneMixer's default tier is 10 credits a second, so a 12-second segment and a 24-second segment differ exactly by that ratio. The timeline card shows the price before you generate; regenerating a longer version is a new charge. Where your credits go in one episode.

| Tier family | Max seconds per segment | Typical use |
| Wan 3.0 · 480P / 720P / 1080P | 30 | Default; drafts to finals |
| Seedance 2.5 · 480P / Lite / Pro | 30 | Longer single takes |
| Seedance 2.0 · 480P / Fast / standard | 15 | Short punchy shots |
| MiniMax H3 / H3 Fast | 15 | Short takes, no 480P |

## FAQ

### Can I make a segment 30 seconds?

Yes, if the episode's storyboard was generated in Long take 30s mode in Step 3 and you generate on a Wan 3.0 or Seedance 2.5 tier. In a 15-second episode segments stop at 15.

### Why does my segment total not match what I typed?

The total is the sum of the shots, kept between 4 seconds and the episode's segment length. Check every shot line.

### Does a longer segment cost more?

Yes, per second, on every tier. The card shows the price before you generate.

### Can I generate 4 seconds?

Yes; 4 seconds is the shortest segment. Make sure a spoken shot still has time for its line.

### Will the site cut my segment to fit?

No. SceneMixer refuses a segment past the tier's cap when you press generate, before anything is charged; shorten it or pick a 30-second tier.

Updated November 3, 2026


---

# AI Character Looks the Wrong Age or Ethnicity: How to Fix It

URL: https://scenemixer.com/guides/character-looks-the-wrong-age-or-ethnicity/
A character came out too young, too old or of a different background than you wrote. Where the look comes from, how to fix the sheet or use a photo, which clips to redo.

# When a Character Looks the Wrong Age, Ethnicity or Skin Tone

Quick answer: A character’s look starts from the description written when your script was read, and the sheet in step 2 is drawn from it. If the age or background is wrong, fix the sheet first: open the character, put the age and background in the first sentence of Appearance (“a Black man in his fifties, grey at the temples”), save and regenerate. Or, on a paid plan, upload a photo; videos then follow the photo’s face, skin tone and clothes. Clips made before the change keep the old look, so regenerate the segments the character appears in.

Open the character in step 2 and read the first sentence of Appearance: whatever it says about age and background is what every shot starts from.

## Where the look comes from

When the script is read, each character gets a written description. If the script states an age or a background, that is used; if it does not, one is chosen to fit the story, and the Target Region you set in step 1 decides where the cast is from. The sheet is drawn from that description, and every video shot of the character starts from the sheet. Target region.

## Fix the sheet first

- In step 2, open the character.
- In Appearance, put the age and background first, in plain words: a Korean woman of about twenty-two, a Nigerian man in his sixties. Keep the distinctive features that follow.
- Save, then press Regenerate. Compare the new sheet with what you want before going further.

Regenerating a sheet on SceneMixer is priced like an image, unless a free sheet covers it. Regenerate one sheet · Describe a face.

## Or use a photo

On a paid plan, Upload & Replace puts your own image on the card. From then on, videos follow the photo’s face, skin tone and clothes rather than the written description. Pick a clear, front-facing photo with even light. Use your own photo.

## Redo the clips

Segments generated before the fix keep the old look. Find the segments the character appears in and regenerate them; the old takes stay in Version History. Regenerate one segment.

## If the sheet is right but a clip is not

When the sheet shows the right person and a clip still shows someone of another age or background, check that the shot names the character with an @ pill. If it does, do not keep regenerating: send feedback with the segment number. My character’s face changed.

## The whole cast is from the wrong place

That is the target region. It is chosen before the reading and fixed afterwards; to move the whole story, start a new project with the same script and another region. For a few characters, fixing their sheets is quicker.

| Problem | Where it comes from | Fix |
| One character's age or background | Their description | Edit Appearance, regenerate the sheet |
| Want an exact face | The sheet | Upload a photo (paid plan) |
| Old clips still show the old look | Made before the fix | Regenerate those segments |
| Sheet right, clip wrong | That clip | Check the @ pill; send feedback with the segment number |
| The whole cast | Target region | New project with another region |

## FAQ

### Why did my character come out younger than I wrote?

Some visual styles draw people younger. State the age plainly in the first sentence of Appearance and regenerate the sheet.

### I uploaded a photo and the video changed the skin tone.

Videos follow the photo. If a clip made after the upload does not, send feedback with the segment number; clips made before the upload need regenerating.

### Does changing the description change existing videos?

No. Only clips generated afterwards.

### Can I write the ethnicity in the script itself?

Yes. A script that states age and background gives the reading something to keep.

Updated November 3, 2026


---

# Character Name Generator for Stories: 250+ Names and AI Video Rules

URL: https://scenemixer.com/guides/character-name-generator-for-stories/
Names for story characters in seven groups, a fantasy-name recipe, nickname pairs, and the naming rules that matter for AI video: unique names, no famous roles.

# Character name generator for stories

Quick answer: Pick a name that is short, easy to say aloud and different from every other name in your cast. This page gives about 250 names to start from in seven groups, plus a recipe for making fantasy names. For AI video, three rules matter: a name must be unique inside a project, because shots refer to characters by name; avoid the names of famous roles, because the face can drift toward the actor; and prefer names the dialogue voice can say plainly. You can rename any card in Step 2 and every reference updates.

Pick three names that start with different letters, say them aloud once, and put them in your AI Create brief so the script keeps them.

## Naming rules for AI video

| Rule | Why |
| One name, one card, per project | Storyboards look characters up by name. A second card with the same name is refused; two characters with one name in the script get a suffix like (2) until you rename them. |
| Spell it the same way every time | Nicknames and short forms are collected on the card as Also called (up to eight), so Danny can still mean Daniel, but a new spelling is a new name. |
| Do not name a character after a famous role | A character named after, or written like, a well-known role can come out looking like the actor, which the video provider's face check may refuse. |
| Use different first letters and lengths | A cast of Mara, Mari and Maya is hard to follow in a long cast list and in dialogue. |
| Choose names the voice can say | The dialogue voice reads the name aloud. Hear a line in the free 5-second preview before you commit. |
| Do not use real people | Using a real person's likeness without consent is not allowed, and real public figures are refused. |
Why names are unique · Also called: nicknames and old names · When a face is flagged · Describing a face.

## Everyday first names, by language

| Language or region | Names |
| English-language | Nell Theo Cass Wren Jonah Tess Hollis Mabel |
| Spanish and Latin American | Lucía Mateo Inés Rafael Camila Joaquín Marisol Tomás |
| French, German and Italian | Camille Étienne Lena Jonas Giulia Matteo Chiara Luca |
| Russian and Ukrainian | Katya Dmitri Oksana Taras Sasha Mila Andriy Olena |
| Turkish | Elif Emre Deniz Selin Kerem Defne Baran Zeynep |
| Arabic | Layla Omar Samira Yusuf Nadia Karim Rania Idris |
| South Asian | Aarav Meera Kabir Anaya Rohan Zoya Dev Isha |
| Chinese | Wei Mei Jun Lin Xiaoyu Haoran Yan Ling |
| Japanese | Haruka Ren Sora Aiko Hiroshi Yuki Kaito Miyu |
| Korean | Min-jun Seo-yeon Ji-ho Ha-eun Do-yoon Soo-ah Tae-yang Ye-jin |
| West African (Nigerian and Ghanaian) | Tunde Folake Femi Chidi Ngozi Obinna Kofi Ama |
Pick names that suit the dialogue language you choose in Step 1; there are 15 to pick from. Choosing the dialogue language.

## Surnames

| Group | Surnames |
| English and Irish | Hale Marsh Calder Finch Ashby Rooke Garrity Lowell |
| Spanish, Italian and Portuguese | Castellanos Navarro Rossi Bianchi Ferreira Lopes Marchetti Vargas |
| Eastern European and Turkish | Novak Ivanova Petrenko Kovalenko Yilmaz Demir Kowalski Horvat |
| Nordic and German | Brandt Lindqvist Aalto Holm Becker Falk Vogel Strand |
| Middle Eastern and South Asian | Haddad Rahman Banerjee Qureshi Nasser Iyer Farouk Malik |
| East Asian | Tanaka Kobayashi Lin Zhou Park Seo Chen Watanabe |
| West African | Okafor Mensah Adeyemi Bello Owusu Nwosu Balogun Boateng |

## Fantasy names: a recipe and 40 names

Join a start, an optional middle and an end, and say it aloud. Kael + ar + ith makes Kaelarith; Lyr + is + wen makes Lyriswen.

| Start | Middle | End |
| Ael Bel Cor Dra Elow Fen Gal Hy Kael Lyr | an ar el en il or ur yth ae is | a ia is on wen rin dor thar ael ith |
Soft vowels and flowing consonants read as gentle; hard stops read as threatening. Twenty of each:

| Sounds like a hero | Sounds like a villain |
| Aerwen Belmoor Corisel Elowen Fenrith Galmere Hyrelle Isolwen Kaelith Lyrawen Merrowyn Nethra Orwyn Pelrin Quillan Rhysael Sorrel Tavrin Ulmara Vaelis | Karzul Brakthar Dromgar Skarn Morgrath Zhaldur Thrull Gorvane Kruxis Vorrak Drazek Xulthar Brannoch Grimsor Tarkun Vexhal Ozmund Rhuk Calvane Harrowick |
Invented names collide often. Search any name you like before you commit, and play a line aloud in the preview to hear how the voice says it. The free preview · An elf or fae fantasy drama · A dark fantasy drama.

## Sci-fi names

| Kind | Names |
| Ships and stations | Kestrel Halcyon Meridian Tethys Orrery Lumen Perihelion Vesper Corvid Anchorage |
| Crew, given name and family name | Ana Vale Kade Orlov Isa Renn Tavi Noor Mirela Quist Juno Hask Teo Marlowe Lio Santoro Priya Wend Oren Lisk |
| Robots and AIs | Tin Ferro Wick Orbit Marlow-9 Unit Bell Pilot Sparrow Sprocket Nib Halo |
A sci-fi short drama · A robot or android drama · A space opera.

## Villains, epithets and animal names

| Kind | Names |
| Villains and epithets | The Auditor Mother Vesper Old Gaunt Mr. Lull Sister Mercy The Collector Captain Brine The Quiet Man Madam Thorn The Locksmith Doctor Hush Aunt Marrow The Tailor Lord Ashgrove The Bell-Ringer Sergeant Hollow The Cartographer Miss Tallow The Moth Warden Pike |
| Animal and toy names for a cartoon | Biscuit Pip Nibs Mochi Clover Button Juniper Bramble Pickle Waffles Tuppence Marmalade Doodle Socks Noodle Pudding Sprout Thimble Cobweb Mittens Toffee Wobble Pebble Fidget |
An animal or toy is a character like any other, with its own sheet and voice. Non-human characters · A cat or pet drama · A kids cartoon channel.

## Nicknames, and how to put names in the brief

A full name and its short forms can all appear in the script; the card keeps them as Also called so the storyboard still knows who is who.

| Full name | Also called |
| Daniel Cole | Danny Dan |
| Elizabeth Hart | Liz Beth Eliza |
| Alexander Voss | Alex Sasha |
| Margaret Ono | Maggie Peggy |
| Mateo Ruiz | Teo |
| Katarzyna Nowak | Kasia |
| Obinna Eze | Obi |
To keep your own names, put them in the brief: the lead is Mara Okafor, 34, a night nurse; her brother is Tunde. Put your names in the brief so the script uses them; AI Create writes the rest. Writing the one-line brief · AI character sheet generator · Character sheet template · Story bible template.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Hospital Ward, Urban Opulence; Costume Style Professional Uniforms, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw your named cast first; the names you settle on are the ones every storyboard will use. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| A cast that blurs | A cast that reads clearly |
| Mara, Mari, Maya, Mia | Mara, Theo, Inés, Oksana |
| Daniel, Danny, Dan (as three people) | Daniel (also called Danny, Dan) |
| A lead named after a famous role | A lead with an ordinary name and a distinct face |
| Two cards called Guard | Guard at the gate, Guard on the wall |

## FAQ

### How do I choose a character name?

Pick a name that is short, easy to say aloud and different in its first letter and length from every other name in the cast.

### Can the AI make up character names?

SceneMixer's AI Create writes names for you when your brief gives none; put your own in the brief and the script uses them. This page gives about 250 names and a fantasy recipe to start from.

### Can two characters have the same name?

Not in one project. SceneMixer refuses a second card with an existing name, and two characters with one name in the script get a suffix such as (2) until you rename them.

### What if my character is named after a famous person or role?

Avoid it. The face can drift toward the actor and the video provider's face check may refuse it. Using a real person's likeness without consent is not allowed.

### Can I rename a character later?

Yes. Renaming a card in SceneMixer rewrites every reference in the project's storyboards, and the old name is kept as an Also called entry.

### What are Also called names?

The nicknames, short forms and old names the script uses for the same character, up to eight, collected on the card so the storyboard knows who is who.

### Which names work in other languages?

Pick names that suit the dialogue language you choose in Step 1. There are 15 to pick from.

### How do I make a fantasy name?

Join a start, an optional middle and an end from the recipe above, say it aloud, and search it before you commit.

Updated November 3, 2026


---

# Character Sheet Template for Stories and AI Video, With an Example

URL: https://scenemixer.com/guides/character-sheet-template/
A character sheet template: name, role, want, face, hair, build, outfit and voice, with a filled example and how SceneMixer turns it into a full-body reference sheet.

# Character sheet template for stories and AI video

Quick answer: A character sheet is a short profile you keep for every main character: who they are, what they want, and exactly what they look and sound like. For AI video the look and voice lines matter most, because they become the reference the picture copies in every shot. In SceneMixer the written sheet is turned into a full-body drawing on a plain background and a voice, and that drawing is reused for the whole series.

Fill the template for your lead, paste it into the character card in Step 2 as the description, and read the drawing it makes before approving.

## The template

| Field | Write | Why for AI video |
| Name and age | MARA OKAFOR, 34 | Anchors the face |
| Role | Night-shift nurse | Anchors clothing and behaviour |
| Want and fear | Wants to leave; fears being needed | Drives the lines |
| Face | Oval, strong brows, chipped front tooth | The identity details |
| Hair and build | Close-cropped hair, tall, narrow shoulders | The silhouette |
| Outfit | Green cardigan over scrubs | Held across episodes |
| Voice | Low, dry, quick to deflect | Chooses her voice |
| Habit | Rolls her ring when lying | A detail to write into shots |

AI character sheet generator · AI actor generator · Picking voices.

## What the tool makes from it

A reference sheet is a full figure on a plain background, which is what every later shot copies. Describe the face, hair, build and outfit in plain words and the drawing follows, in the visual style you chose. If the drawing is wrong, fix it here, once, rather than in shot forty. Members can also upload a reference photo. Using your own photo.

## Write the look so it holds

- Concrete over vague. “Close-cropped hair” beats “stylish”.
- One distinguishing detail. A chipped tooth, a scar, a ring.
- One outfit per state. Changes are explicit states. Costume changes and states.
- No real people or famous characters. They are refused.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Hospital Ward, Urban Opulence; Costume Style Professional Uniforms, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Paste your filled template into the card for your lead and read the drawing before approving it. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Fill in | Skip |
| Face, hair, build, outfit, voice | Backstory the camera will never show |
| One detail to write into shots | A list of ten quirks |

## FAQ

### What is a character sheet?

A short profile of a character: who they are, what they want and exactly how they look and sound.

### What fields matter for AI video?

Face, hair, build, outfit and voice, because they become the reference for every shot.

### Does SceneMixer draw the sheet?

Yes: a full-body reference sheet on a plain background, from your description.

### Can I change a character later?

Yes, on the sheet in SceneMixer; later shots follow.

### Can I use my own photo?

Members can upload a reference photo and have it redrawn by SceneMixer as a sheet in the project's look.

### Are the first drawings free?

The first 15 setup images on an account are free.

### Can I use a real person or a famous character?

No. They are refused.

Updated November 3, 2026


---

# Chargebacks and Payment Disputes: What Happens to Your Account

URL: https://scenemixer.com/guides/chargebacks-and-payment-disputes-what-happens-to-your-account/
Disputing a charge with your bank or PayPal reverses the order here: credits are reclaimed, a membership is voided and cancelled. What happens when the dispute closes.

# Chargebacks and Payment Disputes: What Happens to the Order, the Credits, the Membership, and How to Avoid Needing One

Quick answer: When your bank or PayPal reports a dispute on a payment, the site reverses the order at once. For a credit pack, credits are reclaimed in proportion to the disputed amount; if the balance is short, queued generations that have not started are cancelled first, and any uncovered part is only noted on the order. A disputed membership payment voids the whole order: every period it paid for, everything it granted (including the first-subscription bonus) is reclaimed, and the card or PayPal subscription behind it is cancelled at once. The membership goes under review only if the reclaim cannot be covered. If the dispute is decided for the site (you lose it or withdraw it), the order is restored; if it is decided for you, the money stays with you and the order stays reversed. A refund request to support is resolved in days and never touches your account standing; a dispute takes weeks and does. Write first.

Look at the order in your billing history: a disputed order shows the reversed state, and the credit history shows the reclaim as a Purchase reversed line.

## What a dispute does at once

The order is marked reversed. For a credit pack, the credits are reclaimed in proportion to the disputed amount; if your balance cannot cover it, generations waiting in the queue are cancelled and their credits used, and any remaining shortfall is only noted on the order. For a membership payment, the whole order is treated as disputed: every period it paid for is voided, everything it granted, including the first-subscription bonus, is reclaimed, and the subscription is cancelled at once. Unusual Rows in Your Credit History.

## Membership under review

If the reclaim on a membership order cannot be covered, the shortfall is recorded as a debt against later membership grants and the membership is placed on hold: no member concurrency, no badge switch, and the subscribe buttons refuse with a message that it is under review. SceneMixer’s support reviews it and lifts the hold; it is not lifted automatically when the dispute closes.

## When the dispute closes

Decided for the site (you lost the dispute or withdrew it, so the payment stands): an order voided by the chargeback is restored, the reclaimed credits are returned as they were taken, and a voided membership period is scheduled back with its remaining length. An order that was also refunded is not restored, since that money went back to you. Decided for you: the money stays with you, and the order, the credits and the membership stay reversed.

## Inquiries that are not disputes

Some providers send an inquiry before a dispute. An inquiry moves nothing on SceneMixer; only a filed dispute does. Answer your bank’s inquiry with the receipt and the order number and it usually ends there.

## Why writing to support first is faster

The refund rules cover the cases a dispute would: duplicate charges, an unusable result, a wrong amount. Support resolves those in days, refunds through the original method, and your account is untouched. A dispute takes the provider’s timeline, cancels a disputed subscription at once, and repeated disputes can close an account. Refund Rules.

| Event | Order | Credits | Membership |
| Dispute on a credit pack | Reversed | Reclaimed in proportion | — |
| Dispute on a membership payment | Reversed | Everything the order granted, bonus included | Order voided; subscription cancelled |
| Balance short | — | Queued tasks cancelled; pack remainder noted on the order | Remainder as debt; membership on hold |
| Dispute decided for the site | Restored | Returned | Period scheduled back |
| Dispute decided for you | Stays reversed | Stays reclaimed | Stays voided; any hold is settled by support |
| Inquiry only | Unchanged | Unchanged | Unchanged |

## FAQ

### I disputed by mistake; what now?

Withdraw it with your bank and write to SceneMixer’s support; once the provider reports the dispute closed for the site, the order is restored.

### Why is my membership suddenly on hold?

A disputed membership payment could not be fully reclaimed; the support team on SceneMixer reviews it and lifts the hold.

### Will a dispute close my account?

One dispute, no. SceneMixer treats repeated disputes on delivered work as abuse.

### Does the first-subscription bonus come back when the order is restored?

Yes, with the order.

### Is a refund request the same as a dispute?

No. A refund request on SceneMixer goes to support and never affects account standing.

Updated November 3, 2026


---

# Checking the Script Summary, and What Re-parsing Means | SceneMixer

URL: https://scenemixer.com/guides/checking-the-summary-and-re-parsing/
What the Script Parsed summary shows, what to check, which mistakes are fixed in step 2, which need an episode re-parse or a new project, and what each costs.

# Checking the Summary, and What “Re-parse” Means

Quick answer: When the reading finishes, step 1 shows Script Parsed: title, protagonist, era, adaptation strategy, how many characters, scenes, props and episodes were found, and an outline per episode. Read the cast and the episode count; the rest is easy to fix later. Wrong names, roles or descriptions are edited in step 2. A wrong episode’s storyboard is redone with Re-parse in step 3. A script that was cut into the wrong episodes, or read with the wrong type or region, is best read again in a new project — the reading is priced by length and the free image allowance is account-wide.

A finished sample started from exactly this screen.

## What the summary shows

Title (the project name, editable), Protagonist, Era, Adaptation Strategy — how the reading intends to turn the text into episodes — and the counts: characters, scenes, items, episodes. Below, the script content, and an outline for each episode with Show more to expand them.

## What to check

The episode count. Ten episodes for a story you expected in three means the type or the split is not what you wanted. The protagonist. If the reading picked the wrong person, the roles in step 2 will be off. The era and strategy. A modern story read as period will dress and stage everything wrong. Everything else — a name, a description, a scene that should not exist — is a step-2 edit.

## Fixed in step 2

Names, roles, appearance, clothing, voices; a missing character or place (add it); a place the story does not need (delete it). None of that needs the script read again. Step 2.

## Fixed with an episode Re-parse

Each episode’s storyboard is written from its outline when you press Parse Storyboard in step 3. If the storyboard is wrong, Re-parse writes it again; it costs the storyboard fee and discards that episode’s segments and videos. What re-parsing an episode costs you.

## Needs a new project

The reading itself: where the episode breaks fall, the region the cast was moved to, the world the story lives in. Those are decided when the script is read, and step 1 does not read it again once it has a summary. Start a new project with the same script and the settings you want. SceneMixer prices the reading by script length — 4 credits per 3,333 characters of English — and your free image allowance is counted per account, so images already made do not cost you the allowance twice. Region · Type.

## When the reading failed

Step 1 shows Previous parsing failed with the reason, and the credits are back. Fix what the reason names — usually the script is too long, too short, or a link — and press Next again. The reading.

| Wrong | Fix where | Costs |
| A name, role, description, voice | Step 2, the card | Nothing (regenerating an image is priced) |
| A character or place missing / unwanted | Step 2, Add / delete | Nothing |
| One episode’s storyboard | Step 3, Re-parse | The storyboard fee; that episode’s videos are discarded |
| The episode split, the region, the type | A new project | The reading fee for the script length |
| The project name | Step 1 or the project card | Nothing |

## FAQ

### Can I read the script again in the same project?

No. Once step 1 has a summary it does not read the script again. Start a new project with the same script for a different split, region or type.

### Do I lose my free images if I start a new project?

No. SceneMixer counts the free allowance per account across projects.

### A character has the wrong name.

Edit it in step 2. Every later shot follows the new name.

### The episodes are cut in the wrong places.

That is the reading. Start a new project, choosing the type you want before pressing Next.

### What does Re-parse in step 3 do?

It rewrites one episode's storyboard from its outline, discarding that episode's segments and videos.

Updated November 3, 2026


---

# Checklist Before You Generate Videos in SceneMixer

URL: https://scenemixer.com/guides/checklist-before-you-generate-videos/
Editing text is free and video is not, so check first: sheets, @ links, shot seconds, states, props, background people, dialogue language, tier, and one test segment.

# A Checklist Before You Generate the Videos

Quick answer: Fixing text costs nothing; regenerating video costs every time. Before you generate an episode, run down this list: every character on screen has a sheet you like; their names are @ pills in the shots where they appear; spoken lines have enough seconds; costume changes use a state; the Props line lists only what the shots show; busy places have a background-people line; dialogue is in the project’s dialogue language; and you are on the tier you mean to pay for. Then generate one segment, watch it, and only then press Generate all.

Open the editor of a parsed episode and check segment 1 against this list before you press anything.

## 1. The sheets

Open step 2 and look at every character, scene and prop the episode uses. A face or costume you do not like will be in every shot; fix it now. Regenerate one sheet · Use your own photo.

## 2. Names linked where people are on screen

In each shot, the characters on screen should be @ pills, and people who are only talked about should not be. If you edited a segment, the editor asks before generating when some shots have no linked characters. Shots with no linked characters.

## 3. Seconds

Each spoken line needs time to be said; a shot that is too short rushes or cuts the line. The segment total must stay under the episode’s ceiling and the tier’s cap. Shot seconds and timing.

## 4. States, props and the room

- Costume changes, injuries and damaged rooms: pick the state for the segments that need it. Use a state in one scene.
- The Props line: only objects the shots show. Props and locations.
- Busy places: one background-people line, the same in every segment of the scene. Background people.

## 5. Language and wording

Dialogue should be in the project’s dialogue language; the editor asks if a segment’s lines do not look like it. Describe what is in the frame rather than what is not. Dialogue not in the project language · Say what is there.

## 6. Tier and balance

Check the model in the editor’s top bar: a low tier for drafts, a higher one for finals. The price on each card and the total on Generate all are what you will pay; make sure the balance covers it. Draft low, finish high · Generate all with a low balance.

## 7. One segment first

Use your free preview if you still have it, then generate one segment and watch it with sound. If the faces, the room and the voices are right, generate the rest. The free preview · Generate all.

| Check | Where | Costs to fix now |
| Sheets look right | Step 2 | An image, if you regenerate one |
| Names are pills where people are on screen | Edit the segment | Nothing |
| Lines fit their seconds | Edit the segment | Nothing |
| States, props, background people | Editor left panel and segment header | Nothing |
| Dialogue language | Segment text | Nothing |
| Tier and total | Top bar and Generate all | Nothing |

## FAQ

### Why not just generate and fix afterwards?

Every regeneration on SceneMixer is charged. Most problems are visible in the text and free to fix before generating.

### Do I need to check every segment?

Check the first one closely; for the rest, look at names, seconds and states quickly. Problems tend to repeat.

### What if the test segment is wrong?

Fix the cause in the text or the sheet, regenerate that one segment, and compare. See when a shot comes out wrong.

Updated November 3, 2026


---

# AI Short Drama Dialogue Language and Cast Localization | SceneMixer

URL: https://scenemixer.com/guides/choose-dialogue-language-ai-drama/
Set the spoken language of an AI short drama in SceneMixer, match cast, names and streets to the target market, and ship one script to many markets, no dubbing.

Guide · Languages

# How to Set the Dialogue Language of an AI Short Drama and Localize the Cast (2026)

Published September 12, 2026

Quick answer: In SceneMixer the dialogue language is a project setting chosen in Step 1, before the script is parsed. Pick the language, pick the target region so that names, faces and streets match the market, then parse. Every storyboard copies the lines in that language and every generated segment performs them. To ship the same story to a second market, create a second project from the same script with a different dialogue language and region — the pipeline does the rest, and you pay per segment exactly as before.

## Step by step

- Paste or upload the script. TXT, DOCX or Markdown, in whatever language it was written; the analysis step reads any of the supported languages.

- Choose the dialogue language. In Step 1, next to the visual style and aspect ratio, open "Dialogue language". The list has 15 languages. The setting takes effect at parsing, so choose before you press "Next".

- Choose the target region. Western, East Asian, Southeast Asian, Latin American, Central Asian, or "follow the script". This decides the cast's names and ethnicity, scene architecture and props, and the language of any text visible in the frame. For a market release, match the region to the dialogue language: Korean dialogue with an East Asian region, Spanish with Latin America, Indonesian with Southeast Asia.

- Parse and review. The synopsis, character list and episode outlines appear in your interface language; the dialogue inside them is already in the chosen language. Read the first episode's lines here — this is the cheapest place to fix a phrasing.

- Generate references, then storyboards, then video. Character sheets carry the localized look; storyboards copy the lines verbatim; each segment is generated with the line performed in the dialogue language.

## What changes with the region, and what does not

| Element | Follows the region | Follows the dialogue language | Stays as written |

| Character names | Yes — local names | — | Names you explicitly keep in the script |

| Cast ethnicity and wardrobe | Yes | — | Costume style you choose in Step 1 |

| Street scenes, architecture, props | Yes | — | Visual style (live-action, 3D, painterly…) |

| Text visible in the frame (signs, screens) | Yes | — | — |

| Spoken lines | — | Yes | Plot, beats, cliffhangers |

| Shot timing | — | Yes — computed from that language's speaking rate | Segment length (10–15 s by default) |

Region and dialogue language are independent on purpose. A story about a Western family can be released with Portuguese dialogue for Brazil without changing the cast, or re-cast for an East Asian market with Korean dialogue; both are one setting each.

## Shipping one script to several markets

- One project per market. The dialogue language is fixed at parsing, so a second market is a second project from the same script. Concept art, storyboards and video are generated per project; there is no shared-asset shortcut yet, and that is what keeps the cast consistent within each version.

- Edit in your language. Whatever your interface language, the storyboards you see are translated for reading and editing; when you change a line, the change is written back to the production script and the segment is regenerated in the dialogue language.

- Keep lines short for German and Ukrainian. These two show an occasional mispronounced ending on long sentences. The storyboard prompt already caps sentence length for them; if you rewrite lines by hand, keep the habit.

- Check the first episode before the batch. Generate episode 1, listen to two or three segments, then batch the rest. Regenerating a single segment costs only that segment.

## Common mistakes

- Changing the dialogue language after parsing. Existing outlines keep their old lines; re-parse the project (or start a new one) after changing it.

- Region set to "follow the script" for a translated release. You get the original market's names and streets with the new language on top — fine for a bilingual audience, wrong for a localized launch.

- Pasting lines in a third language into a storyboard. The submit guard will stop the segment and ask you to confirm; if you meant it, confirm, otherwise fix the line.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### Where do I set the dialogue language?

In Step 1 of the project wizard, in the same form as the visual style and aspect ratio. It applies when the script is parsed, so set it before pressing Next; changing it later requires re-parsing.

### Can one project have episodes in different languages?

No — the dialogue language is per SceneMixer project. For a second language, create a second project from the same script; that is also what keeps each version's cast consistent across episodes.

### Do I need to translate the script myself first?

No. Upload the script in the language it was written in; SceneMixer writes the outlines with the dialogue in the language you chose, and the storyboards copy those lines.

### Will the cast look local if the dialogue is Korean?

Only if you also set the target region to East Asian. Region controls names, faces, wardrobe, streets and on-screen text; the dialogue language controls the spoken lines. Set both for a localized release.


---

# How to Choose Visual Style Settings for an AI Drama | SceneMixer

URL: https://scenemixer.com/guides/choose-visual-style-settings/
Four settings decide how your drama looks: visual style, scene style, costume style and content type. What each one controls and which ones you can still change later.

# Choosing Your Style Settings

Quick answer: SceneMixer asks for four settings before it reads your script, and they are deliberately kept apart because they control different layers. Visual style is the medium — how the image is rendered. Scene style is the set — what the space is built from. Costume style is what people wear. Content type is structural — it sets episode length and pacing. Mixing them into one “style” box is what makes AI video look generic; keeping them separate is what lets a wuxia courtyard and a neon-lit office share the same camera language.

Every sample on the site carries its own four settings. Open a finished project and you can see the choices playing out across the character sheets, the location plates and the shots.

## Visual style — how it is rendered

This is the medium and the surface: photographic, painterly, cel-shaded, 3D. It reaches everything — character sheets, location plates, item shots and the video prompt itself — because it is the one layer that has to stay identical in every frame of every episode.

It does not give framing or composition instructions. Telling the system how to paint is fine; telling it where to put the subject is the storyboard's job, and the two fight if both try.

## Scene style — what the set is made of

Space, materials, colour discipline, how densely the set is dressed. This one is injected in exactly one place: the script read. The model digests it once, then writes it into every location description and every shot, so the courtyard and the corridor come from the same design language rather than being re-styled independently.

It is deliberately kept out of the location images and the video prompt. Pasting it in a second time is how you get red lacquer columns growing in an abandoned courtyard — the description already said what the space is; repeating the label overrides it.

Because it lives at the read, it cannot be changed as a setting afterwards: edit the scene descriptions in Step 2 and regenerate those sheets, or start a new project with the same script.

## Costume style — what people wear

Garment construction: silhouette, cut, fabric, craft. Like scene style, it is injected only at the read, and for a second reason: clothing is already described per character. Adding a project-wide wardrobe paragraph on top flattens the class gradient the read worked out — the servant and the magistrate end up dressed with equal expense.

It also protects a fix SceneMixer made earlier: if wardrobe is pushed into the character sheet at full strength, a long coat eats seventy percent of a vertical frame and the face shrinks to nothing.

## Content type — the structural switch

Short drama, long-form or film. It is not a look at all: it sets the per-episode duration ceiling, the performance grammar and the dialogue rhythm. A short drama needs a hook at the end of every episode; a film does not, and the read and the storyboard both branch on this.

## You do not have to decide alone

Paste the script and the settings are pre-filled for you — a small, free pass over your text picks the three visual axes and says why. It only fills in axes you have not touched yourself, and you can override any of them before the read starts.

All three visual axes have to be set before the read can begin. That is on purpose: they are consumed at the read, so a blank one is a decision made by default rather than by you.

| Setting | Controls | Changeable later? |
| Visual style | Rendering medium and surface | No — fixed once you continue from Step 1 |
| Scene style | What the set is built from | Not as a setting — edit the scene descriptions in Step 2 |
| Costume style | Garment construction | Not as a setting — edit the clothing fields in Step 2 |
| Content type | Duration cap and pacing | No — start a new project |

## FAQ

### Why can I not change the scene style after the script is read?

Because it is consumed once, at the read, and written into every location description from there. Changing the label afterwards would leave the descriptions saying something else. Edit the descriptions in Step 2, or start a new project with the same script.

### Can I write my own style instead of picking a preset?

Members can. Presets cover the common ground; a custom line is useful when you have a specific reference in mind.

### Can I change the visual style later?

You cannot change it in the project: it is fixed once you continue from Step 1, so every sheet and clip carries the same style. To try another one, start a new project with the same script.

### Is there a lighting or mood setting?

No, and that is deliberate. Light position, atmosphere and grading are already decided per shot by the storyboard. A project-wide mood setting would fight every shot that asks for something specific.

### What if I do not know what any of these mean?

Paste your script and take the pre-filled suggestion. It reads your text and explains each pick in one line.

Updated November 3, 2026


---

# AI Create: Choosing the Script Length and What It Costs | SceneMixer

URL: https://scenemixer.com/guides/choosing-script-length-and-cost/
The five length options in AI Create, how many episodes each tends to give, how the writing is priced and where the price is shown, and why some options disappear.

# Choosing the Script Length and Cost

Quick answer: Under Length on the AI Create tab: Mini (1k chars), Brief (2.5k), Short (5k), Medium (10k) and Long (20k). Each option in the menu shows its own price, and the Start Writing button shows the one you picked. A short script gives a few episodes; a long one gives many — the read decides the split. Options that would be too long to read afterwards are hidden. For English and other alphabetic languages the same length is more characters, and the price already accounts for it.

The one-line brief matters more than the length; write that first.

## The five options

Mini, 1,000 characters: one scene, a test. Brief, 2,500: a single short episode or two. Short, 5,000: a handful of episodes, the usual first choice. Medium, 10,000: a short season. Long, 20,000: a full series. The count is the script’s length in characters; the read then splits it into episodes of a minute or two each, so more characters mean more episodes rather than longer ones.

## Where the price is

Open the Length menu in SceneMixer: every option carries a ✦ price. Pick one and the Start Writing button shows it. The writing is charged when the script is delivered; a writing that fails is refunded. The read that follows is priced separately, by the script’s length, and is shown on Start Creating in the result window. Where credits go.

## Why an option is missing

The read has a maximum length. Options whose script would not fit are hidden from the menu rather than sold and then refused. In alphabetic languages the same story takes more characters, so the top option can disappear in English while it shows in Chinese.

## How English counts

The lengths are in characters, but an English script says less per character than a Chinese one. The writing targets the equivalent amount of story, and the price on SceneMixer uses the same rule the read does: 3,333 characters of an alphabetic language count as 1,000. Japanese and Korean count like Chinese. Writing in your language.

## Which to pick

First time: Short. It is enough episodes to see the whole path — read, sheets, storyboard, video — without a long read. If the draft is right, write the Long version with the same brief and a request for more episodes. A brief that already names ten episodes needs Medium or Long to hold them.

| Option | Characters | Gives |
| Mini | 1,000 | One scene, a test |
| Brief | 2,500 | One or two short episodes |
| Short | 5,000 | A handful of episodes |
| Medium | 10,000 | A short season |
| Long | 20,000 | A full series |

## FAQ

### Where do I see the price?

On each option in SceneMixer’s Length menu, and on the Start Writing button once you pick one.

### Does a longer script make longer episodes?

No. It makes more episodes; SceneMixer splits by story, and each episode stays a minute or two.

### Why is Long missing from the menu?

It would be too long to read afterwards in your language, so SceneMixer does not offer it.

### Is the writing charged if it fails?

No. SceneMixer charges only a delivered script.

### Is the read included?

No. SceneMixer prices the read separately by length and shown on Start Creating.

Updated November 3, 2026


---

# Clear or Add Background People, Light the Lamps in an AI Video Clip

URL: https://scenemixer.com/guides/clear-or-add-background-people-and-light-the-lamps/
Empty a street that should be deserted, fill a hall that should be busy, or light every lamp in a clip you already have. How it works and when to fix the storyboard.

# Clear Out or Add Background People, and Light the Lamps, in a Finished Clip

Quick answer: Open Edit this clip → AI visuals. Background people → Clear them out leaves only the set and the characters the shot names; Add a crowd puts ordinary passers-by in the background, out of focus and not looking at the camera. Lights → Light them up turns on every lamp, candle and lantern already in the scene. Each is one edit at 2× the per-second price of the clip’s tier for the length of the clip.

Classic case: a scene written as a private conversation that came out with strangers in the background. Clear them out on that one clip.

## Clear them out

The background becomes empty: the furniture, walls and props stay, and the only people left are the characters your storyboard names for that shot. It works best on people who stand or walk in the background for the whole clip. Someone who walks in half-way may not be removed; for that, regenerate. It will not remove a named character or a duplicate of one — those come from the storyboard. Who is in frame.

## Add a crowd

The background fills with ordinary passers-by going about their business, kept out of focus and never looking into the lens, so they read as extras rather than new characters. Use it for a market, a party, a station or a hall that came out empty. Nobody in the crowd is given lines.

## Light them up

Every lamp, candle and lantern already in the scene is switched on and throws warm light on what is near it. It is made for scenes that have these objects; it is not the way to add new lamps to a room. It pairs well with Time of day → Night, which lights a scene only from its own sources: run Night first, then Light them up on the result. Day to night.

## Step by step

- Select the segment and open Edit this clip → AI visuals.
- For a clip over 15 seconds, choose the half.
- Click Background people and then Clear them out or Add a crowd; or click Lights (its only option, Light them up, is selected for you).
- Press Edit into a new version.

## One clip or the whole scene?

These edits change one clip. If every segment of a scene should be empty — or busy — the storyboard is the better place: each segment has a background people line, and when it is right the segments come out consistent, now and on every regeneration. Use AI visuals for the one or two clips that came out differently. Background people in a scene.

| Option | Result | Not for |
| Clear them out | Only the set and the named characters | People who walk in half-way, duplicates of a character |
| Add a crowd | Out-of-focus passers-by in the background | Extras who speak or act |
| Light them up | Every existing lamp, candle, lantern on | Adding lamps that are not there |

## FAQ

### Will the crowd look at the camera?

No. They are kept in the background, out of focus and going about their business.

### Can I clear the background and light the lamps in one go?

No. One change per edit; run the second on the result.

### Does clearing the background remove my characters?

No. Characters the shot names stay.

### Does it change the next segment?

No. Change the storyboard's background people line for a whole scene.

Updated September 30, 2026


---

# How to Composite and Share a Finished Episode | SceneMixer

URL: https://scenemixer.com/guides/composite-and-share-an-episode/
Turn your segments into one file, decide what the AI label does, and put it behind a link anyone can open. What is in the download versus what plays in the browser.

# Compositing and Sharing an Episode

Quick answer: Once the segments you want are shot, compositing joins them into one file for the episode. It is priced by the total length and the price is shown before you confirm. The finished file can be downloaded at full quality, or shared with a link that opens a real page — no account needed on the other end.

A shared link opens a page like the ones behind our sample write-ups: the video, what it was made from, and a way in. That is the same page your viewers get.

## Compositing

The composite uses whichever version of each segment is currently active, so set your versions before you run it. It joins them without re-encoding wherever it can, which means the picture you approved is the picture that ships.

Every segment has to be finished first. The counter in the top bar of the editor shows how many are done out of how many there are, and compositing is refused until it reads all of them — so a half-shot episode cannot be composited by accident. Recomposite later if you change a segment; the earlier file stays until the new one replaces it.

## Two copies of every video, and why

Every finished video on SceneMixer exists in a full-quality copy and a lighter streaming copy. Downloads and the final composite always use the full-quality one; the player in your browser and on a shared page uses the light one, because a 1080p master is a poor thing to stream to a phone.

You never choose between them. The download button gives you the master.

## Music in the finished episode

An episode that has its own score also gets a version of the composite with the music carried across the whole timeline. Each composite in the records has a Music switch: on — the default — playback and download both use the version with the score; off, playback is the plain composite and the download is the dialogue-enhanced version without music. While that version is still being made, the switch shows in progress and the download gives the dialogue version. Both Music switches (this one and the one by the shot player) are saved to your SceneMixer account. Background music in an AI short drama.

## The AI label

Generated video carries a small AI label in the corner. It is applied by us at delivery rather than by the upstream model, so it is consistent and it does not sit on top of the picture in a way that spoils a shot.

On a free account every video carries it. On a paid membership the label is off by default — you can turn it back on if you want it. Either way there is a control next to the model picker in the editor showing whether the next video you generate will carry one, and changing it only affects videos generated after that. A video shared through a SceneMixer link always shows the label, even on a paid plan; your own download is unaffected. Why shared videos carry the badge.

Where you publish may have its own labelling rules. Turning our label off does not remove your obligation to disclose that something is AI-generated where that is required.

## Sharing

Share gives you a link to a real page — not a raw file. The page carries the video, a short explanation of how it was made, and the usual social preview, so it looks right when pasted into a chat or posted somewhere. A shared link always plays the version with the episode’s music, whatever your Music switch says.

Anyone with the link can open it without signing in. There is no expiry on the page.

The link is derived from the thing it points at, so deleting the segment, the episode or the project makes the page stop working — that is how you take one down. Be aware of what it does not do: the video file itself is not removed from storage at that moment, and anyone who already saved the direct file address keeps it.

The page is also a public, indexable page — search engines are allowed to list it. Treat a share link as publishing, not as a private hand-off.

## Limits worth knowing

Each shared link on SceneMixer has a generous daily play allowance. Past it the page still loads and still reads correctly — it just stops streaming the file for the rest of the day. That is a cost guard, not a penalty, and ordinary sharing never comes near it.

If you are pushing a video to an audience rather than to a few people, download the file and upload it to wherever that audience is.

| Thing | What you get |
| Composite | One file for the episode, priced by total length |
| Download | Full-quality master |
| In-browser playback | Lighter streaming copy |
| AI label | Always on for free accounts; off by default on a paid membership |
| Share link | A public page, no sign-in, no expiry |
| Music | Each composite gets a version with the episode's score; playback and downloads follow the Music switch |

## FAQ

### Does compositing cost credits?

Compositing on SceneMixer is priced by the total length of the finished episode and the price is shown on the button before you confirm.

### Can I composite before every segment is shot?

No. SceneMixer refuses compositing until every segment in the episode has a finished video; the editor's top bar shows the count.

### Can I recomposite after changing a segment?

Yes. Set the segment version you want on SceneMixer and run it again; the previous file stays until the new one replaces it.

### Does the downloaded composite carry the music?

Yes by default: the composite’s Music switch is on, and the download gives the version with the episode’s score. Switch it off on SceneMixer for the dialogue-enhanced version without music. Shared links always play the scored version.

### I only want one complete video, not episodes. How?

Composite each episode on SceneMixer once all its segments are done, then join the files in any video editor. For a single file from the start, keep the script short enough to be one episode.

### Do viewers need an account to watch a shared link?

No. The link opens a public SceneMixer page that works without signing in.

### How do I take a shared video down?

Delete the segment, episode or project it points at in SceneMixer. The link stops working immediately.

### Can I remove the AI label?

On a paid SceneMixer membership it is already off by default; free accounts always carry it. Check your local rules before you publish — some jurisdictions require AI-generated content to be disclosed regardless of what the file carries.

Updated November 3, 2026


---

# Composite an Episode: Which Segments Are Joined, and the Missing Ones

URL: https://scenemixer.com/guides/composite-which-segments-are-joined/
An episode composite joins every segment that has an active clip, in order. Which segments are skipped, why a composite is shorter than expected, and how to redo it.

# Compositing an Episode: Which Segments Are Joined, What Happens to Segments Without a Clip, and the Price

Quick answer: Composite joins the segments of the episode in timeline order, taking each segment’s active version, into one file. Every segment needs a clip first: the dialog tells you how many are done. It costs 1 credit per second of the episode’s total length and can be run again after you regenerate segments; each composite is kept as its own record. To leave a segment out, delete it or turn it into something you want — there is no hide switch.

Press Composite on an episode with one segment still ungenerated: the dialog shows how many segments are done instead of starting.

## What gets joined

The active version of each segment, in the order of the timeline, cut hard together. Sound and dialogue come with the segments; nothing is re-rendered, so the composite is quick and its price is small. One composite at a time per episode: a second press while one is running is refused until it finishes. How the result is delivered and shared: Composite and Share an Episode.

## Segments without a clip

A segment that has never been generated, or whose only versions failed, blocks the composite on SceneMixer; the dialog shows how many segments are done, so check the cards to find it. Generate it, or delete it if the episode reads fine without it. A segment that is generating at the moment also blocks; wait for it. Deleting a segment removes its versions and does not refund its generations: Insert or Delete a Segment.

## Leaving a segment out on purpose

There is no per-segment hide switch. If a segment should not appear, delete it (the storyboard text is lost with it — copy it first if you want it back later) or rewrite it into a shot you do want and generate that.

## Price

SceneMixer charges 1 credit per second of the finished episode, whatever tiers the segments used. A ninety-second episode costs ninety times that rate to composite. Recompositing after changing a segment is a new composite at the same rate. Where Your Credits Go in One Episode.

## The file you get

The composite is an original-quality file, plus a lighter play version for the browser and the share page. Each segment carries the AI badge setting it was generated with; switching the badge off later does not remove it from segments already made, so regenerate those first if you want a clean file. Download: the Original vs the Play Version and The AI Badge Switch.

| Segment state | Composite |
| Has an active clip | Joined |
| Never generated / all versions failed | Blocked |
| Generating right now | Blocked; wait |
| Deleted | Not part of the episode |

## FAQ

### Can I composite only segments 1–5?

Not directly. Delete the segments you do not want first in SceneMixer, or rewrite them.

### Which version of each segment is used?

The active one on SceneMixer at the moment you press composite.

### Do I pay per segment or per second?

SceneMixer charges per second of the whole episode.

### Can I composite twice?

Yes. Each composite on SceneMixer is kept; regenerate a segment and composite again.

### Why is the composite blocked when all cards show thumbnails?

A thumbnail can belong to a failed version. Check each card's status on SceneMixer; one is not Done.

Updated November 3, 2026


---

# Content Rules and Moderation: Blocks, Pauses, Suspensions | SceneMixer

URL: https://scenemixer.com/guides/content-rules-and-moderation/
What you cannot generate or upload, where the checks run, what a block costs, how repeated blocks pause generation, when an account is suspended, and how to appeal.

# Content Rules and Moderation: What Is Not Allowed, How Blocks Work, When Accounts Are Suspended

Quick answer: Sexual content, terrorism and extreme violence, drugs and gambling, content that defames or exposes real people, unauthorised film, anime and game IP, and imitations of a real person’s face or voice cannot be generated or uploaded as references. Every generation passes the provider’s moderation; a blocked attempt is normally not billed; repeated blocks in a short time pause video generation; deliberate or repeated violations, getting around moderation, tampering with the AI label or using several accounts to dodge a block lead to removed content, restricted features and account suspension; a balance left on a suspended account is handled under the Terms.

The rules live in the Terms of Service, section 04. This page explains how they play out while you work.

## What is not allowed

The Terms, section 04, prevail; in everyday words: illegal content — sexual or pornographic material, terrorism and extreme violence, gambling, drugs, and anything else the law forbids; harm to real people — defamation, privacy violations, and using someone’s likeness or voice without consent (uploading their photo as a character counts); unauthorised IP — copyrighted film, anime and game characters and footage; the provider runs an “IP suspected” check, and celebrity faces, well-known brands and characters trigger it; evasion and tampering — getting around moderation with respellings or code words, removing or altering the AI-generated label, scraping, and using multiple accounts to dodge a sanction. Any sexualisation of minors is zero tolerance regardless of context.

## Confirming your rights when you upload

The first time you upload a file, a window titled Before you upload asks you to confirm that you have the right to use the material (scripts, images, video, audio) and that every real person whose face or voice appears in it has given consent. Press Confirm and upload to go on; you only need to confirm once, several files in one go ask once, and cancelling stops the upload.

## Where moderation happens

Moderation runs at the model providers, on every generation. For character, scene and prop sheets your description text is checked first, then the image; a block usually names a word in the description. For video, the prompt, the reference images and the clip are checked; the card says the content was blocked by the safety rules, and hovering it shows the provider’s original message, which may name an IP check. Parsing, AI Create and translation check the text. And people look too: feedback, complaints and spot checks. The checks are imperfect — ordinary fight scenes, medical or historical material is occasionally blocked and passes after a rewording. One block is not a finding against you; repeated blocks are the signal.

## What happens after a block

This attempt: refunded automatically when the provider confirms the block, up to 12 blocks in 24 hours; held for review past that or when the outcome is uncertain — see Refunds. Repeated in a short time: more than 12 blocks within 24 hours pause video generation on SceneMixer for 1 hour; the count does not reset inside those 24 hours, so each further block pauses again. This stops moderation being used as a lottery. Parsing, sheets and storyboards keep working during a pause. Serious or deliberate: content is removed, features are restricted, the account is suspended (a balance left on it is handled under the Terms, sections 03 and 09), and we report to the authorities where the law requires it. Credits and membership left on a suspended account are not refunded. That decision is made by a person looking at the generation history, never on a single false block. Real people in uploaded photos: the ad flow removes people from a photo before using it; character and scene photos must be material you have the rights to, and substantiated complaints are handled under the Terms.

## How to fix a blocked generation

Read whether the message names the text or the image. For text, edit the description: remove explicit, gory or real-person terms, brand names and IP characters, and replace “result words” with neutral description of what the camera sees. For an image, the description usually caused it — edit the description and regenerate. For a blocked video, change that segment’s storyboard text or the sheets it uses; resubmitting unchanged only gets blocked again and counts again. For “IP suspected”, replace celebrity faces, known characters and brand marks with original designs; realistic styles hit celebrity likeness more often, so change the style or the appearance text. Subjects that cannot be reworded — heavy gore, sexual content, real political figures — are not made here. Step-by-step rewrites: Blocked by Moderation: What to Change.

## Appealing a block or a suspension

Open Ask AI; if the answer does not solve it, use Still not solved? → Send feedback, category Other, with the project, episode and segment, what you wrote and the message you saw; a person reviews it. A false block gets its held credits returned; the pause itself simply runs out. A confirmed violation is handled as above and you are told why. A suspended account cannot use the site, so a suspension appeal goes by email to support from the account’s email address. How to get help.

| Where | What is checked | What you see when blocked |
| Character, scene and prop sheets | Description text first, then the image | text check failed / image check failed |
| Video generation | Prompt, reference images, the clip | Blocked by content safety rules; hover for the provider’s message |
| Parsing, AI Create, translation | Text | input contains sensitive content |
| People | Feedback, complaints, spot checks | A notification or an email |

## FAQ

### Does one block go on my record?

No. Occasional blocks in normal work are common. Only repeated blocks in a short time pause generation, and only deliberate violations are sanctioned.

### Can I use other features during a pause?

Yes. Only video generation pauses; parsing, sheets and storyboards continue.

### Can I use my own photo for a character?

Yourself, or someone who gave you permission. Not someone else’s face.

### What about fan works or remakes of popular IP?

Not without authorisation. The provider blocks it and repeated attempts are sanctioned.

### Are credits refunded after a suspension?

A balance left on a suspended account is handled under the Terms, sections 03 and 09.

### Can I remove the AI label from my clips?

Members can switch off the visible SceneMixer badge on their own clips. Removing or altering AI marking yourself is not allowed.

Updated November 3, 2026


---

# Copy and Paste Storyboard Text Without Losing Character Links

URL: https://scenemixer.com/guides/copy-and-paste-storyboard-text-without-losing-links/
Copy shots between segments and the @ pills come back; paste from a document and check one message after saving. How to tell names are linked, and how to fix them.

# Copying and Pasting Storyboard Text Without Losing the Character, Scene and Prop Links

Quick answer: Inside the editor, copying keeps the links: copy text from a segment’s storyboard or from the edit box, paste it into an edit box, and the coloured @ pills come back with it. Text pasted from somewhere else — a document, a chat, a note — arrives as plain words. When you save, names that can be matched are linked again, and a message says Names in the text were linked to their characters, scenes and items. If that message does not appear, or a name still shows as plain text, insert it with @ yourself.

Copy one shot from segment 1, paste it into segment 2 in edit mode, and look: the names arrive as pills.

## Inside the editor: pills come back

Select text in a segment’s storyboard or in the edit box and copy it. Paste it into any segment’s edit box in the same project and the character, scene and prop pills come back, along with shot markers and voice tags. This is the safe way to move a shot from one segment to another, or to duplicate a line of dialogue.

## Text from outside the editor

Copy the storyboard into a document, rewrite it there, and paste it back: the pills are gone and the names are plain words. Save anyway. If the editor can match names to entries in the project, SceneMixer links them again and shows the message about linked names. Read the shots after saving: pills mean linked, plain words mean not. Rewritten sentences are the ones most likely to stay plain; link those names with @.

## How to check a segment

- Open it with Edit: every character on screen in a shot should be a pill.
- An amber note above the storyboard after saving can list names written without @.
- Before generating, the editor asks if some shots have no linked characters.

The amber notes · Shots with no linked characters.

## Linking a name by hand

Delete the plain name, type @ in its place, and pick the entry from the menu. Scenes go on the Location line, props on the Props line and in the shot where they are handled. The @ menu.

## Copying out

Text copied out of the editor into a document is readable plain text, which is handy for review or for sharing a storyboard with someone. Keep a copy of a segment before big edits; the editor also keeps a text history you can restore. Text history.

| You paste | Into | Links |
| Text copied in the editor | An edit box in the same project | Come back as pills |
| Text from a document, unchanged | An edit box | Usually linked again on save; check the message |
| Text you rewrote outside | An edit box | May stay plain; link with @ |

## FAQ

### Why did my pasted names turn into plain text?

The text came from outside the editor. Save, then look for the linked-names message; link any plain names with @.

### Can I copy shots between projects?

The text pastes, but pills only link to entries that exist in the project you paste into.

### Does linking names cost anything?

No. Editing and saving text on SceneMixer is free.

### I pasted a whole segment and the video had no faces. What happened?

The names were plain text, so no sheets were used. Link them with @ and regenerate the segment.

Updated November 3, 2026


---

# Costume Changes and Scene States in AI Video | SceneMixer

URL: https://scenemixer.com/guides/costume-change-and-scene-states/
When a character changes clothes or a room burns down, a sentence is not enough. What a state image is, when you need one, and when plain text does the job.

# Costume Changes and Scene States

Quick answer: A model given a reference image copies what is in it — the face, the clothes, the shape of the object — and no amount of text will talk it out of that. So anything that changes the face, the clothes, the form or the structure needs its own reference image, which we call a state. Anything that changes where something is, who is holding it, the light, the weather or what it is covered in is written in words and works fine.

States show up as cards next to the main asset in the project view. They are generated from the script read when the story calls for one, and you can add your own — the asset library in this sample project is where they sit.

## The dividing line

Ask one question: does the thing itself look different, or does the situation around it look different?

A full change of outfit, a transformation, burns after a fire, a collapsed building, a blade out of its scabbard, a snapped staff — the object itself is different, so it needs an image.

Standing versus sitting, held versus on the table, night versus day, in the rain, covered in dust — these ride on top of whatever the reference shows, and the model applies them from text without trouble. Rain in particular is text, not a state.

One more test: if the two versions could appear in the same frame at the same time, they are not states of one thing — they are two separate assets.

## Where states come from

On SceneMixer, the script read proposes them when the story needs them, within tight caps: at most three looks for a character, two for a location, one for a prop. The caps exist because an unbounded read produces states for everything — a cup that is full, a cup that is empty, a lamp that is lit — and none of those change what the object is.

You can also create, rename and delete states yourself, and choose which episodes each one applies to.

## How a state image is made

It is an edit of the base image, not a fresh render. The instruction is four short lines: the state name, what to change, one line holding identity steady, one line for the frame. Nothing else — no style string, no description template.

That brevity is the whole trick. A model given a reference image and a three-thousand-word template reads it as “redraw this reference” and hands back the original almost unchanged. Four lines of edit instruction gets you the change you asked for.

## Which image a shot actually uses

Priority runs from most specific to least: a state you picked for this particular segment, then a custom image you set for this episode, then a state assigned to the whole episode, then the base image. Only states that have actually been generated count — if anything is ambiguous it falls back to the base.

Which segments use which state is worked out automatically after the storyboard lands, per segment rather than per episode. Per-episode would be wrong: a flashback look applied to a whole episode puts the character in flashback clothes in every single shot of it.

## If the state has no image yet

Nothing breaks. The shot uses the base image and the state is described in words instead — the name and the difference are written into the prompt. For a character that replaces the wardrobe phrase while keeping the facial description; for a location or prop it is added to the material's role line.

Generating the image later gives you the stronger version, and shots generated after that will use it.

| Change | State image or text? |
| Full change of outfit | State image |
| Transformation, burns, scars | State image |
| Building collapsed or burnt | State image |
| Blade drawn, object broken | State image |
| Standing, sitting, kneeling | Text |
| Held, dropped, on the table | Text |
| Night, rain, dust, firelight | Text |

## FAQ

### Why is rain not a state?

Rain is applied on top of whatever the reference image shows, and models handle that from text. A state image is for when the subject itself is a different shape.

### Can I have a state for a character standing and one sitting?

You can, but you should not — both could be in the same frame, so they are not two states of one thing. Posture is written into the shot.

### Do states cost extra?

They are priced like any other setup image, and on your first SceneMixer project each type has its own small free allowance, separate from the one for base images.

### What if I regenerate the storyboard?

Per-segment state picks are recalculated. Episode-level assignments are not affected.

### Do I have to generate a state before I can use it?

No. Without an image the state is described in words, which is weaker but works. Generate it when you want the change to be reliable.

Updated November 3, 2026


---

# Create AI Videos From Text: One Sentence, Some Pages or a Novel

URL: https://scenemixer.com/guides/create-ai-videos-from-text/
One sentence, a few pages and a whole novel take different paths to video: what each becomes, where the length limits are, what the reading costs and what to cut first.

# Create AI Videos From Text

Quick answer: To create AI videos from text you do not need a particular kind of text, but the amount of it decides the path. One sentence is written into a script first. A few pages are read as they are, in one pass. A long story or a novel is read in two parts, the cast first and then the episodes, up to 300,000 English characters. Whichever you bring, it ends up as the same thing: a cast, a set of locations, an episode list, a storyboard of numbered shots and then video. Here is each path, the limits on it and what to cut before you paste.

## Three sizes of text, three paths

| You have | It goes in as | What happens | You get |
| One or two sentences | AI Create brief | A script is written to a length you pick, from Mini (about 1,000 counted characters, roughly 3,300 in English) to Long (20,000 counted), then read | A script to review, then a project |
| A paragraph up to a long chapter | Upload Script: paste, or .txt, .docx, .md | Read in one pass, up to 10,000 Chinese-equivalent characters (about 33,000 English characters) | A summary, a cast, episodes |
| A story or novel | Upload Script | Read in two parts up to 100,000 equivalent characters (300,000 English) | The cast first, then the episode list |

“Equivalent” matters in English: SceneMixer counts every length in Chinese-equivalent characters, and 3,333 characters of English count as 1,000. The floor is about a paragraph, roughly 330 characters of English; below it there is nothing to split into episodes. Above the ceiling the read is refused before you pay, rather than silently cut. How characters are counted · What you can feed in.

## One sentence: say five things

A brief that carries the genre, the lead and what they want, the conflict, the ending and the tone comes back as a better script than a longer brief that does not. Pick the length by how much story you want: a longer script gives more episodes, not longer ones, because the read splits by story and each episode stays a minute or two. The script is written in the language your interface is in. Writing the one-line brief · AI story generator.

## A few pages: prose or script, both work

You do not have to format a script. Prose is read, and so is screenplay layout. What changes is how much of your wording survives:

| You write | The storyboard gets |
| Mei (quietly): I saw him. | The line, spoken quietly, on Mei’s shot |
| She told him quietly that she had seen him. | A summary of it; nothing is spoken |
| Father (on the phone, hoarse): Come home. | The line, heard but with the speaker not shown |
Write what must be said as lines, and leave the rest as prose. Paste the text itself: a link is not fetched, and a table or a list is not a story and can fail repeatedly. Write dialogue so it is picked up.

## A novel: why it is read twice

A single SceneMixer read has a fixed budget for what it can write back, and a fifty-episode book does not fit: read in one go, it would deliver a fraction of the episodes without saying so. So from 10,000 equivalent characters the read runs in two parts. The first finds the characters, locations and props. The second builds the episode list with a short outline for each. The first three episodes are prepared in full straight away; any other is prepared when you press Generate storyboard on it, and the button waits and then carries on by itself. You pay the read once, by the length of the text, however it is read.

A text over the ceiling is split into volumes, each its own project, and Pick existing brings a returning character across. Long scripts are read in two passes. For scale: a 30,000-character novel read as ten episodes of 90 seconds at the lowest video tier comes to about 10,026 credits, about $62 at the lowest credit price, nearly all of it video.

## How long the video will be

Text length does not set video length in SceneMixer. The story does: the read splits it into episodes, a Short Drama episode runs up to 240 seconds, a TV Series episode up to 300, and a Film has no cap. Within an episode, dialogue sets the seconds, at about three English words a second. Every segment is up to 15 or 30 seconds, and video is bought by the second, so the budget follows the finished length. Short Drama, TV Series or Film · 15 or 30 second segments.

## What to cut before you paste

- Backstory paragraphs. A camera cannot show them; keep them as one voice-over line or drop them.
- Extra locations. Each place is drawn once and reused, so fewer is steadier and cheaper.
- Extra named characters. SceneMixer gives each a picture and a voice. Introduce a nickname once (“Elizabeth, called Liz”) so the read matches it to the full name.
- Other people’s characters. The IP check refuses pictures and clips that resemble famous characters and real people.
- Everything after your first scene, for a first test. Make the opening alone, watch it, then decide how much more to make.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Classical Elegance, Mediterranean Coast; Costume Style Haute Couture, Minimal Luxury; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the leads with one mark each and a voice; the two or three places as locations; the object the plot turns on as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

## FAQ

### How do I create AI videos from text?

Paste your text into SceneMixer, or write one sentence on AI Create. It is read into a cast and episodes, you check the storyboard, and then video is generated segment by segment.

### How much text do I need?

At least about a paragraph, roughly 330 characters of English. One sentence is enough on AI Create, which writes the script for you.

### What is the longest text it can read?

Up to 300,000 English characters in one project, read in two parts. Beyond that, split it into volumes.

### Does the text have to be a script?

No. SceneMixer reads prose and screenplay layout alike. Dialogue written as lines is kept word for word; speech inside narration is summarised.

### Can I paste a link instead?

No. SceneMixer does not fetch links. Paste the text itself.

### How long will the video be?

It follows the story, not the text length. A SceneMixer Short Drama episode runs up to 240 seconds, a TV Series episode up to 300, and a Film has no cap.

### What does it cost to turn a text into video?

SceneMixer charges 4 credits per 1,000 characters for reading; video is 10 credits a second on the default tier, and joining is 1 credit a second.

### Can I try it without paying?

Yes. Reading costs a few credits from the signup balance, the first 15 setup images and the first storyboard are free, and one preview shot of up to 5 seconds is free.

Updated November 2, 2026


---

# Create an AI Video: Your First 30 Minutes, Minute by Minute

URL: https://scenemixer.com/guides/create-an-ai-video/
What your first half hour of making an AI video looks like: a minute-by-minute timeline from brief to first shot, what is free, what locks, and when the first clip lands.

# Create an AI Video: Your First 30 Minutes

Quick answer: Your first AI video takes longer than 30 minutes to finish and much less than 30 minutes to start. In the first half hour you can write a brief, have it read into a cast, see the cast drawn, read an episode’s storyboard, watch a five-second preview made from your own cast, and have a first segment rendering, while spending only the reading fee. The first finished clip arrives around minute 40 to 50, because one segment takes ten to twenty minutes on its own. The timeline below is built from the timings the product itself gives, and it marks what is free, what locks and what to do while you wait.

Try it: set a timer when you press Get Started, note the minute each of the six checkpoints below appears on your screen, and compare. A busy hour stretches the waits; the order never changes.

## The half hour, checkpoint by checkpoint

| Minutes | You do | The site does | It costs |
| 0–5 | Pick a tab. For a story, paste a page; for an idea, type one line on AI Create and choose the Mini length. Then set up the project: name, three styles (or Let AI choose), 9:16 or 16:9 | On AI Create, writes a script from your line | The writing, priced by length and shown on the button, if you started from a line |
| 5–10 | Press Next. Leave the page if you like; read the summary at Script Parsed | Reads the text: finds characters, scenes and props, splits episodes, shapes the style. A few minutes | 4 credits per 1,000 characters |
| 10–16 | Step 2: read each description, fix any that is wrong, generate the pictures | Draws the cast. About a minute and a half a picture, 3 at a time on a free account | The first 15 pictures on an account are free |
| 16–22 | Step 3: press Parse Storyboard on episode 1, then read every shot and fix the text | Cuts the episode into segments of three to five shots. A few minutes | Free on your first drama; editing is always free |
| 22–26 | Open the editor and press Generate a free preview | Makes about 5 seconds of one shot, from your own cast, at the lowest tier | Nothing. One per account |
| 26–30 | Leave the tier on the default, press Generate on segment 1 | Holds the credits and starts the model. Typically ten to twenty minutes | 10 credits a second, held, spent when the clip arrives |

The sums behind the windows are SceneMixer’s own timings added up, not a promise. As an example, ten setup pictures at three at a time make four rounds of about a minute and a half, so roughly six minutes; a story with a smaller cast is quicker. While the script is being read · Generate the storyboard for an episode · The free preview.

## Two things you cannot change once you press Next

SceneMixer’s visual style shapes every picture and every video, so it is locked when you continue past step 1. The aspect ratio is fixed too: 16:9 landscape by default, 9:16 for a vertical feed. If you want the same story in the other shape, start a new project with the same script; the cast can come across from the asset library. Everything else in the first half hour can be edited later: names, descriptions, pictures, the storyboard text, and the language the characters speak. Choosing visual style settings · Vertical or landscape.

## What to do while the minutes pass

- During the read: nothing. You can browse elsewhere while SceneMixer works; a pill in the corner holds the progress and the project moves on by itself.
- During the pictures: read the descriptions of the characters you have not drawn yet. A wrong age or an outfit that does not match the story is a free edit now.
- During the storyboard: write the brief for your second video. The cast and the storyboard of the first are what teach you how to write it.
- During the first segment: read the next segment’s text and check that each shot says who is in it with an @ chip, because a shot that names nobody gives the model nothing to hold.

## What the first half hour costs

If you paste your own page, SceneMixer’s only charge is the reading, 4 credits per 1,000 characters, where 3,333 characters of English count as 1,000. If you start from a line, add the writing, which is priced by the length you pick. The signup gift is 50 credits. Free in the first half hour: the first 15 setup pictures, episode 1’s storyboard on your first drama, one preview of up to 5 seconds, every edit to text, and switching between takes. Video is the first thing that is charged, from the first second: 10 credits a second on the default tier. What you get for free.

## The four first-time mistakes

- Starting every segment at once. Shoot segment 1 in SceneMixer, check the faces and the room, then use Generate all. Six segments made from a wrong first shot are six regenerations.
- Choosing the Long script length when a test is what you need. A Mini line costs little and teaches the whole path.
- Picking the style in a hurry. It is the one thing you cannot edit later. Open a sample project in the style you are considering first; each sample is labelled with its style.
- Regenerating the storyboard to fix a line. That discards the finished video in the episode. Edit the line instead, which costs nothing. What you lose.

## If you only have ten minutes

Do minutes 0 to 10 and stop. The SceneMixer project card on the home page says Step 1/3 or 2/3 and reopens exactly where you were. Come back after the cast is drawn, which is the part that needs your eyes, and do the rest in a second sitting. Project cards. If you want the five-step method rather than the clock, How to create an AI video has it.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style American Small Town; Costume Style Soft Everyday; Visual Style Realistic HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two neighbours with a description each; the porch and the kitchen as locations; the casserole dish as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Checkpoint | What you hold | What you have spent |
| Minute 10 | A summary, an episode list, and a cast, places and props in words | The reading fee |
| Minute 16 | The cast drawn | Nothing, within the first 15 pictures |
| Minute 22 | Episode 1’s storyboard, editable | Nothing, on your first drama |
| Minute 26 | A preview of about 5 seconds, from your own cast | Nothing |
| Minutes 40–50 | The first finished segment | 10 credits a second for its seconds |

## FAQ

### How long does it take to create an AI video?

Starting on SceneMixer takes about 30 minutes: a brief, a reading, a cast and a storyboard. The first finished segment arrives around minute 40 to 50, because one segment typically takes ten to twenty minutes.

### How long does the reading take?

A few minutes for a typical script, longer in proportion for a long one. You can leave the page and the project moves on by itself.

### Can I make an AI video in 30 minutes?

You can have everything except the video in 30 minutes: cast, storyboard and a five-second preview. A finished 15-second segment takes ten to twenty minutes more.

### What does the first half hour cost?

SceneMixer charges only the reading fee if you paste your own page, plus the writing fee if you start from one line. The first setup pictures, the first storyboard and one short preview are free.

### What cannot be changed after step 1?

The visual style and the aspect ratio. Names, descriptions, pictures, the storyboard text and the dialogue language can all be edited later.

### What should I do while it generates?

Read the next descriptions and storyboard text, and write your next brief. Editing text costs nothing, and you can close the page without losing the work.

### Should I generate everything at once?

Not the first time. Generate one segment, check the faces, the room and the line, and then start the rest.

### Can I stop after ten minutes and come back?

Yes. The SceneMixer project card shows how far you got and reopens at that step.

Updated November 2, 2026


---

# Create Videos Using AI: The 20 Decisions and Who Makes Each

URL: https://scenemixer.com/guides/create-videos-using-ai/
Twenty decisions go into one AI video. Which the AI makes, which it proposes for you to confirm and which only you can make, and where each happens in SceneMixer.

# Create Videos Using AI: Who Decides What, in 20 Decisions

Quick answer: Creating a video with AI is twenty small decisions, and the useful question is who makes each one. In SceneMixer the AI decides eight and lets you override them; it proposes six and waits for you to confirm; and six are yours alone. The six that are yours include the ones that cannot be changed later, so they are worth the most thought. Every step that costs credits starts with a button you press, with its price on it. This guide lists all twenty, says where each happens and how to override it, and ends with the order to check them in.

## How to read the ledger

Three kinds of decision, in the order a project meets them. The AI decides: it makes a choice from your script and you can change it. The AI proposes: it fills something in or drafts it, and you confirm. Only you: nothing is filled in for you, or the choice is yours by nature. Other guides sort the same project differently: Create videos with AI by what you have, and Video AI by which models do which job. This one sorts it by who holds the decision.

## The AI decides, and you can override (8)

| # | Decision | What the AI does | How you override it |
| 1 | Where the episodes break | The reading cuts your script into episodes by story: more script gives more episodes, not longer ones | Cut the script yourself before the read. Afterwards the split is fixed; a different split is a new project |
| 2 | Who is in the cast | Anyone who acts, speaks or is handed something becomes a character with a role: Lead, Supporting or Villain. A part heard and never seen is marked Off-screen | Add, delete or relabel in step 2. Adding a character the script did not produce needs a paid plan |
| 3 | What each person looks like | Writes an appearance and clothing for every character | Edit the text and regenerate the sheet |
| 4 | Which places and props get pictures | Lists the places and the objects that matter to a shot | Add or delete entries in step 2 |
| 5 | The shot list | Writes each episode as segments of shots, with a size, an angle, the line and the seconds | Edit the text (free), change shot seconds, insert or delete a segment |
| 6 | How each person sounds | Writes a voice description per character; by default the video model performs it | Ask for a free voice match and attach a preset you can hear, or upload a voice |
| 7 | The music | Plans a score for the episode once its first shot exists, picks or writes the pieces and mixes them under the voices | Only the Music switch beside the player. There is no track picker or volume control |
| 8 | Joining the segments | Joins the finished segments into one episode file | You press Composite once every segment has a clip |

## The AI proposes, and you confirm (6)

| # | Decision | What the AI proposes | Where you confirm it |
| 9 | The three looks | Pre-fills the visual, scene and costume styles that you have not touched, tagged as AI’s choice | Step 1. All three must be set before the read starts |
| 10 | The script, if you have only an idea | AI Create writes a draft from your line | You read and edit it in a window before it is read |
| 11 | The summary | After the read, step 1 shows the title, protagonist, era, strategy and the counts of characters, places, props and episodes | You check the cast and the episode count |
| 12 | The cast pictures | Draws a sheet for each character, place and prop | Accept, regenerate, AI edit or upload your own. You can continue without making all of them |
| 13 | The next episodes | Let AI continue drafts one to three episodes from where the story stands. AI story continuer | You read and edit the draft, then add it |
| 14 | The cover | Pre-selects a genre mood and your leads, and paints three poster options | You choose one |

## Only you (6)

| # | Decision | Why it is yours | Where |
| 15 | What the story is | An idea, or a script, and how many people and places. Fewer, steadier characters hold up better | The first tab you use |
| 16 | Type: Short Drama, TV Series or Film | It sets how the script is cut, the length cap of an episode (240 s, 300 s, none) and the acting grammar | Step 1. Fixed when the script is read |
| 17 | The shape: 16:9 or 9:16 | The shape of every shot | Step 1 |
| 18 | Dialogue language and the Target Region | The language the cast speaks, and where the names, faces and places are set | Step 1. Fixed when the script is read |
| 19 | Quality and spend, segment by segment | Each segment is priced by seconds and tier. The cheapest tier starts at 10 credits a second | The model button on every segment card |
| 20 | Which takes to keep | You watch each clip, regenerate a failed one, or fix it with Edit this clip: trim, change the visuals, redub | The episode editor. Each segment keeps a version history |
In SceneMixer, decisions 16 and 18 are the two that cannot be taken back once the script is read, so make them on purpose. Changing the region, dialogue language or type after the script is read says what to do if you chose wrong.

## The ones that lock early

| Decision | Fixed when | To change it |
| Type (16) | The script is read | A new project with the same script |
| Dialogue language and Target Region (18) | The script is read | Edit descriptions and lines, or a new project |
| Visual style | You continue past step 1 | A new project |
| Scene style and costume style | The script is read; they are written into the descriptions | Edit the descriptions in step 2 and regenerate the sheets |
| Cast, props, shot text, tier, takes | Never | Any time |
That is why the first four rows get your attention first and the last row can wait. Why scene and costume style are chosen before the read explains the third and fourth.

## The order to check them in

The cheapest place to find a wrong decision is before the expensive step. Four checkpoints cost little or nothing:

- The summary (11). Is it the right protagonist, era and number of episodes? SceneMixer prices the read by length, 4 credits per thousand characters on the default.
- The cast pictures (12). Do the faces look like the people you meant? The first 15 setup images on an account are free.
- The storyboard text (5). Read the shots and fix the lines. Editing the text is free, and episode 1 of the first drama has a free storyboard.
- The preview. One shot of about 5 seconds with your own cast, on the cheapest tier, once per account. A new account also starts with 50 credits.
Only after those does video start, at the price on each card. The longer checklist is Checklist before you generate videos, and Where credits go in one episode adds up a whole episode.

## FAQ

### How do I create videos using AI?

Start with a script, an idea, product photos or a clip. SceneMixer reads it into a cast, places and shots, you check the cast and the text, and the video is made segment by segment and joined.

### What does the AI decide for me?

SceneMixer decides the episode breaks, the cast list and roles, how people look, the places and props, the shot list, the voice descriptions, the music and the joining. You can change most of them.

### What can only I decide?

The story, the Type, the shape, the dialogue language and region, the quality and spend per segment, and which takes to keep.

### Which decisions cannot be changed later?

Type, dialogue language and region are fixed when the script is read. SceneMixer fixes the visual style once you continue past step 1.

### Can I change the music?

SceneMixer has no track picker or volume control. A Music switch beside the player turns the score off in playback.

### Do I have to approve the AI’s choices?

For the looks, the script draft, the summary, the pictures, the next episodes and the cover, yes: they are proposals. Nothing that costs credits starts without you pressing a button.

### Can I check anything before paying for video?

Yes: the summary, the cast pictures, the storyboard text and a short preview, all free or covered by the free allowances.

### How long does one video take to make?

On SceneMixer, the reading takes minutes and each segment runs on the server while you do other things. Closing the page does not stop a generation.

Updated November 2, 2026


---

# Create Videos With AI: Pick the Route That Fits What You Have

URL: https://scenemixer.com/guides/create-videos-with-ai/
Six things you might hold (a script, an idea, product photos, a clip, a photo, a voice): the route for each, what the first step costs and what it will not do.

# Create Videos With AI: Start From What You Already Have

Quick answer: The quickest way to create videos with AI is to start from the material you already hold, because each kind of material has its own door. A script or a story goes in through Upload Script. A one-line idea goes through AI Create. Photos of something you sell or own go through Ad Video. A clip you are allowed to use goes through Remake a video. A photo of a person, a room or an object goes onto a card inside a drama, and a voice recording goes onto a character. The doors differ in what the first step costs and what you get, and after that they share one path. This page maps what you have to the door, and says plainly what none of them do.

Try it: write down, in one line, what you are holding right now (words, an idea, photos, a clip). Find that line in the first table below. The cost column tells you what the first button will ask.

## What you have, and the door it uses

| You have | Door | The first step costs | You get | It will not |
| A finished script or a story | Upload Script | The read: 4 credits per 1,000 characters | An episode list, a cast, and a storyboard per episode | Fetch a link or read a PDF: paste the text |
| One line of idea | AI Create | Writing, priced by the length you pick; then the read | A draft script to edit, then the same path | Write a second draft free: each writing is priced, and every draft is kept |
| Photos of something you sell or own | Ad Video | Reading the photos is free; the ad has one capped price | A 15, 30 or 60-second ad with an end card | Put a readable logo or price into the picture |
| A clip you may use | Remake a video | The read, priced by the clip’s minutes | A storyboard of the clip, with a cast you can change | Reuse the original footage |
| A photo of a person, a room or an object | Step 2 of a drama: Upload & Replace | Optimising is priced like an image; using it as is costs nothing. Uploading is a paid-plan feature | Shots that follow your picture’s face, room or object | Animate a single still on its own |
| A voice recording | Upload Voice on a character | A paid-plan feature | That character speaks in that voice | Turn audio alone into a video |

All four main doors end in the same kind of project, with a cast, locations, props, episodes and one editor. Four ways to start a project is the short version of this table.

## If you have the words: Upload Script

Paste the text into SceneMixer, or upload TXT, DOCX or Markdown. Anything from a novel chapter to a screenplay works, and the read splits it into episodes itself; the project’s Type (Short Drama, TV Series or Film) decides how: short episodes with hooks for a short drama, longer ones for a series. Under about 10,000 characters the whole script is read in one pass. Above that it is read in two, the cast first and the episodes second, up to 100,000 characters, or up to 300,000 raw characters of English, because every 3,333 characters of English count as 1,000. Long scripts are read in two passes.

Two habits repay the effort. Write the lines you need spoken as Name: words, one sentence per line. And use one spelling for each character and each place. Writing dialogue so it is picked up.

## If you have a premise: AI Create

One or two sentences are enough if they carry five things: the genre, the lead and what the lead wants, the conflict, the ending you want, and the tone. SceneMixer’s Inspire Me fills the box with an example you can edit. The length options run from Mini, about 1,000 characters, to Long, 20,000. Mini is the cheap test, and a Short script is the usual first real choice. If the draft is off, sharpen the brief and write again; every draft stays in Writing History, and you can take any of them into a project. Writing the one-line brief.

## If you have photos: Ad Video

Upload one to four photos of the same subject to SceneMixer, with an optional line saying what it is. The first photo is the main subject. A product, a dish or a garment is redrawn as a clean studio shot on white so it can be lit the same way in every segment. A shop, a restaurant, a car, a house or an app screenshot is not redrawn; if people are in a photo of a place or a vehicle they are removed first.

Length is 15, 30 or 60 seconds, which is one, two or four segments. Eight vibes set the storytelling. Quality is 480P, 720P or 1080P, and the menu shows the all-in price of each. The button is a cap, not an estimate: you are charged for the steps that actually happen. Absolute claims such as “number one” are stripped from the script, every segment has a spoken line, and the brand, product name, tagline and call to action go on the end card rather than into the picture. How to make an ad from one photo · Ad vibe, length, format and quality.

## If you have a clip: Remake a video

SceneMixer takes an MP4 or MOV you have the rights to, up to 300 MB and up to 5 minutes. It is read into a scene-by-scene script: who is there, what happens, every line. The original footage is never given to the video model; what comes out is generated from the description, with your cast. The read does not give you a shot list. You pick the look, open the episode and generate the storyboard, priced like any episode.

Expect the result to run longer than the source. Real editors cut every second or two and real actors talk fast; the generated dialogue is paced at reading speed and each shot gets a minimum of a few seconds, so a 26-second clip with nine lines becomes about 45 to 60 seconds. After the read, generate the storyboard · Recasting a remake.

## If you have a photo or a voice for one character

Open the card in SceneMixer’s step 2 and press Upload & Replace. Optimize into a reference sheet reads the face, hair and clothing and redraws them in the project’s style on a clean background, priced like an image. Use this image as is puts your picture on the card untouched and costs nothing. From then on every shot that names the character is made from that picture. A recognisable public figure or a franchise character is refused when you generate. Using your own photo.

A voice recording is a separate upload on the character, and the voice travels with the character if you reuse it in another project. Reusing a character, scene or voice.

## Mixing routes, and what none of them do

- Script plus your own faces. Read the script, then replace the character pictures with your photos in step 2. Videos made after the replacement follow the photos.
- Remake plus a new cast. Read the clip, then rewrite the characters and upload your own faces before any image is generated.
- Ad plus a line. If your line and the photo disagree, the line wins: write our family restaurant and the photo is treated as a place, not a product.
None of the doors takes a link, a PDF, or audio on its own, draws readable text inside the picture, or animates one still in a single click. Image to video AI explains the photo routes.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Soft Romance; Costume Style Smart Casual; Visual Style Filmic Live-Action; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with a description each; the rooftop café and the tram stop as locations; the umbrella as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Route | Best when | Watch out for |
| Upload Script | The words already exist | One name per person and per place; dialogue as Name: line |
| AI Create | Only the idea exists | Start at Mini; each draft is priced |
| Ad Video | The goal is to sell one thing | The first photo is the main subject; your line outranks the photo |
| Remake a video | A reference clip exists and you hold the rights | The result runs longer than the source |
| Photo or voice on a card | A specific face, room, object or voice matters | A paid-plan feature; public figures are refused |

## FAQ

### How do I create videos with AI if I only have an idea?

Use SceneMixer’s AI Create: type a line with the genre, lead, conflict, ending and tone, pick the Mini length to test, and a script is written for you to edit before the read.

### How do I create a video with AI from photos?

Use SceneMixer’s Ad Video: upload one to four photos of one subject, add a line if the photo is ambiguous, pick a length of 15, 30 or 60 seconds, and the ad is made with an end card.

### Can I create a video from my own script?

Yes. Paste it or upload TXT, DOCX or Markdown on SceneMixer’s Upload Script tab. The read splits it into episodes, a cast, locations and props.

### Can AI recreate a video I already have?

SceneMixer’s Remake a video reads a clip you hold the rights to into a storyboard and generates a new video with your cast. The original footage is never given to the video model.

### Can I use my own photo as a character?

Yes, in step 2 of a drama, with Upload & Replace. It is a paid-plan feature; optimising the photo into a reference sheet is priced like an image.

### Can I create a video from a link or a PDF?

No. SceneMixer does not fetch links and refuses PDFs. Paste the text or export to text or Word first.

### What does the first step cost on each route?

On SceneMixer, the read costs 4 credits per 1,000 characters; AI writing is priced by length; reading ad photos is free; a remake is read by the clip's minutes.

### Can I animate a single photo?

Not with one button. In SceneMixer, photos become an ad, or the face, room or object that a drama's shots follow.

Updated November 2, 2026


---

# "Credits Held for Review": What Is Being Checked and What Happens Next

URL: https://scenemixer.com/guides/credits-held-for-review-what-happens-next/
A few failures cannot be refunded on the spot because the clip may already exist. Which cases, how the review ends, why not to resubmit meanwhile.

# “Credits Held for Review”: What Support Is Checking, How Long It Takes, What You Should Do

Quick answer: Most failures refund themselves. A few cannot: the request had reached the model, something went wrong afterwards, and the site cannot tell whether a clip was produced. Refunding blindly would pay for a clip twice, so the credits stay held and a person checks. The review ends one of two ways, shown on the task’s line in your credit history: Refunded (nothing was delivered), or the tag goes and it counts as spent (the clip was delivered and appears on the segment). Do not generate the same segment again while it is held; that would hold a second amount. You do not need to write in.

Open the credits panel: the held amount is listed separately under Frozen, and the task’s own line in Credit History is tagged Under review.

## Why some failures are held instead of refunded

The site refunds on the spot only when it knows the model never did the work: the provider refused, or the request never left. When the request reached the model and the failure came after — a connection dropped mid-generation, the finished clip could not be stored, the provider cannot report the result — there may be a clip on the provider’s side that was already paid for. That is the case that goes to review. The three families of failure messages: When a Generation Fails.

## What the review does

Support looks at the task: whether the provider produced a clip, whether it can be fetched and delivered, what the error was. If the clip exists it is delivered to your segment and the held credits become a normal charge; if not, they are refunded. A task whose provider can still be asked is checked automatically first; only the ones that cannot be resolved that way reach a person.

## How long

Automatic checks run within minutes to hours. Support reviews what is left, and the result appears in your history without any action from you. If a held amount is older than a few days, send feedback naming the project, episode and segment — How to get help.

## What you should and should not do

Do: keep working on other segments; check the segment later — a delivered clip lands there by itself. Do not: regenerate the same segment to “fix” the hold; that is a separate task with its own hold, and if the first clip is later delivered you have two. Deleting the project does not settle a held task either: it stays held until support decides it.

## How the line changes after review

If the review finds nothing was delivered, the task’s line changes from Under review to Refunded. If a clip was delivered while its charge had failed, the tag simply goes and the line stays as spent. No new line appears either way. Reading Your Credit History; the full picture of what refunds and what does not: Refund rules.

| Outcome of the review | The task’s line | On the segment |
| Nothing was delivered | Refunded | Still failed; regenerate when you like |
| The clip was delivered | Spent, no tag | The clip appears as a version |
| Delivered earlier, charge had failed | Spent, no tag | Already there |

## FAQ

### Did I lose the credits?

No. They are held, not spent, until the review ends in a refund or a charge.

### Should I open a ticket?

Not for a hold under a few days old; the review runs by itself. If it is older, write in and name the project, episode and segment.

### Can I regenerate the segment meanwhile?

You can, but it is a second task with a second hold. Better to wait.

### Why does this take a person?

Because paying the provider twice is worse than waiting a day; only the ambiguous cases go this way.

### Does a hold count as a moderation strike?

No. Holds are about delivery, not content.

Updated October 13, 2026


---

# Custom Visual Style: How to Write Your Own Look

URL: https://scenemixer.com/guides/custom-visual-style/
The + Custom Style box in step 1: who can use it, how to write a style description that holds up, what it shapes, and what to do when it goes wrong.

# Custom Visual Style: How to Write Your Own Look

Quick answer: In step 1, open the Visual Style card and press + Custom Style at the bottom of the picker: a text box appears under the card, and the description you type replaces the preset for the whole project. It is a member feature — on a free account the button carries a small PRO tag and opens the membership dialog instead. Your words are used exactly as written, into every character sheet, scene image, prop image and video, and writing them costs nothing. The catch is timing: once you continue past step 1 the style is locked, so decide before you press Next.

Open step 1, click the Visual Style card, and look at the bottom row of the picker: that is where + Custom Style lives.

## Where it is, step by step

- Open the project’s step 1, Set Up Your Project. Visual Style is the first of the three style cards (Scene Style and Costume Style sit next to it).
- Click the card. The picker opens with the grid of preset samples and one line of hint text: the style sets the rendering medium and texture of the whole piece, and the character sheets, scenes and video all inherit it.
- Along the bottom of the picker is the row of special choices. On the Visual Style card that is + Custom Style. (Scene Style and Costume Style also have a None there.)
- Press it. The picker closes and a text box appears under the card, marked Enter custom style description.... Type your description there.
- The card now reads Custom Style instead of a preset name, with your sentence under it. If you change your mind, open the picker again and click a preset — that clears the custom text and goes back to the preset.

Nothing is committed until you press Next: before that you can switch between presets and your own text as often as you like, and edit the text in place.

## What your text replaces

- It is one or the other. The project’s style is either a preset or your text — the box does not add to a preset. If you like a preset but want one thing different, describe the whole look in your own words.
- Used as written. Your sentence is passed on exactly as typed: nothing translates it, rewrites it, or expands it. What you name is what the generators get.
- Length. The box shows no counter; the project keeps up to 200 characters and quietly cuts anything past the 200th. Two or three sentences cover everything a style needs.
- Empty box. If you continue with the box empty (spaces only), the preset you had selected before is saved — by default that is Digital HD. So an empty box is never a broken project, just a preset project.
- No charge. Writing the description costs nothing. Your images and videos cost exactly what they cost under any preset.
- The AI suggestion never fills this box. When you let the AI pick a style, it chooses among the presets; wording your own is yours to do.

## What it shapes

- Every character sheet, scene image and prop image in step 2 — the first thing you will see carrying your look.
- State variants and AI redraws you generate later: anything produced for the project afterwards keeps the same style.
- Every video segment of the project: each video prompt carries it.
- The script read, too: it rides along when your script is read, so the descriptions written for your cast and sets keep to the same look.
- What it never changes: what happens in a shot. Framing, camera, blocking and dialogue come from the storyboard; the style only decides how the frame is drawn.

## How to write a description that holds up

The test is one question: could you point at each phrase in a finished frame? A medium, a surface, the way lines and shadows behave, a palette — all pointable. Praise, moods and orders are not. Two or three sentences is plenty. Six examples that pass the test:

- Film: 35mm film look: visible fine grain, a soft glow around lights, slightly faded warm shadows, muted greens.
- Watercolor: Hand-painted watercolor: visible paper grain, pigment pooling along the edges, thin broken outlines, a pale earth palette with one warm accent.
- Cel animation: Flat cel shading: hard-edged color fields, bold uniform outlines, no gradients, a high-saturation primary palette.
- Stop-motion clay: Clay stop-motion: every surface keeps fingerprints and tool marks, rounded corners, matte toy colors, the shallow focus of a tabletop miniature.
- Picture-book painting: Gouache picture-book painting: matte paint texture, visible brush shapes, a short palette of indigo, ochre and cream, soft paper tooth.
- Ink: Black ink and grey wash: confident brush lines, paper white kept as large empty areas, no color, soft grain in the greys.

Keep it about how, not what: the people, the places and the events live in your script. The description of the look is the one place where you talk about painting, not about the story.

## Three ways descriptions go wrong

- Praise instead of description. “Beautiful, cinematic, 8K, masterpiece quality” names nothing visible; these words are noise to the render at best. Write the cause instead: “soft window light, muted pastel palette, visible paper texture.”
- Camera and framing orders. “Extreme close-ups throughout, the heroine always fills the frame” — every shot already gets its own framing from the storyboard, and a project-wide order fights each of them. A style describes how a frame is painted, not how it is shot.
- Names, and the wrong layer. “In the style of” a studio, franchise or artist: those names can get your generations rejected by the video moderation, so describe the look in plain traits instead. And scene content like “rainy neon street at night” is not a rendering style — it belongs to the scene row or the script.

One trap is specific to this box: your text also goes into the scene and prop images, which are supposed to show no people at all. Person words in the style (“a lone swordsman”, “a pretty girl”) can pull figures into those empty sets. Keep the description to medium, material and light.

## Locked once you continue — and who can use it

Under the three style cards sits the notice, and it means what it says: the style cannot be changed once you continue, because it shapes every character sheet, scene and video; to switch styles, start over with a new script. Custom text is under the same lock — there is no edit entry after step 1, so your last chance to change a word is before you press Next.

On membership: custom style is one of the member features. On a free account the + Custom Style button carries a small PRO tag and picking it opens the membership dialog instead of the box; the requirement is also enforced when the project is saved, not just on the button.

## The same box on the other two rows

Scene Style and Costume Style each have their own + Custom Style, with the same member rule, the same length limit and the same lock at continue. They then work differently once the project runs: they are folded into the descriptions at the script read rather than pasted in front of every image, which is explained in Scene Style and Costume Style Apply at Parse Time, Not Later. Their example placeholders show the flavor: an art direction like “old industrial plant, steel frames and rusted sheet metal, cold green palette”, or a costume system like “Miao silver regalia, pleated skirts and horned crowns, indigo batik”.

## If the result looks wrong

| What you see | Where it comes from | What to do |
| The images look like a standard preset instead of your text | Either the box was left empty when you continued (the previous preset is saved), or a preset was clicked after typing (clicking a preset clears the custom text) | Check the card before continuing — it reads Custom Style with your sentence when the text is in play. After continuing, a new project is the way to change it |
| Videos come back blocked by moderation | A studio, franchise or artist name in the description | Rewrite the look in plain traits; keep brand names out. What gets blocked, and what to change |
| People appear in scene images that should be empty rooms and streets | Person words in the style text | Describe medium, materials and light; leave people to the script and the cast |
| Every shot is framed strangely — all close, or the subject crowding the frame | Framing or camera instructions in the style text | Take them out: framing is decided per shot in the storyboard, not by the style |
| The look is close to the text but oddly generic | Praise words and moods took the room the concrete traits needed | Trade adjectives for visible things: grain, brush marks, palette, light |

## FAQ

### Who can use + Custom Style?

It is a SceneMixer member feature. On a free account the button carries a small PRO tag and picking it opens the membership dialog instead of the text box, and the requirement is enforced when the project is saved, not only on the button.

### Does writing a custom style cost credits?

No. The description itself is free; your images and videos cost exactly what they cost under any preset.

### Can I extend a preset instead of replacing it?

No. The project's style is either a preset or your text — the box replaces the preset completely. If you like a preset but want one thing different, describe the whole look in your own words.

### Can I change the style after step 1?

No. It locks when you continue, because it shapes every character sheet, scene and video. To change it, start a new project with a new script.

### How long can the description be?

Two or three sentences. SceneMixer keeps up to 200 characters and quietly cuts anything past the 200th, so keep it short and concrete.

### What language should I write it in?

Any — the text is used exactly as written and nothing translates it. What matters is that the words name visible things: a medium, materials, lines and shadows, a palette.

### What happens if I leave the box empty?

The preset you had selected before is saved — by default Digital HD. An empty box is not an error, just a preset project.

### Can the AI write the style text for me?

No. When you let the AI pick a style, it chooses among the presets; custom text is yours to write.

### Do Scene Style and Costume Style have the same box?

Yes — each has its own + Custom Style, with the same member rule, the same length limit and the same lock at continue. How those two take effect once the project runs is covered in the scene and costume style guide.

Updated November 3, 2026


---

# Default Tier and Recommended Badge, Explained | SceneMixer

URL: https://scenemixer.com/guides/default-and-recommended-badges-explained/
What the Recommended badge on SceneMixer’s model tiers means, which tier is the default for video, image and text, and how your own choice fits.

# The default tier and the Recommended badge

Quick answer: In Settings → Model Preferences, one tier per category carries a Recommended badge: the tier the site suggests for most finished work in that category. It changes nothing by itself. The default tier carries no badge; it is simply the tier you are on until you pick another, and the one the site falls back to if the tier you picked is withdrawn. Your own choice overrides both.

Which model is best is the page to read once you know what the labels mean.

## Default and recommended today

| Category | Default tier | Recommended badge |
| Video | Wan 3.0 · 480P | Wan 3.0 · 720P |
| Image | Qwen-Image 3.0 | Qwen-Image 3.0 Pro |
| Text | Doubao Seed 2.1 Turbo | Doubao Seed Evolving |

The site adjusts these as tiers change; the Settings list always shows the current Recommended badge.

## The default tier

Three jobs, no badge. It is the tier a new account starts on for that category. It is the tier used when you have never picked one. And it is the tier the site falls back to when the tier you picked no longer exists, with the price on the generate button updated to match.

## The Recommended badge

Advice, shown as a badge in the Settings list. It marks the tier SceneMixer suggests for a typical keeper — a 720P segment to share, a Pro image tier for faces, a base-rate text tier for a hard episode. Nothing is switched for you; you pick it when you want it. The editor’s tier menu shows no badges.

## The First one free badge

A second badge you may see: First one free. It appears on the default text tier while the free storyboard for episode 1 of your first project is still unused, and disappears once it is used. The free storyboard only applies on that tier.

## How your choice fits

SceneMixer’s Settings → Model Preferences holds one choice per category. For video, the tier button in the episode editor changes the same setting. Whatever you pick wins over the default, and picking the Recommended tier there is all it takes to make it your everyday tier.

## FAQ

### Is Recommended the same as the default?

No. The default on SceneMixer is what you get without choosing; Recommended is a suggestion for finished work.

### Is there a Default badge?

No. The default tier on SceneMixer is not labelled; if you have never changed a category, you are on its default.

### Does picking a tier in the editor change my setting?

Yes. The video tier you pick in SceneMixer's editor is saved as your video preference and applies to the segments you generate next.

### Which tier does the fallback use?

The default for that category; for video in SceneMixer’s editor, the tier the episode last used comes first if it is still listed.

### Does the badge affect price?

No. Each SceneMixer tier has its own price; the badge only labels it.

Updated November 3, 2026


---

# Delete a Character, Scene or Prop in an AI Drama Project

URL: https://scenemixer.com/guides/delete-a-character-scene-or-prop/
How to delete a character, location or item in step 2 of a SceneMixer project, what happens to lines, storyboards and videos, and when to regenerate instead.

# Delete a character, scene or prop from your project

Quick answer: Open the character, scene or item in step 2 (Characters & Scenes) and press the red Delete at the bottom left of its card. After you confirm, it gets no more images, is not attached to any new video, and disappears from the list. Storyboards already generated stay as they are, deleting is free, and credits already spent on its images are not refunded. If a character still has lines or appearances, the site shows how many first and suggests regenerating the image instead — deleting a character who is still in the story leaves them in the videos with no reference to follow.

Want a different look rather than no character? Regenerate one sheet.

## When deleting is the right move

- The reading created someone who is only a passer-by — “the waiter”, “a guard” — and you would rather they be background people. See background people in a scene.
- The same place was read twice under two names, and you only want one set.
- A prop that is only mentioned, never seen, got its own card.
- A character was cut from the story and you have already removed their lines.

If you only dislike how someone looks, do not delete them: edit the description and regenerate the image. Their lines, voice and appearances stay connected.

## Step by step

- Open the project and go to step 2. Pick the Characters, Scenes or Items tab.
- Open the card you want to remove.
- Press Delete at the bottom left of the card (it is red).
- Read the confirmation and press Delete again. The card closes and the list updates.

The confirmation reads: It won’t get images or appear in videos anymore, and it’s removed from this list. Storyboards already generated stay as they are. Re-reading the script will create it again. Credits already spent on its images aren’t refunded.

## When the character is still in the story

If the character still has dialogue or appears in the storyboard, you see a different dialog first: “Name” is still in the story, with the counts — how many lines of dialogue, how many episodes they appear in, how many generated storyboard segments they are on screen in.

Deleting then removes the character from the list, along with their character image and voice, but their lines and appearances stay in the story. They will still appear in videos, just with no character image or voice to follow, so their look and voice may not match from clip to clip. You can press Keep character, or Delete anyway.

To take a character out of the story properly, remove their lines and mentions from the storyboard first — editing the storyboard — then delete the card.

## What changes and what does not

| | After deleting |
| Its card in step 2 | Gone from the list |
| New images (sheets, state images) | None; nothing can be generated for it |
| New videos | It is not attached as a reference |
| Storyboards already generated | Unchanged; the text is not rewritten |
| Videos already made | Unchanged |
| Credits spent on its images | Not refunded |
| Free sheets it used | Not returned to the allowance |
| Re-reading the script | Creates it again, by name |

## When you cannot delete yet

- Something here is still generating or syncing — delete it once that finishes. An image is being drawn for it, the episode’s storyboard is being written, or the script is being read. Wait for it to finish; nothing is cancelled for you.
- In a SceneMixer ad project, the product itself cannot be deleted; other cast and places can.

## There is no undo

A deleted character, scene or item cannot be restored from the project. If you might want its image again, download it before you delete. Re-reading the script brings the card back by name, but as a new card: its image has to be generated again.

If the character was on your promo cover, the cover you already have stays; the next time you make one, you will be asked to pick the cover characters again.

## FAQ

### How do I delete a character in SceneMixer?

Step 2, open the character's card, press the red Delete at the bottom left, and confirm.

### Does deleting a character remove their lines?

No. Storyboards already generated are not rewritten. If the character still has lines, SceneMixer warns you and shows how many before you delete.

### Do I get credits back?

No. Credits spent on the deleted item's images are not refunded, and free sheets it used are not returned.

### Does deleting cost anything?

No. Deleting is free.

### Can I undo a delete?

No. Download the image first if you may want it. Re-reading the script creates the card again by name, without its old image.

### Why can't I delete it right now?

Something for it is still generating or syncing, such as its image or the episode's storyboard. Delete it once that finishes.

### Will a deleted character still appear in my videos?

Videos already made are unchanged. In new videos, a character who still has lines or appearances in the storyboard can still appear, but without their image or voice to follow.

### Can I delete a location that the episode uses?

Yes. Shots set there keep their storyboard text, but new videos get no location image for them; generate those segments after giving the shots another location.

Updated November 3, 2026


---

# Delete Your Account: How to Ask, What Is Removed, What Blocks It

URL: https://scenemixer.com/guides/delete-your-account/
How to ask support to delete your SceneMixer account from the website, what is removed and what is not refunded, what to settle first, and signing up again.

# Deleting Your Account

Quick answer: There is no self-service delete button on the website: the deletion is carried out for you by support. Write from the email address on your account to support@scenemixer.com; if you are signed in you can also send the request from inside the app — Ask AI → Send feedback (on a phone, Ask AI is in the menu), or My feedback in the account menu — and ask for the account to be deleted. Deletion is final: you are signed out everywhere, your projects with their sheets, storyboards and videos are removed, share links stop working, and your email is freed for a new account. Remaining credits and any membership time are not refunded. Before you ask: finish or cancel running tasks, cancel an auto-renewing plan on the Membership page, and settle any partner payout and any payment that is still being confirmed.

What you get to keep: download what you want before you delete.

## Before you ask

- Download what you want to keep. Downloading is free, and nothing in the account can be reached once it is closed; what you get to keep covers the details.

- Use up the balance, or accept the loss. Remaining credits and paid membership time are not refunded.

- Finish or cancel running tasks. A task that has not started can be cancelled; a video that is already generating cannot be stopped — wait for it to finish or fail.

- Let payments land. One that is still being confirmed usually clears in a few minutes; a membership checkout opened in the last day has to complete or expire.

- Cancel an auto-renewing plan. The plan on the Membership page carries a Cancel control (its tooltip reads Cancel subscription); it stops the next renewal, and the paid period runs to its end. The deletion cannot go through while a plan is still set to renew, because it would keep charging a closed account. More in subscriptions, upgrades and cancellations.

- Partner payouts. Cash commission that has not been paid out cannot be claimed once the account is gone; settle anything in progress before you ask.

## How to ask for deletion

- Write to support@scenemixer.com from the email address on your account, with a subject like “Delete my account”. Signed-in users can send the same request as in-app feedback — Ask AI → Send feedback (on a phone, Ask AI is in the menu), or My feedback in the account menu; it arrives with the account attached.

- Mention anything that needs handling first — a task you are waiting on, a plan that is already cancelling, a partner payout in progress — and ask if you want a copy of your data before it goes.

- Support typically replies within two business days; the request is checked against the email on the account.

- Verified requests are honoured within 30 days, as the Privacy Policy states. Once the account is closed, every session ends at once and nothing inside it can be reached any more.

## What is removed

Every project — cast sheets, storyboards, segment videos, assembled episodes — and every upload in your asset library. Share links to your videos stop working. Your email address (and phone number, if one is on the account), nickname and other personal details are wiped; the address can be used for a new account. Payment records are kept as the law requires, without your personal details.

## What is not refunded

Credits left on the balance, and the remaining time of a plan you paid for. Use them up, or let the plan run to its end, before you delete.

## What can delay it

Deletion cannot go through while something is still unresolved, and support will tell you which one applies: a task still running (wait for it, or cancel one that has not started — a video already generating cannot be stopped); an auto-renewing plan (it would keep charging a closed account; cancel it on the Membership page first); a membership payment that has not finished (let it complete or expire); a payment still being confirmed (a matter of minutes, as a rule). If the account owns a featured sample project, that is handled with you directly in the same request.

## Signing up again

You can, with the same email address — and the same Google account starts a new account too. It is a new account: nothing from the old one comes back, and the welcome credits are not granted a second time to an address or number that already received them.

| | After deletion |
| Sign-in | Ended everywhere, at once |
| Projects, sheets, videos, uploads | Removed |
| Share links | Stop working |
| Email | Freed for a new account |
| Credits | Not refunded |
| Membership | Cancelled; remaining paid time not refunded |
| Referral code and partner earnings | Code retired; no commission on later payments; unpaid cash cannot be claimed |
| Payment records | Kept as required, without personal details |

## FAQ

### Where is the delete button?

SceneMixer has no self-service button: support carries the deletion out for you. Write from the email address on your account to support@scenemixer.com, or send in-app feedback while signed in (Ask AI → Send feedback, or My feedback in the account menu), and ask.

### How long does it take?

Support typically replies within two business days. SceneMixer honours verified deletion requests within 30 days, as the Privacy Policy states.

### Are my credits refunded?

No. Remaining credits and remaining plan time are forfeited; use them first.

### What happens to my membership?

It ends with the account and the remaining paid time is not refunded. If it renews automatically, cancel it on the Membership page before you ask.

### A task is still running — what now?

The deletion waits for it. Cancel a task that has not started; a video that is already generating cannot be stopped, so wait for it to finish or fail.

### Can I get my projects back later?

No. Deletion is final. Download what you want to keep first.

### Can I sign up again with the same email?

Yes, as a new account; the same Google account starts a new account too. Nothing from the old account comes back, and the welcome credits are not granted twice.

### What happens to my referral link and partner earnings?

The referral code is retired with the account and stops earning; commission on payments made after deletion is not counted. Cash commission that has not been paid out cannot be claimed once the account is gone — settle it before you ask.

### What about my data and payment records?

Personal details are erased or anonymised as the Privacy Policy sets out; payment records are kept as the law requires, without your personal details. Ask in the same email if you would like a copy of your data first.

### Can I use the site on my phone?

Yes — everything runs in the browser, including on a phone; there is nothing to install. Using SceneMixer on a phone.

Updated November 3, 2026


---

# Deleted a Project or File by Mistake? The Seven-Day Window

URL: https://scenemixer.com/guides/deleted-projects-and-files-the-seven-day-window/
Deleting removes the item from your account at once; its files are kept seven days before permanent removal. What that window is for and what to do after a mistake.

# Deleted Projects, Episodes, Sheets and Videos: What Happens Right Away and What Happens After Seven Days

Quick answer: Deleting a project, an episode, a sheet, a clip or a library upload removes it from your account immediately: it is gone from the lists and nothing on the site can open it. The files behind it (images, videos) are not erased at that moment; they are queued and permanently removed about seven days later. That week exists so that a mistaken deletion can still be looked into by support; it is not a recycle bin you can open, and restoring is not guaranteed. If you deleted something by mistake, write to support within the week, with the project name and roughly when; after seven days nothing can be brought back. In-flight work in a deleted project: queued work is refunded, clips already sent are settled by what the provider reports, and other work already sent is held for review.

Not deleted anything, but a video seems to be missing? Check the episode cards first: where are my videos.

## What deletion does right away

The item leaves your account: project cards, episode lists, the library. Links you shared to its clips stop working. Its storage is released from your quota at once. Nothing on the site shows it again. Project Cards: Resume, Rename, Delete.

## What the seven days are

The files are held for about a week before permanent removal. In that week, support can look into a deletion you report as a mistake; whether anything can be reassembled depends on what was deleted and how long ago, so treat it as a chance, not a promise. After the week, the files are removed from the site’s storage.

## What is never deleted by a delete

A sheet that is also used in another project (copied from the library, or a reference image shared between projects) stays where it is still used; the site checks references before removing a file. Your credits, ledger and orders are account records, not project files, and are untouched. Generated results do not count toward storage, so deleting to free quota only matters for your own uploads. The Asset Library.

## Deleting while something is generating

SceneMixer cancels and refunds generations still waiting in the queue, or not yet sent to the model. Clips already sent are settled by what the provider reports: charged if delivered, refunded if it confirms a failure. Other work already sent (sheets, storyboards, parses) is held for support to review. Clips delivered after the deletion are not attached anywhere and are cleaned up with the rest. Stop a Generation.

## Deleting the account

Account deletion follows the same week for files, and additionally clears the account itself; sign-up bonuses are not granted again to the same email or phone if you re-register. Delete Your Account.

| Deleted | Gone from account | Files removed | Credits |
| Project | At once | About 7 days later | Queued work refunded; clips in flight settled by the provider; other sent work held for review |
| Episode | At once | About 7 days later | Same |
| Sheet or clip | At once | About 7 days later, if not used elsewhere | — |
| Library upload | At once | About 7 days later, if not used elsewhere | — |
| Account | At once | About 7 days later | Balance forfeited |

## FAQ

### Is there a recycle bin?

No. Deletion is immediate on your side; the week is a support window, not a bin.

### I deleted the wrong project ten minutes ago.

Write to support now with the project name; the sooner the better.

### Do I get credits back for a deleted project?

For queued work, and for running work that had not reached the model yet. Clips already sent are settled by what the provider reports; other work already sent, and any task already held, is held for review.

### A sheet I deleted is still showing in another project.

It is used there and is kept there; deleting from one project does not remove it from others.

### Does deleting free storage?

Your own uploads, yes, at once. Generated results were never counted.

Updated November 3, 2026


---

# Describe a Face So It Isn't a Generic Face: Bones, Marks, Anchor

URL: https://scenemixer.com/guides/describe-a-face-so-it-is-not-a-generic-face/
Character sheets come from the appearance text. What makes a face specific (bones, marks, one anchor), what pulls it back to a stock face, which words ruin the skin.

# How to Describe a Character’s Face So the Model Draws This Person and Not a Generic Pretty One

Quick answer: A face becomes specific through things that can be drawn: the shape of the jaw and brow, the set of the eyes, the nose bridge, the lips, skin tone, hair line and cut, age as it shows on the face, and one unmistakable anchor — a scar through an eyebrow, a chipped tooth, a mole under the left eye. Words like beautiful, handsome, charismatic describe nothing drawable and pull the model toward its default good-looking face. Put the face first, then build and posture, then clothing; the sheet is a casting photo, not an outfit shot. Two traps: makeup results (‘flawless porcelain base’) give a doll, and any line describing a result on the skin (‘a hint of red at the nostrils’) gets painted literally. Describe the person; leave finish and lighting to the sheet.

Open a character card and read the appearance: if you could not pick this person out of a crowd from the text alone, the model cannot either.

## What the model does with a vague face

Given ‘a beautiful woman in her twenties with long black hair’, the model draws the most probable such woman — the same one it draws for everyone. Every later shot is then matched to that sheet, so the whole series inherits a stock face. Specific text is the only lever: the parse writes detailed appearances for this reason, and you can sharpen them on the card. Same Face in Every Shot.

## What to write, in order

Face: face shape (long, square jaw, round), brows (straight, arched, heavy), eyes (deep-set, hooded, wide-set, upturned), nose (high bridge, broad, hooked), mouth (thin, full lower lip, downturned), skin tone and texture as a person has it (freckled, weathered, olive), hair (line, cut, colour, grey at the temples), the age you see. Anchor: one thing nobody else has. Build and posture: height, frame, how they carry themselves. Clothing: last, one outfit. Chinese readers: traditional bone-structure terms work well and need no translation.

## What quietly ruins the sheet

Adjectives of effect: beautiful, striking, intense. Makeup results: porcelain base, pores hidden, blush placed on the apples; the model paints the result and you get a doll. Say instead what the makeup is for (heavy stage makeup for a performer) and let texture stay. Skin results: redness at the nose, shine on the forehead; these are painted as written. ‘Neutral expression’: it erases whatever made the character readable; write the habitual expression instead (a permanent slight frown). The general rule: Describe the Cause, Not the Result.

## Priority: face over outfit

A long coat or an elaborate costume, described at length, makes the model frame for the clothes and shrink the head. Keep clothing to one sentence on the sheet; costume detail belongs in the storyboard where it matters. One Outfit per Character Sheet.

## Regenerate and compare

After editing the appearance, regenerate the sheet and check the anchor first: is the scar there, is the mole on the correct side? Then the bones. A sheet that gets the anchor and the face shape right will hold across shots. Regenerate One Sheet.

| Instead of | Write |
| A beautiful woman, long black hair | Long face, straight heavy brows, hooded eyes, high nose bridge, thin mouth; black hair cut blunt at the jaw; a small mole under the left eye |
| Handsome, charismatic man | Square jaw, deep-set eyes, a broken nose set slightly left, grey at the temples; habitual half-smile on one side |
| Flawless porcelain skin | Fair skin, fine texture kept; no retouching |
| Neutral expression | A permanent slight frown, eyes alert |
| Elegant and refined | Slim, upright, hands still; speaks little |

## FAQ

### The parse already wrote a detailed appearance; should I still edit?

Read it. If it contains result words (porcelain, hidden pores, redness) or no anchor, fix those; otherwise leave it.

### Can I describe makeup at all?

Yes, by purpose and style (heavy stage makeup, dark lips), not by what it hides.

### Why did my character get a red nose?

A line described redness on the skin; the model painted it. Remove the line.

### Does the visual style change any of this?

No. Style changes the drawing; the description still has to name the person.

### Can I upload a photo instead?

On a paid SceneMixer plan, yes, if you have the rights; it then serves as the sheet and the text matters less.

Updated November 3, 2026


---

# Describe the Cause, Not the Result: Writing for AI Image and Video

URL: https://scenemixer.com/guides/describe-the-cause-not-the-result/
Write what produces the picture, where the lamp is and what the fabric is, not the picture you hope for. Why 'light on her cheekbones' paints a yellow patch.

# Describe the Cause, Not the Result: The One Rule Behind Good Sheets and Good Shots

Quick answer: Image and video models take a description as a list of things to put in the frame. Write a result — “warm light on her cheekbones”, “an air of menace” — and the model paints exactly that as an object: a yellow patch on the cheek, a menacing face pasted on. Write the cause instead — “a warm ceiling lamp above and to the right”, “a scar through the left eyebrow, eyes narrowed” — and the light, the shadow and the mood follow on their own. This one rule fixes most odd sheets and most odd shots.

Take one of your character descriptions and strike every word that judges (beautiful, imposing, cinematic); what is left should still describe a person.

## The rule in one line

A description is not a review of the picture you want; it is an inventory of what is in front of the camera and where the light comes from. Everything else — glow, shadow, mood, drama — is a consequence the model computes. Name the cause, get the consequence; name the consequence, get a literal painting of it.

## Light: place the lamp, never the highlight

“Warm light on her forehead and nose” produces a smear of yellow on the forehead and nose, precisely where you said, like a torch held to the face. “A warm ceiling lamp above and to the right, a window to the left” produces a lit face with a shadow side and a rim of daylight, because that is what such lamps do. Two extra habits: do not name a boundary (“where the two colours meet on her nose” draws a hard line), and avoid saturated colour words for light — a tungsten lamp on skin is warm white, and “yellow” gets you yellow paint.

## Faces: features, not verdicts

“Strikingly beautiful”, “handsome”, “fierce” are verdicts; the model resolves them to an average pretty face and the character loses her identity. Write the bones: eye shape, brow, nose bridge, jaw, a mole, a scar, the way the hair falls. Then one or two things she does with the face: eyes narrowed, mouth set. Identity survives because it was described, not requested. Write a Description That Works.

## Rooms and objects: material and layout, not atmosphere

“A gloomy, oppressive corridor” gives you a dark smudge. “A hospital corridor, green tiles to waist height, fluorescent tubes every three metres, one flickering, doors on the left” gives you a corridor that is gloomy because of what is in it, and it will be the same corridor in every shot. Props the same way: “a worn leather satchel, brass buckle, strap repaired with tape”, not “an important-looking bag”.

## Shots: actions from a starting posture, directions in frame terms

In the storyboard, an action needs a starting state and a direction: “seated, she stands and turns toward frame left” rather than “she reacts dramatically”. Someone outside the frame is a direction (“looks toward frame right”), not a name. “Tension”, “energy” and “a powerful moment” describe nothing the camera can see and are ignored or painted as noise. How to Write One Shot.

| Result (avoid) | Cause (write this) |
| Warm light on her forehead and nose | A warm ceiling lamp above and to the right; window light from the left |
| An air of menace | Scar through the left eyebrow, eyes narrowed, jaw set |
| A gloomy corridor | Green tiles to waist height, fluorescent tubes every three metres, one flickering |
| Volumetric god rays | A layer of cigarette smoke hanging in the room; one small window |
| She reacts dramatically | Seated; she stands and turns toward frame left |

## FAQ

### Is this only for sheets, or for the storyboard too?

Both. SceneMixer draws sheets from descriptions; shots are drawn from the storyboard text. The same rule applies to both.

### Can I say 'cinematic' at all?

It adds nothing the model can place in the frame. Say what makes it cinematic: the lens, the light, the framing.

### What about the visual style preset?

That is where the look lives: live action, anime, 3D. Descriptions should not repeat it.

### Why do my sheets come out with a yellow face?

Almost always a lighting result written in the description. Move the lamp instead.

### Does the rule apply in Chinese and English alike?

Yes; the models read both, and verdict words fail the same way in both.

Updated November 3, 2026


---

# AI Video With Audio: Dialogue and Sound in One Generation | SceneMixer

URL: https://scenemixer.com/guides/dialogue-and-sound-in-one-generation/
The voice, the moving lips and the ambient sound are generated with the picture. What you write to get them, what the voice tag does, and how to change a line later.

# Dialogue and Sound in One Generation

Quick answer: A SceneMixer segment comes out with its sound already in it: each line spoken in that character’s voice, the speaker’s lips moving with the words, and the ambience of the place — rain, traffic, a clock. You do not record, dub or mix anything. You write the line with its voice tag, you write what the place sounds like, and the segment is generated with the picture and the sound as one thing. The download has the audio in it.

Open a finished episode, play a segment with dialogue, and watch the speaker's mouth.

## What comes out

Three layers, all generated with the picture. Voices: every tagged line, spoken in the voice chosen for that character, by the person whose lips move. Ambience: the sound of the place, from the Ambient line of the shot. Silence where you asked for it: a shot marked No dialogue has nobody talking. On top of those, the episode gets its own background score — chosen for the episode and mixed in under the voices, not generated with the clip — and the Music switch by the player turns it off. Background music in an AI short drama.

## What you write to get it

A line: @Mara (low, steady): the words. (voice: 🔊Mara). The voice tag from the @ menu’s Voice tab is what says whose voice speaks; without it the line may be spoken in some other voice, and the editor warns you after saving. Dialogue lines.

Ambience: at the end of a shot, Ambient: rain on the glass, a clock in the hall. Say what is audible, in a few words. Nobody speaking: No dialogue.

Lips: the person speaking moves their mouth; give anyone else in the shot lips closed, no lip movement so they do not mouth the line.

## Where the voice comes from

Each character has one voice, chosen in step 2 — matched to the character, previewed, changeable, or uploaded by you. Every line tagged with that character is spoken in it, in every episode. To change how a character sounds, change the voice on the character, then regenerate the segments you want updated. How character voices work.

## Which language

Lines are spoken in the project’s dialogue language, set when the project was created. Write the lines in that language. A German project speaks German; the interface you use to write it does not matter. Choosing a dialogue language.

## Changing the sound afterwards

SceneMixer has no mixer and no separate track to edit; the sound is part of the clip. You can still change it without regenerating: Redub, under the player of a finished segment, changes the lines or a character’s voice, and Trim cuts a stray word at the start or end. Or change the words or the delivery note and regenerate the segment. If the voice is wrong, check the tag first, then the voice on the character. Change a line after the video is made · Change a character’s voice.

## When the sound is off

Wrong voice on a line: the tag names the wrong character, or is missing. Listener moving their mouth: add lips closed, no lip movement. Line rushed or cut: the shot is too short for the words — about three words a second in English. Shot seconds. Too quiet a room: the shot has no Ambient line; add one. Then regenerate that segment; the old take stays in Version History.

| You want | Write |
| A line in her voice | @Mara (how): the words. (voice: 🔊Mara) |
| Rain and a clock under the scene | Ambient: rain on the glass, a clock in the hall |
| Nobody talking | No dialogue |
| The other person not mouthing the line | @Jules lips closed, no lip movement |
| A voice from off screen | (voice-over) at the start of the delivery note |

## FAQ

### Do I need to add voices afterwards?

No. The lines are spoken in the generated clip, in the character's chosen voice, and the lips move with them.

### Can I change a line after the clip is made?

Yes. Use Redub under the player; it lists the segment's lines for you to edit.

### Can I use my own voice?

Yes, on a paid SceneMixer plan. Upload a voice on the character in step 2; every tagged line for that character is then spoken in it.

### Is there background music?

Yes. SceneMixer gives every newly made episode its own score, chosen for the episode and mixed in under the voices; episodes that had clips before scoring rolled out, in early October 2026, are not scored. The Music switch by the player turns it off — the switch shows only on clips that have a scored version — and downloads follow the switch.

### The wrong character spoke the line.

Check the voice tag after the line: it names the speaker. Fix it and regenerate the segment.

### Does the downloaded file include the sound?

Yes. Voices and ambience are in the clip, and in the composited episode, with the episode’s score unless you turned the Music switch off.

Updated November 3, 2026


---

# "Dialogue Is Not in the Project's Dialogue Language": Fix or Confirm

URL: https://scenemixer.com/guides/dialogue-is-not-in-the-project-language-what-to-do/
Before a clip is generated, the segment's lines are checked against the dialogue language you set. Why lines end up in another language, how to fix them, how to confirm.

# ‘This Segment’s Dialogue Is Not in the Project’s Dialogue Language’: Why, and the Two Ways Out

Quick answer: Every project has a dialogue language, set in Step 1. Before a clip is generated, the segment’s spoken lines are checked against it; if they are in another language the click is refused with this message and nothing is charged. Two usual causes: you edited a line by hand into another language, or the writing model kept a line in the language of the original script. Two ways out: fix the lines by editing them in the segment, or, if the mixed language is intended — a foreign phrase, a song — confirm the dialog and the clip is generated as written.

Open the segment and read the lines after the speaker tags: the language you see there is what the check saw.

## Where the check happens

At the moment you click generate on a segment, on that segment only. It looks at spoken lines (text after a speaker tag), not at the picture description, and it compares their script to the project’s dialogue language. The rest of the episode is not affected. Choose the Dialogue Language.

## Why lines end up in the wrong language

Hand edit: you typed a line in another language; retype it in the project language. Model slip: the storyboard came out with a line in the wrong language despite the setting — often a line the original script had in another language; SceneMixer catches most of these before saving; this check catches the rest. Edit the line. Regenerating the storyboard does not help, because it copies the outline’s lines word for word; and the dialogue language itself cannot be changed after the script is read. What you can change after the read.

## Fixing the lines

Edit the segment text and change only the words after the speaker tag; keep the tag and the delivery note. Save; generate again. Only the segments you generate are checked, so you can fix them one at a time. Dialogue Lines, Voice-Over and Captions.

## Generating anyway

If the other language is intended, confirm in the dialog and the clip is generated with the lines as written; the model speaks them in that language as best it can. The confirmation is per click; the next generation on that segment asks again unless the lines change.

## What is not the cause

The interface language and the language you read the storyboard in do not matter here; the check reads the working text. Captions and picture descriptions are not checked. Remakes have their dialogue language fixed at the read, so this message on a remake means a hand edit.

| Cause | Fix |
| A line edited by hand into another language | Retype the line in the project language |
| Model wrote one line in the wrong language | Edit that line |
| Foreign phrase or song on purpose | Confirm and generate |

## FAQ

### Was I charged?

No. The check runs before anything is sent.

### Does confirming change the project's dialogue language?

No. It only lets this generation proceed.

### Will the clip pronounce the other language correctly?

Best effort; a line in a language the project was not set up for may not sound natural.

### The message appears on every segment of the episode.

The outline’s lines are in another language, usually the original script’s. Edit the lines, or confirm if that language is what you want; regenerating the storyboard keeps the outline’s language.

### Can I turn the check off?

No; confirming per click is the override.

Updated November 3, 2026


---

# Dialogue Language and Video Tier: Which Need Care | SceneMixer

URL: https://scenemixer.com/guides/dialogue-language-and-video-tier/
How spoken dialogue relates to the video tier on SceneMixer: why no tier is marked better for a language, where German and Ukrainian occasionally slip, and what helps.

# Dialogue language and video tier

Quick answer: Every video tier on this site speaks the dialogue you wrote, in the language you chose in step 1, with each character’s voice. The site flags two languages: German and Ukrainian occasionally mispronounce a word on the default tier — a verb ending, a case ending — and the step 1 dropdown does not warn about it. No tier is designated as better for those two; the fixes are shorter sentences, a regeneration, and checking names.

Work in your own language covers the four language settings a project has.

## Where dialogue language is set

In step 1, before the script is read. It decides the language the characters speak in every segment; the storyboard copies lines as written, and the setting is fixed once the script is read; for another language, start a new project with the same script. Changing the video tier does not change the language; it only changes who renders the speech.

## How the tiers compare

| Language group | Default tier (Wan 3.0 · 480P) | Other tiers |
| German, Ukrainian | Flagged: an occasional mispronounced word (verb or case endings) | No tier is marked better; the same care applies |
| The other dialogue languages | Not flagged | No tier is marked better |

The picture is the same across the Wan, Seedance and H3 families: none of them is flagged as the fix for German or Ukrainian, so switching tiers for pronunciation alone is not a good use of credits.

## What actually helps

Shorter sentences. A short line leaves less room for a slip than a long one. Regenerate the segment. A slip is not fixed; the same line can come out right the next time, and every version is kept. Check names. A foreign name in the middle of a line is where most slips happen; spell it the way it is pronounced in the dialogue language, or give the character a name native to that language in step 1.

## A warning you may see before generating

If a segment’s lines are in a different language from the project’s dialogue setting — because you edited them by hand — SceneMixer asks you to confirm before rendering. Confirm to render as written, or fix the lines first.

## FAQ

### Does a more expensive tier pronounce German better?

No tier on SceneMixer is marked as better for German or Ukrainian. Use shorter lines and regenerate.

### Can characters speak different languages in one episode?

A SceneMixer project has one dialogue language; lines you write in another language are spoken as written after a confirmation.

### Why did a Japanese name come out with the wrong reading?

Names are the most common slip. Write the name the way it is pronounced, or pick a name with one reading.

### Does regenerating cost the full price?

Yes; every render on SceneMixer is a new charge. If one fails, its credits come back or are held until support checks it.

### Is the voice the same across tiers?

On SceneMixer, every tier uses the character’s voice setting, up to the tier’s limit on voice clips: at most 3 on the H3 tiers and 5 on the others, and no more than about 14 seconds of voice samples in total, which in practice is often three voices. A character beyond that limit speaks from the written voice description.

Updated November 3, 2026


---

# AI Video Dialogue: Lines, Voice-Over and Captions | SceneMixer

URL: https://scenemixer.com/guides/dialogue-lines-voice-over-and-captions/
How a line is written so it is said by the right voice, in the right way, by the person whose lips move. Voice-over, on-screen captions, and the tag every line needs.

# Dialogue Lines: Delivery Notes, Voice-Over, Captions and the Voice Tag

Quick answer: A line is written inside the shot where it is spoken: @Mara (low, steady): Read it before you say no. Then the voice chip from the @ menu — (voice: 🔊Mara) — which says whose voice speaks it. The delivery note in brackets is how it is said. A line spoken from outside the picture starts its note with (voice-over) or (off-screen). Text that should appear on screen is a caption, not a line: On-screen caption reads “Three years later”. The dialogue is generated with the picture, in the character’s voice, and the person speaking is the one whose lips move.

Open a finished episode and play a segment with dialogue, then read its storyboard. Every line you hear is written this way.

## The shape of a line

@Mara (low, steady, a beat before ‘no’): Read it before you say no. (voice: 🔊Mara)

Four parts. The speaker’s chip. The delivery note in brackets. The words after the colon. The voice chip. Write the line inside the shot where it is spoken, at the moment it is spoken; what happens after it — she closes her lips and looks away — follows in the same shot. If you write it another way — Mara says: “…”, brackets without a colon — saving rewrites it into this shape and says which shots it tidied: writing dialogue in the storyboard, and what save tidies.

## The delivery note

How it is said, in a few words, physical rather than emotional: low, steady; through her teeth; fast, out of breath; calling from the stairs; stress on ‘never’. One note per line. A note that describes a whole mood — heartbroken — gives less to work with than one that describes the voice.

## The voice chip

Every line needs one. Type @ after the words, open the Voice tab and pick the speaker. The chip reads (voice: 🔊Mara) and says the line is spoken in Mara’s voice — the one chosen for her in step 2. A line without a chip may be spoken in some other voice; the editor warns you after saving. How a character gets a voice.

The Cast chip and the Voice chip are different things. The Cast chip puts her in the picture; the Voice chip gives the line her voice. A line usually needs both: her in the shot, and her voice on the words.

## Voice-over and off-screen lines

When the speaker is not in the picture — a narrator, a phone call, a shout from another room — start the delivery note with (voice-over) or (off-screen), keep the voice chip, and do not put the speaker’s Cast chip in the picture text. The line is heard over the shot; nobody on screen mouths it. Who is in frame.

## The listener

Whoever is in the shot and not speaking gets lips closed, no lip movement. Without it, a listener can end up mouthing the other person’s words.

## On-screen captions

Text the viewer should read on the picture — Three years later, a place name, a date — is not dialogue. Write it as a caption in the shot: On-screen caption reads “Three years later”. Give that shot a couple of seconds so the text can be read. Keep captions short.

## Length and language

A line takes as long to say as it takes to say: about three words a second in English, so a twelve-word line needs a four-second shot. Shot seconds. Lines are spoken in the project’s dialogue language, chosen when the project was set up; write them in that language. Dialogue language.

| You want | Write |
| A line said on camera | @Mara (low, steady): the words. (voice: 🔊Mara) |
| A line from off camera | @Jules (off-screen, calling): the words. (voice: 🔊Jules) — no Cast chip for him in the picture |
| A narrator | @Narrator (voice-over): the words. (voice: 🔊Narrator) |
| Text on the picture | On-screen caption reads “Three years later”. |
| The other person in the shot | @Jules lips closed, no lip movement |
| Nobody speaks | No dialogue |

## FAQ

### Is the dialogue really spoken in the video?

Yes. SceneMixer generates each line with the picture, in the voice chosen for that character, and the speaker's lips move with the words.

### What happens if I forget the voice chip?

SceneMixer's editor warns you after saving that the line's audio may drift. Add the chip from the Voice tab.

### How do I write a phone call?

The person on the phone: their line with (off-screen) in the note and their voice chip, no Cast chip in the picture. The person we see: normal line.

### Can I put subtitles or a title card on the screen?

Write it as an on-screen caption in the shot. Keep it short and give the shot a couple of seconds.

### Can I write the line in my own language?

Write it in the project's dialogue language. That is what the voice will speak.

Updated November 3, 2026


---

# Direction Words, Not Names: Writing Who Is Off-Screen in a Shot

URL: https://scenemixer.com/guides/direction-words-instead-of-naming-who-is-off-screen/
Naming a character who is not in the shot invites the model to draw them in. How to write gaze and movement as directions, and the two ways to frame two people.

# Write Directions, Not Names: How to Refer to Someone Who Is Off-Screen in a Single-Person Shot

Quick answer: In a shot that shows one person, do not name the person they are talking to or looking at. ‘Mei glares at Jun’ makes Jun a candidate for the frame, and the usual result is a second person appearing or Mei’s face drifting toward Jun’s. Write the relation as a direction: Mei glares toward frame left, eyes on the door behind the camera, turns to the voice off-screen right. The site already rewrites off-screen names into directions before generating, but text you write as directions from the start is clearer and survives editing. When two people must be seen together, there are only two honest framings: a side-on two-shot, or over one shoulder at the other; ‘A faces B, B in the background’ is a camera position that does not exist.

Open a segment, find a shot listing one person in frame, and read its text: any other character’s name in it is a risk; replace it with a direction and regenerate.

## Why naming an off-screen person is dangerous

The models draw what is named. A name in the text is matched to that character’s sheet and the sheet is a strong pull; in a single-person shot the pull lands on the only person available, so features blend, or a second figure is invented. Directions carry the same staging information without the pull. Who the shot shows is decided by the in-frame list, and the text should agree with it: Who Is in Frame.

## How to write it

Gaze: eyes toward frame right, looks straight into the lens, looks down and to the left. Address: speaks toward the door off-screen. Reaction: flinches from something at frame left. Movement: walks out of frame right. Left and right are the picture’s left and right, as the audience sees them. Keep the other person’s name for the shots they are actually in. How to Write One Shot.

## What the site does for you

Before a clip is generated, SceneMixer replaces names of characters who are not in the shot with direction words, and does not attach their sheets. The same for a character who exists only as a body or a corpse in that shot. It is a safety net: a shot written with names and later edited may slip past it, and directions you wrote yourself read better.

## Two people who must share the frame

Side-on two-shot: both in profile or three-quarter, facing each other across the frame. Over-the-shoulder: one person’s shoulder in the near corner, the other’s face beyond it. Those are the two ways a camera can show two people facing each other. ‘A faces B, with B behind her in soft focus’ asks for A to face away from B and toward B at once; the model picks one and the staging breaks. Depth words (in the background, behind) are camera positions; use them only when the geometry works. Spatial Consistency.

## Keeping sides consistent

Once A looks frame-left at B in one shot, B looks frame-right at A in the reverse. Keep that across the segment; swapping sides between shots reads as the two having changed places. The storyboard’s axis rule is written for this.

| Instead of | Write |
| Mei glares at Jun (Jun not in frame) | Mei glares toward frame left |
| He answers Father on the phone | He answers, eyes down, phone to his ear |
| She watches Jun leave | She watches something exit frame right |
| A faces B, B in the background | Over A's shoulder: B's face beyond it |
| The corpse of Wei at her feet | A body at her feet, face down |

## FAQ

### If the site fixes names anyway, why should I care?

The fix is a net, not a guarantee; and a direction tells the model where to look, which a removed name does not.

### Can I name someone off-screen who is speaking?

Mark the line as off-screen; the speaker's name stays on the line, not in the picture text.

### Which left is 'frame left'?

The audience's left, looking at the picture.

### Why did a second face appear in a close-up?

Almost always a named off-screen character. Replace the name with a direction.

### How do I show two people talking?

Side-on two-shot or over-the-shoulder; nothing else.

Updated November 3, 2026


---

# Do AI Movies Make Money? The Numbers You Can Actually Check

URL: https://scenemixer.com/guides/do-ai-movies-make-money/
Do AI movies make money? No earnings promises: YouTube's published Partner Program thresholds, its monetization wording, what 4,000 watch hours means in views, and costs.

# Do AI movies make money?

Quick answer: Whether an AI film makes money is decided by an audience and a platform's rules, not by the tool, and we have no earnings data for AI films to give you. What can be checked: the thresholds YouTube publishes for its Partner Program, the wording of its monetization policy on mass-produced and templated content, the arithmetic of what those thresholds mean in views, and what a finished minute costs to make here. All YouTube pages checked November 2, 2026; they change, and they are the authority.

Sources: YouTube Partner Program overview & eligibility · YouTube channel monetization policies.

## Where money can come from

- Platform ad share. On YouTube that is the Partner Program (below). Other platforms have their own programs and terms; check their own pages, which we have not summarised here.
- Clients. Short ad spots, trailers, training videos and explainers made for someone else. What to charge for an AI video project · AI ad video generator.
- Pitching. A trailer or pilot to show a series before it exists. A pilot or pitch trailer.
- Festivals and sales. Whoever you sell to or submit to sets its own rules for AI-assisted work. A short film for a festival.

## YouTube's published thresholds

YouTube's Partner Program overview says a channel can become eligible with either of two routes: 1,000 subscribers with 4,000 qualified watch hours in the last 12 months, or 1,000 subscribers with 10 million qualified Shorts views in the last 90 days. It also says you need to follow the channel monetization policies, have no active Community Guidelines strikes, have 2-Step Verification on, and have an AdSense for YouTube account linked. The page says a decision comes after review, “typically in about 1 month”. Its first lines add that updates to the Partner Program start on February 1, 2027, with new terms to accept in YouTube Studio by January 31, 2027. Source: YouTube Help, YouTube Partner Program overview & eligibility (checked November 2, 2026).

## What 4,000 watch hours means in views

4,000 hours is 240,000 minutes. How many views that takes depends on how much of each video people watch, on average:

| Average minutes watched per view | Views for 4,000 hours |
| 1 | 240,000 |
| 2 | 120,000 |
| 3 | 80,000 |
| 5 | 48,000 |
| 10 | 24,000 |
That is arithmetic, not a forecast. A three-minute film that people watch to the end needs about 80,000 views for the watch-hours route, and no tool can promise them. The watch hours also have to be public and “qualified” in YouTube's sense, and the page notes that watch hours from Shorts views in the Shorts feed don't count towards the 4,000.

## What the monetization policy says about the kind of video that earns

YouTube's monetization policies say content should be original and authentic. Under generic or repetitive content they list as allowed “a series following a set of characters across episodes” where each video has a distinct storyline, focus or concept, and “using AI to visualize a unique character and narrative you invented”. Among what's not allowed: “Videos where characters are put in the same situation over and over again with the same outcome”, and AI-generated content made with “generic or unoriginal templates giving the impression of mass production”. Source: YouTube Help, YouTube channel monetization policies.

Questions to put to your own channel, not a ruling on it: is each episode a different story? Did you write the characters and the plot? Would two episodes be hard to tell apart? In SceneMixer the script, the cast and every storyboard change come from you or are edited by you, and each episode is its own script. What YouTube says about AI videos.

## What a finished minute costs to make

Video is 10 credits a second on the default tier (about $0.06 a second at the lowest credit price), and joining an episode is 1 credit a second. The worked example in the cost guide is ten 90-second episodes, fifteen minutes of runtime, from a 30,000-character novel: about 10,026 credits, about $62 at the lowest credit price. Retakes are extra, so budget for them. What a micro-drama costs · Spend fewer credits · AI video cost calculator.

## What we will not tell you

We will not quote typical earnings, say how fast a channel grows, or call a niche ‘worth it’. We do not have those numbers for AI films, and a made-up one would be a promise. The market page has what we do have about the short-drama business. AI micro-drama market data.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Streetwear, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Make each episode its own story, with the same few people; the policy's own example is a series of distinct storylines. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Number | Value | Source |
| Partner Program, watch-hours route | 1,000 subscribers and 4,000 qualified watch hours in 12 months | YouTube Partner Program overview |
| Partner Program, Shorts route | 1,000 subscribers and 10 million qualified Shorts views in 90 days | YouTube Partner Program overview |
| Review time | “typically in about 1 month” | YouTube Partner Program overview |
| 4,000 hours in minutes | 240,000 | Arithmetic |
| Cost of 15 minutes of runtime here | About 10,026 credits, about $62 | The cost guide's worked example |

## FAQ

### Do AI movies make money?

That depends on the audience and the platform's rules, not on the tool. We have no earnings data for AI films, so this page gives the published thresholds and the cost side.

### Can I monetize AI-generated videos on YouTube?

YouTube's disclosure page says disclosing AI content alone won't limit eligibility to earn money, and its monetization policy asks for original, authentic content. Our longer answer.

### How many subscribers do I need?

YouTube's Partner Program overview says 1,000 subscribers, together with either 4,000 qualified watch hours in 12 months or 10 million qualified Shorts views in 90 days.

### How long until I get paid?

Applying starts a review that YouTube says typically takes about a month; the rest depends on your audience and the program's terms.

### Is an AI drama series worth it?

That depends on your goal and your budget. Start with one episode: its storyboard is free on your first drama, and a 5-second preview is free, before any video is charged.

### What does it cost to make a video?

Video is 10 credits a second on the default tier plus 1 credit a second to join, before retakes.

### Do I need a paid plan?

Not to make a video: new SceneMixer accounts start with 50 credits. Free accounts' videos carry a small AI badge, and removing it is a membership benefit, from $7.99 a month.

### Can I sell the film itself?

The site's Terms allow commercial use of results to the extent permitted by law; the buyer sets its own rules. Ownership, with sources.

Updated November 3, 2026


---

# Do Credits Expire? Balance, Held Credits, Transfers and Monthly Grants

URL: https://scenemixer.com/guides/do-credits-expire-and-other-balance-questions/
Purchased and membership credits do not expire or reset monthly. What 'held' means, why credits cannot be transferred, and what welcome and bonus credits are.

# Do Credits Expire? And Other Questions About Your Balance

Quick answer: No. Credits you buy and credits your membership grants stay in your balance until you spend them; there is no monthly reset. That holds for every kind: purchased, membership, welcome, bonus and referral credits alike. What can have a deadline is an offer, such as the first-purchase bonus, which is added only to purchases made in time; once added, those credits do not expire either. Your balance is one number for all of them. Part of it can be held while a task runs; held credits come back or turn into a charge when the task ends. Credits cannot be transferred to another account or cashed out.

Open the credits panel in the top bar: available, held and the history are all there.

## Where the balance lives

The number in the top bar is your available balance. Open it and you see three things: available, held (credits reserved by tasks still running), and your latest Credit History entries, with View all for the full list. New accounts start with 50 credits; purchases and membership grants add to the same number. Reading the rows: Reading Your Credit History.

## Expiry

Purchased credits on SceneMixer do not expire. Membership credits, granted each month (on a yearly plan, each month of the year), also stay; an unused monthly grant carries into the next month and is not cleared. What has a date is an offer: the first-purchase bonus stages and a promotion code each work only within their dates, and the page that shows them says until when. The credits they add stay like any others. What a Paid Plan Changes.

## Held credits

When a task starts, its price moves from available to held. It is not spent yet: on success it becomes a charge, on failure it comes back. One pool holds all your running tasks, so a large held number simply means several tasks are running. A held amount that stays for a long time belongs to a task waiting for review; the task card says so. Refund rules.

## Transfers, sharing, cashing out

SceneMixer credits belong to the account and cannot be moved to another account, pooled between accounts, or converted back to money except under the refund exceptions. The referral programme rewards you with credits for people you bring in; that is the one way credits arrive from someone else’s activity. The Referral Program.

## What happens on cancellation or deletion

Cancelling a card or PayPal membership stops the next renewal; on a quarterly or yearly plan the monthly grants you already paid for keep arriving until the term ends. Credits already granted stay. Deleting your account (by writing to support) gives up the remaining balance; it is not refunded. Delete Your Account.

| Credits from | Expire? | Notes |
| A purchased pack | No | Stay until spent |
| Membership monthly grant | No | Carry over; not cleared monthly |
| Welcome credits | No | 50 on a new account |
| First-purchase / promotional bonus | No | The offer has a deadline; the credits it adds do not |
| Referral rewards | No | Same balance |

## FAQ

### Do my membership credits reset at the end of the month?

No. Unused SceneMixer credits carry over.

### Why is part of my balance 'held'?

Tasks are running. Held credits return or become charges when they finish.

### Can I give credits to a friend?

No. SceneMixer credits stay on the account that holds them.

### Can I convert credits back to money?

Only under the SceneMixer refund exceptions (duplicate charge, unusable purchase, minor).

### If I cancel my plan, do I lose the credits it gave me?

No. SceneMixer stops the next renewal; granted credits stay, and a yearly or quarterly plan keeps granting monthly until its term ends.

Updated November 3, 2026


---

# Dolly Zoom in AI Video: The Hitchcock Zoom for a Moment of Shock

URL: https://scenemixer.com/guides/dolly-zoom-ai/
The dolly zoom (Hitchcock zoom) in AI video: what it looks like, the story moment it is for, the minimum shot length, and how SceneMixer applies it.

# Dolly zoom in AI video

Quick answer: The dolly zoom keeps a person the same size in frame while the world behind them stretches away: the camera tracks backward while the lens zooms in at the same rate. It is the camera move for the moment a character freezes and realises the truth. SceneMixer is one of the few AI video tools with it built in: when your storyboard reaches a moment of shock or realisation it can apply the move, and the shot needs at least 3.5 seconds to play it.

A shot to try: Close-up on @Mara as she reads the name on the letter; the camera tracks back while the lens pushes in; the room behind her seems to fall away. Ambient: a single held note.

## What a dolly zoom does

Two things happen at once, in opposite directions: the camera moves away and the lens zooms in. The person does not change size, but the background stretches or compresses behind them, which reads as a sudden loss of footing. Hitchcock made it famous; films since use it for the second a character understands something terrible.

## When to use it

- The shock: a name on a page, a face in a crowd, a sentence that changes everything.
- The realisation: the character connects two facts.
- Freezing: the moment someone stops moving because they have understood.
Use it once or twice in a film. Used every episode it stops meaning anything.

## How it works here

SceneMixer has eight special camera moves that the storyboard applies at matching story moments; the dolly zoom is for shock and realisation. Give it a shot of at least 3.5 seconds, a single person in frame, and a line or a held beat to land on. Ordinary camera moves (push-in, pan, track) are written into the shot in plain words. All eight special moves · How to write camera movement.

## Mistakes

- Two people in frame. The effect needs one clear subject.
- A shot that is too short. Under 3.5 seconds there is no room for the move.
- Using it as decoration. It should answer a story beat.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Neon Nights; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the one person whose shock you are filming and the one room she is in. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Move | Camera | Lens | Effect |
| Dolly zoom | Tracks backward | Zooms in at the same rate | The person stays the same size; the background stretches away |

## FAQ

### What is a dolly zoom?

A shot where the camera moves backward while the lens zooms in at the same rate, so the subject keeps their size and the background stretches.

### Is it the same as the Hitchcock zoom?

Yes. Both names mean the same move.

### Does SceneMixer have it built in?

Yes. It is one of eight special camera moves the storyboard applies at matching story moments.

### How long must the shot be?

At least 3.5 seconds.

### When should I use it?

For a moment of shock, realisation or freezing, once or twice in a film.

### Can I ask for ordinary camera moves?

Yes. SceneMixer also takes ordinary moves: write push-in, pan or track into the shot in plain words.

### Is it free to try?

Editing the storyboard is free, and SceneMixer gives you a 5-second preview.

Updated November 3, 2026


---

# Doubao Seed 2.1 Pro: The Text Tier for Hard Episodes | SceneMixer

URL: https://scenemixer.com/guides/doubao-seed-2-1-pro-tier/
Doubao Seed 2.1 Pro on SceneMixer: the base-rate text tier, what it does better than the default on storyboards and long scripts, and when to use it.

# Doubao Seed 2.1 Pro

Quick answer: : Doubao Seed 2.1 Pro is the family’s full-size tier at 100% of the base text rate, twice the default. Same three jobs — reading, writing, storyboards — with more headroom on long, crowded episodes: it holds the storyboard’s strict format more reliably and keeps track of many characters and locations in one pass. Switch to it for the episodes that trip the default, not for the whole series.

The default tier is where most text work belongs; this page is the step up.

## Where you pick it

Settings → Model Preferences → Text Model. The choice applies to the three text jobs that cost credits — reading a script, writing a script with AI, and generating an episode’s storyboard — from the next task on. Translations of what you see on screen and the small helper calls do not follow it. Nothing already generated is redone by switching.

## What improves over the default

Storyboards for long or crowded episodes come back in the required format more often on the first draft, with fewer characters dropped and the sequence of shots holding together across a 30-second take. Long scripts keep their episodes. It is not a different style of writing; it is the same family with more room.

## What it costs

On this tier, reading a script costs 4 credits per thousand characters (a thousand equivalent characters for alphabet-based scripts, see the pricing page), a storyboard costs 1 credit per hundred characters of the episode’s share of the script, and AI writing costs 4 credits per thousand characters written. That is 100% of the base rate. Most failed tasks are refunded automatically; a few kinds of failure hold the credits until support checks the task.

SceneMixer's default tier is at 50%. Because the preference applies per task, you can pay this rate for one episode’s storyboard and return to the default for the rest.

## When to use it

An episode whose storyboard failed twice on the default; a script with dozens of scenes or a large cast; a project where every episode is long. For an ordinary series the default tier and this one produce comparable storyboards and the default is half the price.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| A long script or a dense episode fails | The reply ran past what one call can return; the site already ran it once more | The credits stay held until support checks the task. Trim the script or split it before running again |
| Storyboard generation failed | Neither of two drafts came back as a usable storyboard | Nothing was saved or charged; the credits stay held until support returns them. Regenerate, or switch tier |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Try again in a few minutes |
| The storyboard came back in the wrong language | The tier ignored the language setting | The site redoes it once automatically; if it still arrives wrong, regenerate or switch tier |

## When to switch again

To Evolving (same rate, marked Recommended in Settings) if you want the site’s suggested tier. To Qwen3.8 Max for the hardest scripts, at a higher rate. Back to the default for everything routine.

| | This tier |
| Role | Base-rate tier for hard episodes |
| Rate | 100% of base |
| Long scripts | Read in two passes automatically |
| Format checks | Retried once before a failure is reported |

## FAQ

### Is Pro twice as good as the default?

It is twice the price. It is noticeably more reliable on long, crowded episodes and about the same on ordinary ones.

### Do I have to use Pro for the whole series?

No. The preference applies per task; switch for one episode and switch back.

### Does Pro re-read my script?

No. SceneMixer reads a script once per project; switching affects later tasks such as a storyboard.

### Is Evolving better than Pro?

Same rate; Evolving carries the Recommended badge. Try either for a hard episode.

### What if Pro also fails twice?

Try Evolving or Qwen3.8 Max for that episode. SceneMixer does not charge a failed storyboard; its credits stay held until support returns them.

Updated November 3, 2026


---

# Doubao Seed 2.1 Turbo: The Default Text Tier | SceneMixer

URL: https://scenemixer.com/guides/doubao-seed-2-1-turbo-tier/
What the default text tier does on SceneMixer: reading scripts, writing with AI and generating storyboards at half the base rate, where it is enough, and when to step up.

# Doubao Seed 2.1 Turbo: the default text tier

Quick answer: : Doubao Seed 2.1 Turbo is the tier every account uses for text work until you change it: it reads the script, writes with AI when you ask, and generates each episode’s storyboard, at 50% of the base text rate. It handles most scripts and most episodes; when a storyboard fails its format check twice or a long script strains it, the Pro and Evolving tiers are the usual next step.

Which model is best puts the text tiers next to the image and video ones.

## Where you pick it

Settings → Model Preferences → Text Model. The choice applies to the three text jobs that cost credits — reading a script, writing a script with AI, and generating an episode’s storyboard — from the next task on. Translations of what you see on screen and the small helper calls do not follow it. Nothing already generated is redone by switching.

Left alone, this is the tier you are on.

## What it does on this site

Three jobs. Reading: it turns a script into the cast, locations, props and episode outlines you approve in step 1; very long scripts are read in two passes automatically. Writing: when you give it a premise instead of a script, it writes the script in your interface language. Storyboards: it breaks an episode into segments and shots in the strict format the video tiers read.

## What it costs

On SceneMixer's default tier, reading a script costs 2 credits per thousand characters (a thousand equivalent characters for alphabet-based scripts, see the pricing page), a storyboard costs 1 credit per hundred characters of the episode’s share of the script, and AI writing costs 2 credits per thousand characters written. That is 50% of the base rate. Most failed tasks are refunded automatically; a few kinds of failure hold the credits until support checks the task.

Pro and Evolving are at 100% of the base rate, twice this tier.

## Where it is enough

Scripts of ordinary length in Chinese or English, episodes with a handful of characters, storyboards where the site’s own checks pass on the first or second draft. That is most projects. The storyboard is checked for format and language before it is saved; if a draft fails, SceneMixer asks for one more automatically and only then reports a failure.

## Where to step up

When an episode’s storyboard fails twice in a row, when a very long or unusually structured script comes back with episodes missing. Switch the preference and regenerate just that episode; nothing else is redone.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| A long script or a dense episode fails | The reply ran past what one call can return; the site already ran it once more | The credits stay held until support checks the task. Trim the script or split it before running again |
| Storyboard generation failed | Neither of two drafts came back as a usable storyboard | Nothing was saved or charged; the credits stay held until support returns them. Regenerate, or switch tier |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Try again in a few minutes |
| The storyboard came back in the wrong language | The tier ignored the language setting | The site redoes it once automatically; if it still arrives wrong, regenerate or switch tier |

| | This tier |
| Role | Default for reading, writing and storyboards |
| Rate | 50% of base |
| Long scripts | Read in two passes automatically |
| Format checks | Retried once by the site before a failure is reported |

## FAQ

### Does the text tier change my videos?

Only through the storyboard it writes. The video tiers render whatever storyboard you approve.

### Is the default tier good enough for a whole series?

For most series, yes. Step up per episode when a storyboard keeps failing.

### Why is it half the base rate?

Tier rates follow what each tier costs to run.

### Does switching tiers re-read my script?

No. A script is read once per project; switching affects later tasks. To use the new tier for a storyboard, switch before you regenerate it.

### Which tier writes the free first-episode storyboard?

Only this one. The free first-episode storyboard of your first SceneMixer project applies when it is generated on the default tier; on any other tier that episode is charged at the tier’s rate.

Updated November 3, 2026


---

# Doubao Seed Evolving: The Recommended Text Tier | SceneMixer

URL: https://scenemixer.com/guides/doubao-seed-evolving-tier/
Doubao Seed Evolving on SceneMixer: why it is the Recommended text tier, its base-rate prices, what it does for storyboards, and how it compares with Pro and the default.

# Doubao Seed Evolving: the recommended text tier

Quick answer: : Doubao Seed Evolving is the text tier marked Recommended in Settings: base rate (100%), and the tier the site suggests when the default is not enough for an episode. It reads, writes and storyboards like the rest of the family, with the same headroom on long and crowded episodes as Pro. The badge is advice; the default tier stays the default until you change it.

Pro is the same-rate sibling; the two are interchangeable for most hard episodes.

## Where you pick it

Settings → Model Preferences → Text Model. The choice applies to the three text jobs that cost credits — reading a script, writing a script with AI, and generating an episode’s storyboard — from the next task on. Translations of what you see on screen and the small helper calls do not follow it. Nothing already generated is redone by switching.

It carries the Recommended badge in the Settings list.

## What it is for

The episodes the default tier stumbles on: long ones, crowded ones, ones where the storyboard failed its format check twice. It keeps the strict storyboard format on the first draft more often, holds a large cast in one pass and keeps long scripts’ episodes intact. For ordinary episodes it produces storyboards comparable to the default at twice the price.

## What it costs

On this tier, reading a script costs 4 credits per thousand characters (a thousand equivalent characters for alphabet-based scripts, see the pricing page), a storyboard costs 1 credit per hundred characters of the episode’s share of the script, and AI writing costs 4 credits per thousand characters written. That is 100% of the base rate. Most failed tasks are refunded automatically; a few kinds of failure hold the credits until support checks the task.

SceneMixer's default tier is at 50%. Switch per task: this tier for the hard episode, the default for the rest.

## Evolving or Pro

Same rate, same jobs. Evolving carries the site’s recommendation; Pro is the family’s full-size fixed tier. If one fails twice on an episode, try the other before moving to a higher rate.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| A long script or a dense episode fails | The reply ran past what one call can return; the site already ran it once more | The credits stay held until support checks the task. Trim the script or split it before running again |
| Storyboard generation failed | Neither of two drafts came back as a usable storyboard | Nothing was saved or charged; the credits stay held until support returns them. Regenerate, or switch tier |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Try again in a few minutes |
| The storyboard came back in the wrong language | The tier ignored the language setting | The site redoes it once automatically; if it still arrives wrong, regenerate or switch tier |

## When to switch again

Up to Qwen3.8 Max for the hardest scripts. Back to the default for routine episodes.

| | This tier |
| Role | Recommended tier for hard episodes |
| Rate | 100% of base |
| Long scripts | Read in two passes automatically |
| Format checks | Retried once before a failure is reported |

## FAQ

### Is Recommended the same as default?

No. The default for text stays the cheaper tier; Recommended is the badge on the tier the site suggests for hard episodes.

### Should I set it as my default?

If most of your episodes are long or crowded, yes; otherwise switch per episode and keep the cheaper default.

### Does it change how the script is read?

A script is read on whichever tier you had when the read ran; switching later affects new tasks only.

### Evolving or Pro for a failing episode?

Either; same rate. Try the other one if the first fails twice.

Updated November 3, 2026


---

# Download: the Original vs the Play Version | SceneMixer

URL: https://scenemixer.com/guides/download-original-vs-play-version/
Every finished clip exists twice: a lighter play version for the player and share page, and the original-quality file that Download and the episode assembly use.

# Download: the Original vs the Play Version

Quick answer: When a segment is delivered, two files are kept. The play version is what the player on the site and the share page stream: same resolution, smaller, quick to start. The original is the full-quality file: it is what Download gives you, and what the assembled episode is built from. So the download is larger than what you watched, and looks at least as good. Download sits under the player for a segment and on the assembled episode’s row for the full cut; on a phone, Share… hands the file to another app.

What comes out lists the resolution, frame rate and audio; this page is only about the two files.

## Two files per clip

The player streams the play version so a segment starts at once and scrubs smoothly. Download fetches the original. Both come from the same generation; nothing is re-rendered when you download. If the play version is missing for a clip, the player falls back to the original, which starts a little slower.

## Where Download is

A segment: select it on the timeline; Download is under the player next to Share. The assembled episode: Composite History lists every cut; each row has play, Download and Share. The asset library: Assets → Videos lists every clip across projects with a download on each. Assembling and sharing.

## The badge

Whether the file carries the corner badge was decided when the clip was generated. On a free SceneMixer account the original and the play version both carry it; on a paid plan with the badge switched off, neither does. Downloading does not add or remove it. The badge switch.

## Sizes

Roughly: a fifteen-second 480P segment is a few megabytes as an original and less as a play version; 720P and 1080P scale up. An assembled ninety-second episode at 720P is tens of megabytes. The numbers depend on the picture, so check the file after downloading.

## If Download plays instead of saving

Some browsers open a video inline when the link is on another domain. The Download button uses an address that asks the browser to save; if yours still plays it, right-click the player and choose save, or use Share… on a phone. If a video will not play or download.

| | Play version | Original |
| Used by | The player on the site, the share page | Download, the assembled episode |
| Resolution | Same as the original | As generated |
| Size | Smaller | Larger |
| Badge | As decided at generation | As decided at generation |

## FAQ

### Which file does Download give me?

The original-quality file, not the lighter one the player streams.

### Why is the download bigger than what I watched?

You watched the play version; the download is the original.

### Is the assembled episode built from originals?

Yes. The play versions are only for watching on the site.

### Does downloading remove the badge?

No. The badge was decided when the clip was generated.

### Where do I download the whole episode?

Composite History in the editor: each cut has a Download.

Updated November 3, 2026


---

# Draft Cheap, Finish Sharp: When It Saves Credits and When Not

URL: https://scenemixer.com/guides/draft-at-480p-finish-at-1080p/
Drafting on the default tier adds one extra pass that pays back only through retakes. The arithmetic for a higher-tier finish, and when to go straight to the final tier.

# Draft at 480P, Finish at 720P or 1080P: When It Saves Credits, When It Does Not

Quick answer: Drafting means generating every segment at Wan 3.0 · 480P (10 credits a second), fixing the storyboard and the sheets where a shot is wrong, and only then regenerating the segments you keep at Wan 3.0 · 720P (19/s) or Wan 3.0 · 1080P (38/s). The draft pass is one extra full pass at the cheap rate, and it pays back only through retakes: each retaken second costs the draft rate instead of the final rate. Finishing at 720P, the gap between the two rates is too small for that to pay back unless most segments need more than one retake; finishing at 1080P, a few retakes make it roughly even and more retakes make it cheaper.

On the timeline, hover a segment that has been generated twice: the version switch shows both takes and their tiers.

## The workflow

1. Generate every segment at the default 480P tier. 2. Watch each one and note what is wrong — a face, an action, a room, a line. 3. Fix the cause where it lives: the sheet in Step 2, the description, the shot text, the state pick. 4. Regenerate the fixed segments, still at 480P, until the shot is right. 5. Switch the selector to 720P or 1080P and regenerate the segments you keep, once. Since October 2026 there is a second way to finish a segment you want to keep exactly as it is: Enhance raises the same take one resolution step instead of making a new one (enhance or regenerate). 6. Composite. The mistakes are found at the cheap tier; the expensive tier is paid once per segment — at the price of paying for the whole episode once more at the cheap tier.

## The numbers for a ten-segment episode

Ten segments of 12 seconds is 120 seconds of video. Drafted at 480P: 120 × 10 credits. Say four segments need a second draft: another 48 × 10. Then a final pass of all ten at 720P: 120 × 19. Compare generating straight at 720P and redoing those four there: 168 × 19. Here the draft route costs more: the extra 120 seconds at 10 outweigh the four retakes it made cheaper. Run the same episode to 1080P and the two routes come out roughly even: 120 × 38 plus the drafts against 168 × 38. The rule: the draft pass costs the whole episode once at the draft rate, and saves the difference between the two rates on every retaken second. It wins in credits when the final tier costs several times the draft tier and you expect several retakes. A full episode budget, line by line: Where Your Credits Go in One Episode.

## What the draft cannot tell you

480P shows composition, action, faces, timing and dialogue — everything that comes from the storyboard. It cannot show fine texture, and a stylised look flattens more at 480P than a realistic one. Judge the shot, not the pixels: if the face and the blocking are right, the 1080P take will be the same shot, sharper. If the face drifts at 480P, it will drift at 1080P too — My Character’s Face Changed.

## Versions, and which one the composite uses

Every generation of a segment is kept as a version. The timeline shows the active one; switch between them from the segment’s version control, and the composite joins the active version of each segment. So a 480P draft can stay active for a segment you are still deciding on while others are already at 1080P. Downloads follow the active version too. Composite and Share an Episode.

## When to skip the draft

A single short segment, a shot you have already generated successfully in a neighbouring episode, or a project where the first draft came out right for every segment: go straight to the final tier. Finishing at 720P, drafting is a working method, not a saving: it keeps the look-and-fix loop cheap per attempt, and costs a little more overall. Finishing at 1080P with a new cast or a new set, where several retakes are likely, it is the cheaper route.

| Route | Seconds generated | Where the money goes |
| Draft 480P, retake four, final 720P | 120 + 48 at 480P; 120 at 720P | More than going straight to 720P |
| Straight to 720P, retake four | 168 at 720P | Cheaper here: the rates are close |
| Draft 480P, retake four, final 1080P | 120 + 48 at 480P; 120 at 1080P | Roughly even with straight 1080P; cheaper with more retakes |
| Straight to 1080P, retake four | 168 at 1080P | Every mistake paid at 38/s |

## FAQ

### Can I upscale the 480P draft instead of regenerating?

Yes, since October 2026: Enhance under the player keeps the same take and raises it one step, 480P to 720P. A higher tier is a new generation of the shot instead. Enhance has its own price, shown on its card.

### Will the 1080P take look the same as the 480P draft?

Same storyboard, same sheets, same references, so the same shot in most cases; small differences in motion are normal.

### Do drafts count against anything?

Only credits. There is no count limit on versions; SceneMixer removes versions that are not active and older than 90 days.

### Is 720P enough for phones?

Yes for vertical short drama on phones; 1080P for landscape or larger screens.

### Does the composite cost depend on the tier?

No. SceneMixer prices the composite per second of the episode, whatever tiers the segments used.

Updated November 3, 2026


---

# Drone Shot AI Video: How to Write an Aerial or FPV Shot

URL: https://scenemixer.com/guides/drone-shot-ai-video/
Get a drone shot in AI video by writing it: the aerial line that works, an FPV run cut into short shots, which segment length fits, and what cannot be done.

# Drone shot AI video

Quick answer: There is no drone mode and no flight-path editor, so a drone shot is something you write into the shot: one high wide framing, one movement, one place, for example High wide shot, moving forward over a fishing harbour at dawn. An FPV run is the same idea at speed, but it has to be cut into short shots of two or three seconds, because a move that changes halfway through a line reads as two moves. Test the shot as its own segment on the default tier, and say plainly which way the camera goes: a very slow move can come out still.

A shot to paste into the storyboard and try: [Shot 1] 5s: High wide shot, moving forward over a fishing harbour at dawn; small boats below, mist on the water. Ambient: gulls and a distant engine.

## A drone shot, written as a shot

A storyboard shot is a marker with its seconds, a framing, whether the camera moves and how, who or what is in the picture, the one thing that happens, and the sound. A drone shot is the framing high wide shot plus a movement, over a place. It is the “aerial” row in the camera-moves guide, and every line below follows that grammar: one framing and one movement per shot.

| The drone shot | Write it as | Seconds |
| Flyover that opens a scene | High wide shot, moving forward over a fishing harbour at dawn; small boats below. | 4–6 |
| Rise to reveal | High wide shot, the camera rising straight up from the cliff edge; the whole bay opens out below. | 4–5 |
| Pull-back reveal | Wide shot, pulling back from the lone figure on the roof to show the empty city around them. | 3–5 |
| Orbit | High angle, the camera circles the lighthouse once, slowly; waves below. | 4–6 |
| Overhead track | Shot straight down; the camera moves along above @Mara as she runs, the street sweeping backward beneath her. | 3–5 |
| Descent into a place | Wide shot, the camera descending toward the market square; stalls and awnings growing larger. | 3–4 |
Write what the viewer sees and where things are. “Slight” and “very slow” moves tend to come out as a still shot, so make the move clear enough to see and say which way it goes. How to write one shot.

## Two built-in moves that come close

Of the eight special camera moves the storyboard can apply, two are straight down from above, which is the nearest thing to a drone move that is built in:

- Overhead track. The camera moves along above one person, keeping them centred while the ground sweeps backward beneath them. It is meant for someone walking away alone, running or being chased; the shot needs at least 2.5 seconds.
- Overhead spin and rise. Shot straight down, the camera turns slowly clockwise while rising, and the person shrinks in the middle of the frame. It is meant for someone lying on the ground and helpless; at least 3 seconds.
Both need one clear subject in frame, and both are applied by the storyboard at matching story moments rather than chosen from a menu. They are part of the shot text, so you can read them, edit them or take them out. All eight special moves.

## FPV: a run is short shots, not one flight

A real FPV drone flies one unbroken line through gaps and around corners. Here a move that changes direction mid-line reads as two moves, and two shots of the same place at the same framing look like a skip, so an FPV run is written as a short chain, each shot with one movement and a different framing from the one before:

| Shot | Write |
| 1 | [Shot 1] 3s: Low tracking shot, moving fast forward along an empty alley between warehouse walls. Ambient: wind rushing past, a distant siren. |
| 2 | [Shot 2] 2s: High angle, the camera diving through the open loading-bay shutter. Ambient: the wind drops to an echo. |
| 3 | [Shot 3] 3s: Wide shot, low and handheld, moving forward across the warehouse floor between stacked crates. |
| 4 | [Shot 4] 2s: Close-up from a low angle, rushing toward the lit window at the far end. Ambient: the siren, louder. |
That is ten seconds of “flight” made of four cuts, and the cuts are the honest price. If you need it unbroken, write one shot with one move (moving fast forward along the alley) and hold it for the seconds it needs; the segment can only be as long as the tier allows (below). Two seconds is the usual floor for a shot; shorter than that tends to read as a flicker. Shot seconds and timing.

## Choose the segment length first

A drone glide wants seconds, and the segment is what is generated at once. The segment length is set per episode in Step 3 (Standard 15s or Long take 30s), and the tier has to be able to deliver it:

| Video tier family | Longest segment |
| Wan 3.0 (480P / 720P / 1080P) | 30 seconds |
| Seedance 2.5 | 30 seconds |
| Seedance 2.0 | 15 seconds |
| MiniMax H3 | 15 seconds |
The same shot text goes to every tier, and none of the tier guides on this site ranks the tiers for aerial shots, so this page does not claim one looks better. What it does say: draft the shot on the default tier at 10 credits a second, judge whether the camera went where the line said, and only then spend a sharper tier on it. Which tier for what · 15 or 30 seconds · Draft cheap, finish sharp.

One more rule shapes a flight: a segment stays in one place and one timeline. A glide from the harbour to the old town is two segments, one place each. A segment stays in one place.

## Give the camera a place that can be flown over

A moving camera shows every vague part of a place. The place comes from the scene you set up (its sheet and its description), not from a map, so write where things are: water on the left, the old town on the right, the lighthouse at the far end of the breakwater. The scene style sets the look: Pastoral Healing for fields and villages, Medieval Castle for ramparts and keeps, American Small Town for a main street, Neon Nights for a city after dark. Your scene image becomes the set · Scene styles, one by one.

## What it cannot do

- No flight controls. SceneMixer has no waypoints, no altitude or speed values and no gimbal settings; the move is the words in the shot.
- No real map. It will not fly over a real address or a real skyline you name; the place is drawn from your scene description.
- No drone footage in. Uploads are references, not source clips; your own flights cannot be edited here. What you can feed in.
- No readable lettering. Street names, signs and map labels are not drawn reliably. Text on screen.
- No free test of the shot. The free 5-second preview is one shot picked for you, once per account, so it will not be your aerial unless it happens to be the opening shot. Generate the aerial as its own segment on the default tier instead. The free preview.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the place the camera flies over as a scene, with where things are written in; name the one person the camera finds, if any. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Move | Camera line | Built in? |
| Flyover | High wide shot, moving forward over the place | Written in the shot |
| Rise reveal | The camera rising straight up | Written in the shot |
| Orbit | The camera circles the subject once, slowly | Written in the shot |
| Overhead track | Straight down, moving along above one person | Yes, at least 2.5 seconds |
| Overhead spin and rise | Straight down, turning clockwise while rising | Yes, at least 3 seconds |
| FPV run | Short low or high shots, one move each, different framing every cut | Written as a chain |

## FAQ

### Can I make a drone shot with AI video?

Yes, as text: write a high wide shot with one movement over a place you have set up. There is no drone mode or flight-path editor.

### How do I write an FPV drone shot?

As a chain of short shots, two or three seconds each, one movement per shot and a different framing every cut. A single unbroken FPV flight is not something the shot text can promise.

### Does SceneMixer have a drone camera move built in?

Not by that name. Two of the eight built-in moves are straight down from above: the overhead track and the overhead spin and rise.

### Can I set the altitude or speed?

No. There are no numeric flight controls; the movement is described in words in the shot.

### Can it fly over a real city or my address?

No. The place is drawn from your scene description and sheet, not from a map.

### Why did my aerial come out as a still shot?

A very slow or vague move can read as no move. Make it clear and say which way it goes, then regenerate that segment.

### How long can a drone shot be?

A segment can run up to 15 seconds, or up to 30 on the tiers that take long takes; the shot has to fit inside one place and one segment.

### How much does it cost?

Video is priced by the second: 10 credits a second on SceneMixer's default tier, plus 1 credit a second to join the segments.

### Is it free to try?

New SceneMixer accounts start with 50 credits. The storyboard is free to write and edit; the free preview is one shot picked for you, once.

Updated November 3, 2026


---

# How to Fix One Shot Without Redoing the Episode | SceneMixer

URL: https://scenemixer.com/guides/edit-one-shot-and-regenerate/
Editing the storyboard is free, regenerating one segment costs one segment, and every version is kept. What each button actually changes.

# Fixing One Shot Without Redoing Everything

Quick answer: Open the episode, click into the segment you want to change, edit the text, save, and regenerate that one segment. Editing costs nothing; only the regeneration is charged, and only for that segment. Every version you shoot is kept, so you can compare and switch back. The one button that throws work away is “regenerate the storyboard” — that rewrites the whole episode's segments.

The three-pane editor is the same one every project uses. Open a finished episode to see the storyboard text, the shot list and the preview side by side.

## What the editor is showing you

Left is the asset library for this episode — the characters and locations this episode draws on. Middle is the storyboard for the segment you have selected: the shot-by-shot text with durations. Right is the video for that segment, or the button to generate it. Along the bottom is the timeline of every segment in the episode.

The storyboard text is a working document, not a log. It is what gets turned into the video prompt, so editing it is how you change the picture.

## Editing the text

Hit edit, change what you want, save. SceneMixer charges nothing and calls no model — your edit is re-read locally and only the shots you actually touched are re-interpreted.

In a non-Chinese interface the structural labels appear as chips carrying their original values. Type around them freely; the chips keep the machine format intact underneath so the downstream rules still recognise the shot.

The shot durations in the text are real — they add up to the length of the segment. Cutting a line of dialogue without adjusting the seconds leaves a shot with dead air in it.

## Regenerating one segment

The generate button on the segment is priced by that segment's duration, and the price is on the button before you press it. It does not touch any other segment.

Changing the script on SceneMixer does not change a video you have already shot — the shot you have is the shot that was generated from the old text. Regenerate it to pick up the edit.

## Versions

Every generation of a segment is kept. The history control on the segment lets you flip between them and put a different one back in the episode, which is also what the final composite will use.

This is why regenerating a segment is safe: a worse result never costs you the better one you already had. Regenerating the storyboard is a different button, and it is not safe — see below.

## The one destructive button

Regenerating the storyboard for an episode replaces all of its segments — and the videos you have already shot for that episode go with them, along with their version history. Those credits are not refunded. SceneMixer refuses the request and asks you to confirm first, but there is no undo once you do.

Per-segment choices you made by hand, such as which state image a particular segment uses, are recalculated. Anything assigned at the episode level, and any episode you have already composited, is unaffected.

Use it when the breakdown itself is wrong. When one shot is wrong, edit that shot.

## Two things worth knowing

You can insert a segment or delete one, and neither costs anything — no model is called. A new segment starts out empty, so write its shots before you generate it. A segment whose video is still generating cannot be deleted, which stops you paying for a shot and then throwing it away.

If you delete an episode or a project while a video is still generating, the ones that never started are refunded and the ones already running are charged — the model has already been called by then, and that call is already paid for upstream.

| Action | Cost | What it touches |
| Edit the storyboard text | Free | One segment's text |
| Regenerate one segment | That segment's duration | One segment |
| Switch to an earlier version | Free | Which version is active |
| Insert or delete a segment | Free | One segment |
| Regenerate the storyboard | One storyboard fee | Every segment in the episode |

## FAQ

### Does editing the script cost credits?

No. Editing is free and involves no model call. Only generating video is charged on SceneMixer.

### Will my edit change the video I already have?

No. Regenerate that segment on SceneMixer to pick up the change; until then you keep the shot you had.

### Can I get an old version back after regenerating?

Yes. SceneMixer keeps every generation, and you can switch back at any time.

### What happens to my videos if I regenerate the storyboard?

They are deleted along with the segments, and the credits are not refunded. SceneMixer asks you to confirm first. An episode you have already composited is not affected.

### What happens to my manual choices if I regenerate the storyboard?

SceneMixer recalculates per-segment picks from the new text. Episode-level assignments survive.

### Can I change the duration of a segment?

The shot seconds in the storyboard on SceneMixer add up to the segment length, so editing them changes it. Keep the shots and the total consistent or you will get dead air.

Updated November 3, 2026


---

# Edit This Clip: Every Message in Trim, AI Visuals and Redub

URL: https://scenemixer.com/guides/edit-this-clip-messages-and-what-they-mean/
Wait for this clip to finish, another change still running, made by an older pipeline, too much dialogue, choose which half: every message in Trim, AI visuals and Redub.

# Every Message in the Edit This Clip Panel, and What to Do

Quick answer: Most messages in Edit this clip mean “not yet”: the clip is still generating, another change is running, or the version on screen just changed. A few mean “not this clip”: it was made by an older version of the site, or it has no dialogue. Whenever a fix does not go through, your current version is untouched and any credits held for it come back. The Enhance tab’s own messages are listed in Enhance a clip.

If a tab says something unexpected, click another segment and back: the panel re-reads the clip's state.

## While something else is happening

- Wait for this clip to finish, then edit it. / …then trim it. — the segment is generating. The tabs open again when the new clip lands.
- Another change to this clip is still running. — an AI visuals, Redub or Enhance change to the same clip has not finished (perhaps started in another tab). Wait for it; Trim still works, except while an Enhance is running.
- This clip is already being trimmed. Wait a few seconds. — one trim per segment runs at a time.
- This clip is already being changed. Wait for that to finish. — the same for AI visuals and Redub.
- The version on screen just changed. Check the player, then trim again. / This segment just switched versions. Pick again. — someone (you, in another tab) switched versions, so the fix would have been made from a clip you are not looking at. Look at the player and press again.
- The episode’s storyboard is being generated — wait for it; clips cannot be changed while their segments may be rebuilt.

## This clip cannot be changed this way

- This clip was made by an older pipeline — regenerate it first. — a few clips from an earlier version of the site cannot be used as a starting point. Regenerate the segment, then fix the new clip.
- AI edit isn’t available for this clip. — AI visuals is paused for the moment, or the panel could not load. Try again later; Trim is unaffected.
- This segment has no dialogue. — Redub lists only lines written as dialogue in the storyboard. How to write dialogue lines.
- This clip is too short to trim. — one second or less.
- This change isn’t available for this clip. — for example Re-perform on a clip longer than 15 seconds. Use Lip sync.

## Something to fix in your choice

- Keep at least 1s / That range can’t be trimmed. Keep at least 1 second and try again. — move the handles so more is kept.
- This clip is longer than 15 s — choose which half to edit. — pick First half or Second half. Why.
- Too much dialogue for this clip’s length. Shorten the lines. — the new lines would take longer to say than the clip lasts.
- Lines can’t contain brackets or line breaks. — remove them from the text box.
- Pick a voice for every line. — a voice pill is orange; choose a voice for that line.
- That edit isn’t available. Pick one from the menu. — choose one of the eight AI visuals changes again.

## Limits and pauses

- You’ve trimmed a lot this hour. Try again a bit later. / Too many edits in a short time. Please try again later. — an hourly limit per account. Limits that are not about credits.
- Video generation paused after repeated content blocks — SceneMixer pauses AI visuals and Redub together with generation. The pause rules.
- Trimming is temporarily unavailable., AI edit is temporarily unavailable., Redub is temporarily unavailable. — switched off for the moment; nothing was charged.
- The buy-credits window opens — your balance does not cover the price on the button (a trim is 5). Top up and press again. Buy credits.

## After you press

- Trimming…, AI is changing the visuals…, Redubbing… — working; you can leave the page, it carries on.
- Trimmed and saved as a new version., Edited and saved as a new version., Redubbed and saved as a new version. — done; see Version History.
- The new lines were also written into the storyboard. — your changed words are now in the segment’s text.
- The trim didn’t go through. Your current version is unchanged, so you can try again. (and the same for edits and redubs) — the credits come back.

| Message | Means | Do |
| Wait for this clip to finish… | Still generating | Wait |
| Another change to this clip is still running. | An edit, redub or enhance is running | Wait; Trim still works unless it is an enhance |
| The version on screen just changed… | Versions switched meanwhile | Check the player, press again |
| This clip was made by an older pipeline… | Old clip | Regenerate, then fix |
| This clip is longer than 15 s… | Needs a half | Pick First or Second half |
| Too much dialogue for this clip’s length… | Lines too long | Shorten them |
| …didn’t go through… | Failed | Credits come back; try again |

## FAQ

### Can I leave the page while a fix runs?

Yes. SceneMixer runs it on the server; the result is in Version History when you come back.

### Why does Trim work when AI visuals or Redub is waiting?

Trim only cuts the existing file, so it does not wait for an AI visuals or Redub change to finish. Enhance is the exception: while an Enhance is running, Trim waits too.

### A message is not listed here.

See every error message on the site, or ask in the Ask window.

Updated November 3, 2026


---

# Edit This Clip: Trim, AI Visuals or Redub — Which Fix to Use

URL: https://scenemixer.com/guides/edit-this-clip-trim-ai-visuals-or-redub/
Want changes to a finished AI video clip? Start here. Trim it, change what is in the picture, or change the lines and voices, without regenerating. Which fix to use.

# Edit This Clip: Trim, AI Visuals or Redub, and When to Regenerate Instead

Quick answer: Under the player of every finished segment sits Edit this clip, with four tabs. Trim cuts the start, the end or a moment in the middle. AI visuals changes what is in the picture — clothes colour, time of day, weather, background people, lamps, a scar, an expression, or who plays a part. Redub changes the spoken lines or the voices. Enhance keeps the take and raises its resolution one step. Each one saves a new version of the segment; the clip you started from is kept. None of them can move the camera, change the framing or move people around — for that, edit the storyboard and regenerate.

Open any episode with finished segments, click a segment, and look under the player on the right: the four tabs are there as soon as the segment has a clip.

## Where it is

On a computer: in the episode editor, the right-hand pane holds the player. Directly under it is the Edit this clip panel with four tabs: AI visuals, Redub, Trim and Enhance. It appears once the selected segment has a finished clip and opens on Trim; click another tab to switch. The panel belongs to the segment you have selected — pick another segment on the timeline and it follows.

On a phone: open the Preview tab. Under the player there is a row labelled Edit this clip with four buttons. Trim turns the strip under the player into a trimming track while the player stays in view; AI visuals, Redub and Enhance open a sheet from the bottom of the screen. Using SceneMixer on a phone.

## Which tab fixes what

Trim is for timing problems: the clip starts a beat too early, runs on after the last line, or has a stumble in the middle. Nothing new is generated: what you keep is the picture you already had. Each trim costs 5 credits. Trim a clip.

AI visuals is for the content of the picture when the shot itself is right: the dress should be red, the scene should be at night, it should be raining, the street should be empty, the lamps should be lit, the hero should have a scar, the face should look terrified, or a different character should be in that place. It re-renders the clip from itself and costs 2× the per-second price of the tier the clip was made on. What AI visuals can change.

Redub is for the sound of the dialogue: a line should say something else, or a character should have a different voice. Lip sync keeps the picture and moves the mouths to the new audio (1 credit per second of clip); Re-perform plays the clip again with the new lines, at the same price as AI visuals. Change a line after the video is made.

Enhance is for a clip that is right but looks soft: it raises the same take one resolution step (480P to 720P, 720P to 1080P, up to the 2K top step) with Instant HD, AI Detail or Cinema Restore. It is priced per second by option and target resolution; the price is on each card. Enhance a clip.

## When to regenerate instead

All four tabs start from the clip you already have, so they keep how it was shot. They cannot change the framing (close-up versus wide), the camera angle, where people stand or which way they face, and they cannot remove a person who should not be there or fix text that appeared in the picture. Those come from the storyboard: edit the shot and generate the segment again. When a shot comes out wrong and How to fix one shot.

Price matters too. SceneMixer charges AI visuals at twice the per-second price of the clip’s tier, so on the higher tiers one edit can cost more than generating the segment again on a cheaper tier. What each fix costs.

## What happens to the clip you started from

Nothing. Every fix is saved as a new version of the segment, labelled Trimmed, AI edit, Redubbed or Enhanced in Version History. If the clip you edited is still the one showing when the fix finishes, the new version becomes the active one straight away; otherwise it waits in the history for you to switch to. A segment keeps one version of each kind: a second trim replaces the first trimmed version, and so on. The generated clip a fix was made from stays for as long as that fix exists. Trimmed, AI-edited and redubbed versions.

## When the tabs are not available

While the segment is still generating, while its episode’s storyboard is being generated, and while another AI visuals or redub change to the same clip is running, those tabs say so and wait. A few older clips, made by an earlier version of SceneMixer, ask you to regenerate them first. Ad projects in advanced mode have the same tabs; in an ad scored the new way, Lip sync replaces the clip’s own soundtrack and the ad’s music is mixed back under the new version by itself. In an ad made before the new scoring, the clips carry the video model’s own music, and a lip-synced clip comes back without it — re-perform that clip or regenerate the segment instead. In sample projects the panel is shown for reference but greyed out. Every message in the panel.

| The problem | Use | Price |
| Starts too early, ends too late | Trim → Keep | 5 per trim |
| A stumble or a dead pause in the middle | Trim → Remove | 5 per trim |
| Wrong colour, time, weather, lamps, scar, expression, crowd | AI visuals | 2× the tier’s per-second price × clip length |
| Wrong actor in the part | AI visuals → Swap person | 2× the tier’s per-second price × clip length |
| A line should say something else | Redub | Lip sync 1 per second, or Re-perform |
| A character sounds wrong | Redub → pick a voice | Lip sync 1 per second, or Re-perform |
| The clip is right but looks soft | Enhance | Per second, by option and resolution; shown on the card |
| Wrong framing, angle, positions, an extra person | Edit the storyboard and regenerate | The normal generation price |

## FAQ

### Do these fixes change my storyboard?

Trim and AI visuals do not. Redub on SceneMixer writes a changed line back into the storyboard text, so the next regeneration says the new line too.

### Can I undo a fix?

Yes. Open Version History on SceneMixer and switch back to the version you started from; switching is free.

### Can I combine them?

Yes, one after another: for example AI visuals first, then Trim on the edited version. SceneMixer charges each on its own.

### Am I charged if a fix fails?

No. The credits are held when you press the button and come back if the fix does not go through. A trim on SceneMixer costs 5 credits; the button shows the price before you press.

### Does the fixed clip carry the AI badge?

A trimmed clip keeps whatever the clip had. AI visuals and Redub follow the same rule as generating: SceneMixer adds or leaves off the badge according to your plan and your AI badge setting.

Updated November 3, 2026


---

# Editing a 30-Second AI Video Segment: First Half or Second Half

URL: https://scenemixer.com/guides/editing-a-30-second-segment-first-or-second-half/
Clips longer than 15 seconds are edited one half at a time: choose the half, it is changed and stitched back to the untouched half. What it costs and how to do both.

# AI Visuals on a 30-Second Segment: Pick the First or the Second Half

Quick answer: AI visuals works on up to 15 seconds at a time, so on a clip longer than that the panel first asks for First half or Second half. Only that half is changed; the other half is not touched, and the two are joined back into one clip automatically. You pay for the half: a 30-second segment is charged as 15 seconds. To change the whole clip, run the second half on the result of the first. Re-perform in Redub is not offered for clips this long; Lip sync is.

A 30-second clip on Wan 3.0 · 480P: one half changed to Night costs 15 × 20 credits; both halves, two edits, twice that.

## Which clips are split

Any clip longer than 15 seconds — in practice the 30-second segments made on the tiers that allow them. A 15-second or shorter clip is edited whole and the choice does not appear. 15-second or 30-second segments.

## How the split works

The clip is divided exactly in the middle. The half you pick is re-rendered with your change; the other half is not touched at all. The two are then joined in their original order, so the result is one clip of the original length, saved as an AI edit version. You see the hint under the choice: Over 15 s — pick the half to edit; it’s stitched back automatically.

## Mind the join

Because only one half changes, anything that should be continuous can jump at the middle. Turn only the first half to night and the light switches mid-clip; recolour a dress in the second half and it changes colour at the cut. That is useful when the change should happen there — the lights go on half-way — and a problem when it should not. For a change that must hold through the whole clip, edit both halves: first one half, then the other half of the new version.

## Doing both halves

- Run the edit on the First half. When it lands, the AI-edited version becomes the active one.
- Open AI visuals again. The panel now works on that new version; choose Second half and the same change.
- The second edit is made from the first one, so the result carries both halves; it replaces the first AI-edited version in the history.

Each run is charged for its half: two runs on a 30-second clip cost the same as one 30-second edit would, 2× the tier’s per-second price × 30.

## Redub on long clips

Lip sync handles the whole clip in one go and is charged per second of the whole clip, 1 credit a second. Re-perform, which re-plays the clip with new lines, is only offered for clips of 15 seconds or less; on a longer clip its option is greyed out. Lip sync or Re-perform.

| Clip length | AI visuals | Charged for | Re-perform |
| Up to 15 s | Whole clip | The whole clip | Available |
| Over 15 s | First or second half | Half the clip | Not available (Lip sync is) |

## FAQ

### Can I choose where the clip is split?

No. It is always the middle. If you only need part of the clip, trim it to 15 seconds or less first; a clip that short is edited whole.

### Does the untouched half lose quality?

Not visibly. It is not generated again; it is joined to the changed half as it is.

### Why can't I re-perform a 30-second clip?

SceneMixer's Re-perform replays the clip from itself, and it can take only 15 seconds at a time. Use Lip sync, or regenerate the segment with the new lines.

### Do I pay twice for a whole 30-second clip?

You pay for each half once, which adds up to the whole clip.

Updated November 3, 2026


---

# Editing a Sheet: Write the Change, Not the Whole Picture

URL: https://scenemixer.com/guides/editing-a-sheet-write-the-change-not-the-picture/
An edit starts from the sheet you already have. Why a full description gives the same picture back, how to write a one-line change, what to do when it changed too much.

# Editing a Sheet or Making a State: Write the Change, Not the Whole Picture

Quick answer: When you edit a sheet (change the hair, swap the coat, add a scar) or generate a state (the drawn blade, the burnt room), the model starts from the existing image and reads your text as what to change. A long description of the whole character or room is read as ‘repaint this’, and the answer is the same picture again. Write one short instruction naming only the difference: hair cut to the jaw, same colour; the same room after the fire: walls blackened, roof beams fallen. One change per edit; identity, framing and everything you did not mention are held by the reference. If the result changed things you did not ask for, the instruction named them somewhere — shorten it. AI edit with your own instruction is a paid-plan feature; states work the same way on any plan.

On a paid plan, open a character card and use AI edit with a five-word instruction, then with a paragraph: the short one changes the thing; the paragraph tends to return the original.

## Why a long instruction gives you the same picture back

Editing is image-to-image: the reference carries the face, the pose, the framing and the style; the text is a delta. When the text restates everything, it agrees with the reference on almost every point, and the model’s best fit to ‘everything the same’ is the original. The site keeps its own edit instructions to a few lines for this reason: the state, the difference, one line holding identity, the aspect ratio.

## How to write a change

Change X to Y. Name the part and its new form; say ‘everything else unchanged’ if you like, but do not list what stays. Structure words work; mood words do not: collar torn, left sleeve missing rather than looks like she has been through a fight. For a room: the same room, ceiling collapsed at the far end, daylight through the hole. For an object: blade drawn, scabbard beside it. AI Edit a Sheet.

## One thing at a time

Hair and coat and scar in one instruction gives the model three places to drift; do them in three edits, checking each. Each edit on SceneMixer is a generation at the sheet price, so plan the order: the change most likely to alter the face last.

## When the edit changed too much

The face moved: the instruction mentioned the face, or age, or expression; remove that. The framing changed: the instruction described the full figure; drop it. The style changed: a style word slipped in; the style is set by the project, not by the edit. Regenerate from the original sheet, not from the drifted result. Your Own Reference Images.

## States follow the same rule

A state’s description is a difference sentence, not a second description of the thing. Write what is structurally different from the main sheet; leave lighting and mood to the shots. One Outfit per Sheet.

| Instruction | Result |
| Hair cut to the jaw, same colour | Hair changes; face, clothes, pose hold |
| A woman in her thirties with a long face, straight brows… (full description) with shorter hair | The original sheet, or the original with a small drift |
| Torn collar, left sleeve missing | The clothing changes as asked |
| Looks like she has been through a fight | Anything: bruises, mess, a new pose |
| The same room after the fire: blackened walls, beams down | The room, burnt, same layout |

## FAQ

### Should I paste the character's description into the edit box?

No. The reference already carries it; the box is for the change only.

### Can I change two things at once?

You can; separate edits are safer and easier to check.

### The edit changed the face.

Remove any word about face, age or expression and regenerate from the original sheet.

### Does the visual style need to be repeated?

No. It is the project's; naming it in an edit can shift it.

### Are state descriptions different?

Same rule: write the difference, in structural words.

Updated November 3, 2026


---

# Editing the AI Storyboard: What You Can Change for Free | SceneMixer

URL: https://scenemixer.com/guides/editing-the-storyboard-what-you-can-change/
Every sentence of a storyboard is yours to rewrite, free. What the chips are, what Save does, why a save can be refused, and the one button that rewrites everything.

# Editing the Storyboard: What You Can Change, and What You Can’t

Quick answer: Open a segment, click Edit Script, change any text, click Save. Nothing is charged and no video is generated. The grey chips — Location, [Shot 1] 4s, a character’s name, a voice tag — are the structure of the script: you can move or delete them, and new ones come from the @ menu, but you do not retype them by hand. A save re-reads only the shots you touched. The video you already have does not change until you regenerate that segment.

Newer storyboards. In storyboards written or regenerated since late October 2026 the header is just Location and a Time · Weather line with a Change link (free; it changes the next video you generate, not the text), and the text has no Opening frame, Set dressing, Props, In frame only or Ambient lines. Notes under the text say what a segment will show and what blocks generation. The segment header, field by field · Notes under a segment.

Open a finished episode and click into a segment. The storyboard in the middle pane is exactly what you would be editing.

## Opening the editor

In an episode, click a segment on the bottom timeline. The middle pane shows its storyboard: a short header (where, when, weather, what is on set), then the shots, numbered, each with its seconds. Click Edit Script above the text to start editing. On a phone you can read and generate, but detailed storyboard editing is a desktop job.

## What the chips are

In an English interface the structural labels appear as small grey chips: Location, Time, Weather, Set dressing, Props, Opening, the shot markers [Shot 1] 4s, and inside a shot things like In frame only, lips closed, Ambient and No dialogue. Character, scene and prop names are chips too, and so is the voice tag after a line of dialogue.

A chip is one unit. You can put the cursor before or after it, delete it, or cut and paste it. You cannot edit the letters inside it. To add a name, a voice tag or a shot duration, use the @ menu — how the @ menu works. If you delete a chip by mistake, Discard puts the segment back the way it was before you started.

## What you can change

Everything that is not a chip. Who does what, the framing, the camera move, the light, the dialogue and how it is delivered, the seconds of each shot (through the chip menu), which characters are in the picture, the header fields. You can add a shot, remove a shot, or split one into two.

Keep two habits. Name every character who is in the picture with an @ chip, so the shot knows who to draw — who is in frame. And keep the shot seconds honest, because they add up to the length of the segment — shot seconds.

## What a save does

Save stores the text. SceneMixer charges nothing, calls no model, and re-reads only the shots you actually changed. Three warnings can appear afterwards, none of them fatal: a name that is not a character, scene or item in this project (no image will be attached for it); a character who has no image yet; a line of dialogue with no voice tag. Fix them or not, as you like.

The video on the right does not change when you save. It is the video of the old text. Regenerate the segment to get one from the new text; every version is kept. If you click another segment with unsaved changes, you are asked whether to save first — Save & Switch does both.

## Three saves that are refused

The script is too long. Trim it. The header could not be reattached safely. The save is cancelled so the script is not corrupted; your draft is kept — copy it, close the editor, reopen and paste. The header does not match the segment being edited. The segment changed under you; close the editor and reopen it. In all three cases nothing is charged and nothing is lost.

## The one button that rewrites everything

Regenerate on a segment remakes that segment’s video and keeps the old one. Regenerating the storyboard of an episode is a different thing: it rewrites every segment’s text and drops the videos with them. Which button does what.

| Action | Cost | What it touches |
| Edit any sentence and save | Free | The text of one segment |
| Delete or move a chip | Free | The structure of one segment |
| Add a name, voice tag or seconds via @ | Free | One shot |
| Regenerate the segment | That segment’s duration | One segment’s video (old version kept) |
| Regenerate the episode storyboard | One storyboard fee | Every segment and its videos |

## FAQ

### Does editing the storyboard cost credits?

No. Editing and saving are free and call no model. Only generating video costs credits on SceneMixer.

### Why can't I type inside a chip?

Chips are the fixed structure of the script. Move or delete them; add new ones from the @ menu.

### I saved, but the video looks the same.

Saving changes the text, not the video. Regenerate the segment to get a video from the new text.

### The save was cancelled with a message about the header.

Copy your draft, close the editor, reopen the segment and paste. Nothing was charged.

### Can I write in my own language?

Yes. The interface labels follow your language, and the storyboard text is shown in the project's display language. Write in the language you are reading.

Updated November 3, 2026


---

# The Episode Editor Layout for Vertical and Landscape Projects

URL: https://scenemixer.com/guides/editor-layout-for-vertical-and-landscape-projects/
The editor arranges itself around your video: tall side columns for 9:16, a full-width timeline for 16:9. Where the player, buttons and timeline sit, and why.

# How the Episode Editor Lays Out Vertical (9:16) and Landscape (16:9) Projects

Quick answer: The editor has three columns — assets, storyboard, player — and a timeline, and arranges them by the project’s aspect ratio. In a vertical (9:16) project the left and right columns run from top to bottom, and the timeline sits under the storyboard in the middle, so a tall video gets the whole height. In a landscape (16:9) project the timeline runs the full width along the bottom and the player fills the right column’s width. Under each segment’s storyboard is one row of buttons; under the player, Edit this clip.

Open a 9:16 project and a 16:9 project side by side: the timeline moves from the middle to the full width.

## The three columns

Left: the episode’s assets — characters, locations, props. Middle: the selected segment’s storyboard. Right: the player for that segment, with its controls and the Edit this clip panel. Timeline: one card per segment. A tour of the editor.

## Vertical projects

A 9:16 clip is tall, so the right column runs the full height of the window and the player is as tall as the column allows, after leaving room for the controls under it. The left column runs full height too. The timeline sits only under the storyboard in the middle. On a wide screen the right column widens to fit the video.

## Landscape projects

A 16:9 clip is wide, so the timeline gets the full width of the window along the bottom, and the side columns stop above it. The right column takes a fixed share of the width and the video fills it edge to edge. Vertical or landscape.

## The row under the storyboard

Every segment’s storyboard card ends with one row: Edit, AI edit and History on the left, Insert and Delete after them, and a single solid button on the right — Generate or Regenerate — with its price. On a narrower window the smaller buttons shrink to icons first, and only then does the row wrap into two lines. AI edit and History.

## Under the player

One line with Play continuously, the version number and length, share and download; on a narrow column share and download become icons. Below it, whenever the segment has a clip, the four tabs of Edit this clip: AI visuals, Redub, Trim and Enhance. Edit this clip.

## The timeline

When there are more segments than fit, small arrows appear at the two ends; click one to scroll about a screen’s width, or scroll with the mouse wheel — it moves sideways. The selected segment is scrolled into view when you change segments, and the timeline stays where you left it otherwise. Every icon on a timeline card.

## Full screen

The player’s full-screen button shows the whole clip, vertical or landscape, with the controls, and Play continuously keeps going from segment to segment in full screen.

## On a phone

The columns become tabs — Assets, Script, Preview. On the Script tab the buttons sit in two rows of three, with the main button’s price under its label. Using SceneMixer on a phone.

| | Vertical (9:16) | Landscape (16:9) |
| Side columns | Full height | Stop above the timeline |
| Timeline | Under the storyboard only | Full width along the bottom |
| Player | As tall as the column allows | Fills the column's width |
| Right column width | Follows the video | A fixed share |

## FAQ

### Can I switch between the two layouts?

On SceneMixer, the layout follows the project's aspect ratio, which is set when the project starts.

### The buttons turned into icons. Where did the words go?

The window is narrow; hover an icon to see its name, or widen the window.

### Why does the timeline sit in the middle in my project?

It is a vertical project; the side columns run full height to give the tall video room.

Updated November 3, 2026


---

# Emails SceneMixer Sends and How to Unsubscribe | SceneMixer

URL: https://scenemixer.com/guides/emails-we-send-and-how-to-unsubscribe/
Every email SceneMixer sends: sign-in and reset codes, a welcome, four reminders, and replies to feedback. What the unsubscribe link turns off, and what it cannot.

# The emails we send, and how to stop the reminders

Quick answer: We send sign-in and password-reset codes when you ask for one, a single welcome email after you sign up, replies to feedback you sent, and four kinds of reminder: a project you read but never turned into video, a balance running low, a plan you stopped from renewing, and a plan about to end without renewing. Every reminder has an unsubscribe link at the bottom that works without signing in; one click stops all four reminders. Codes and replies you asked for keep coming.

How to get help explains feedback and how a reply reaches your inbox. Announcements are not on this list: they stay in the bell and on the announcements page.

## Everything we send

| Email | When | Stopped by unsubscribe? |
| Sign-in code | Each time you ask to sign in with email | No; you asked for it |
| Password reset code | Each time you ask to reset a password | No; you asked for it |
| Welcome | Once, after you sign up with a verified email | Sent only once anyway |
| You read a script but made no video yet | Once per account, a couple of days after a script is read, if the account has never generated a video | Yes |
| Your balance is running low | When the balance drops low while you are actively using the site; at most about once a month | Yes |
| You turned off plan renewal | Shortly after you cancel, while the plan still has time left | Yes |
| Your plan ends soon | A few days before a plan that will not renew ends | Yes |
| Reply to your feedback | When support answers and you chose to receive replies by email | No; you asked for it |
| Reply to an email you sent us | When support answers a message you sent to the support address | No; you asked for it |

## Which language they arrive in

The language you use the site in. If you switch the interface to Japanese, the next reminder arrives in Japanese. Sign-in codes add an English copy underneath when your interface language is not English.

## How the unsubscribe link works

Every reminder carries a link at the bottom. Opening it stops all four reminders for your account at once, without signing in; the page confirms it. Mail apps that show an unsubscribe button at the top of a message use the same mechanism. There is no confirmation email and nothing else to click.

Unsubscribing does not affect your account, your projects, your credits or your plan. It only stops the reminders.

## What it cannot stop

Sign-in and password-reset codes, because they only arrive when you request one. Replies to feedback, because you choose whether you want them each time you send feedback. Receipts from the payment provider, which come from the provider, not from us.

## If you do not receive our emails

Check spam and promotions folders first; sign-in codes are the most likely to be filtered. Add the sending address to your contacts. If a sign-in code still does not arrive after a minute, request another; codes expire, so use the latest one.

## FAQ

### Can I turn the reminders back on after unsubscribing?

SceneMixer has no switch for it yet; write to support from your account’s email and ask.

### Do reminders contain prices or offers?

No. They point you back to your project, your balance or your plan page.

### Why did I get a plan reminder after I cancelled?

Cancelling stops renewal; your plan runs to the end of the period. The reminder is about that remaining time.

### Does unsubscribing delete my account or data?

No. It only stops the four reminders.

### Do you send marketing newsletters?

No. The emails listed on this page are the ones we send.

Updated November 3, 2026


---

# Enhance Clarity: Make a Finished Clip Sharper, One Step Up

URL: https://scenemixer.com/guides/enhance-a-clip-to-720p-1080p-or-2k/
Raise a clip you made on SceneMixer one resolution step: the same take, sharper, no new render. The three options, the price, versions, and what to do if it fails.

# Enhance Clarity: Make a Finished Clip Sharper, One Step Up

Quick answer: Since October 2026, Edit this clip under the player has a fourth tab, Enhance. It takes the clip you already have and raises it one resolution step — 480P to 720P, 720P to 1080P, up to the 2K top step — without making a new take: same framing, same acting, same sound. You choose one of three options, Instant HD, AI Detail or Cinema Restore; each card shows what this segment will cost and roughly how long it takes, and the cheapest option is selected for you. Press Enhance to 720P (or whichever step is next): the credits are held, the result is saved as a new version labelled Enhanced, and the original stays in Version History. It works on segments made on SceneMixer, not on videos you upload.

Open any episode with a finished segment, click the segment, and under the player choose Edit this clip → Enhance. Reading the tab is free; nothing is charged until you press the button.

## Where the Enhance tab is

On a computer: open the episode editor and click a segment on the timeline. The right-hand pane holds the player; directly under it sits Edit this clip with four tabs: AI visuals, Redub, Trim and Enhance. The panel opens on Trim; click Enhance.

On a phone: tap the segment to open its preview; the same four tabs are in the sheet under the player.

The tabs appear once the segment has a finished clip. Before that the panel says Generate this clip first, then edit it here., and while the clip is still being made, Wait for this clip to finish, then edit it. Edit this clip: which fix to use.

## What the tab shows

Top to bottom, the tab is short:

- The step. One line such as Now 480P → enhance to 720P. If the version you are looking at was itself enhanced, the line reads Original 480P → enhance to 720P, above a note like This version was enhanced to 720P with AI Detail.
- Three cards side by side. Each has the option’s name, the price for this segment with the ✦ mark, and an estimate such as about 3 min. A card that cannot take this clip is greyed with a short reason under it.
- One sentence about the selected option, under the cards.
- The button: Enhance to 720P · ✦ N, with the target resolution and the price of the selected card.
- Notes under the button, only when they apply: Enhancing again replaces the previous enhanced version. and, when the step is to 2K, Full-episode exports top out at 1080P.

## The three options

| Option | What it does | Faces | Wait |
| Instant HD | “Quickest — a sharper picture without redrawing anything.” Edges and fine lines get crisper at the higher resolution. | Unchanged: nothing is redrawn | The shortest |
| AI Detail | “AI redraws fine detail; faces may change slightly.” Skin, hair, fabric and small textures are filled in rather than stretched. | Can shift a little; check the close-ups | Several minutes for a typical segment |
| Cinema Restore | “Most thorough restoration — cleans up noise and blur; takes the longest.” For a clip that looks grainy, smeared or soft. | Kept; noise and blur are cleaned around them | The longest |
The exact wait for your segment is on each card. How to choose:

- Posting to a phone feed, the clip already looks right: Instant HD. It is the safe default and the one SceneMixer selects for you when it is the cheapest available.
- The clip looks soft or plasticky at 480P and you want texture: AI Detail. Watch the faces in close-ups afterwards; if a face moved too far from the character, switch back to the original version.
- A hero shot, a noisy night scene, a clip with motion blur you want cleaned: Cinema Restore.
Compare the three prices on the cards before you press: the card you have selected is exactly what is charged.

## Enhance a segment, step by step

- Pick the segment on the timeline and make sure the version you want is the active one (it is the one playing in the player).
- Open SceneMixer’s Edit this clip → Enhance and read the step line: Now 480P → enhance to 720P.
- Choose a card. The cheapest option that can take this clip is already selected; click another card to switch. Its one-line description appears under the cards.
- Press Enhance to 720P · ✦ N. The credits are held, not yet spent.
- Wait or leave. The panel says Enhancing clarity… with a progress bar sized to the estimate on the card. You can work on other segments or close the page; it keeps running.
- Done: Enhanced and saved as a new version. If the version you enhanced was still the active one, the enhanced version becomes active and plays straight away. If you switched to another version in the meantime, the new one waits in Version History; press Use on it when you want it.

## One step up, always from the original

The step is worked out from the clip’s original resolution: 480P goes to 720P, 720P to 1080P, up to the 2K top step. A clip of an unusual size goes to the next standard step above it. SceneMixer has no menu to pick a bigger jump.

Enhancing an enhanced version again does not stack. The tab says Original 480P → enhance to 720P, works from the original again, and the new result replaces the earlier enhanced version — useful for trying AI Detail after Instant HD, not for climbing from 480P to 1080P. To end at 1080P from a 480P clip, regenerate the segment on a 1080P tier, or regenerate it at 720P and enhance that to 1080P. Enhance or regenerate at a higher tier.

A clip that is already 2K shows This clip is already at the highest resolution (2K). and there is nothing to enhance.

## What it costs, and when it is charged

SceneMixer prices Enhance per second of the clip, at a rate that depends on the option and on the resolution you are going to; the segment’s length is rounded up to whole seconds. The number on the card and on the button is the full price for this segment — there is nothing else to add. There is no free allowance, and it is open to every account, not only members.

- On pressing: the credits are held.
- When the enhanced version is delivered: they are charged. Your credit history shows one line, Enhance clarity · AI Detail (or the option you chose).
- If it fails: see the last section — most failures refund automatically.
- If the price or the active version changed since you opened the tab (another tab or device switched versions), nothing is charged; the tab refreshes so you can look again and press once more.
Compare it with the alternative before you press: regenerating the same segment at Wan 3.0 · 720P costs 19 credits a second, and the price is on the Generate Video button when that tier is selected. What each clip fix costs.

## How it fits with Trim, AI visuals and Redub

| Situation | What happens |
| An enhance is running on this segment | The Trim tab shows Another change to this clip is still running. AI visuals and Redub wait too. Finish or wait for the enhance first. |
| An AI visuals, Redub or Trim job is running | Enhance waits with the same message. |
| You trim an enhanced version | The trimmed version stays enhanced; you are not charged for enhancing again. |
| You use AI visuals or Redub on an enhanced version | They work from the original clip, so the new version is not enhanced. The tab then says This version was edited from the original clip, so it isn’t enhanced — you can enhance it again. Nothing is charged automatically. |
The practical order: fix what is in the picture and the sound first (regenerate, AI visuals, Redub), enhance last. Trim can come before or after. Trim a clip.

## Versions, downloads and the episode

- SceneMixer’s Version History lists the enhanced take with the label Enhanced. Use on the original switches back, free and instantly. Each segment keeps one enhanced version; enhancing again replaces it.
- Download and share follow the active version: if Enhanced is active, the download is the enhanced file.
- Compositing joins the active version of every segment. Enhanced and un-enhanced segments can sit side by side; the episode is assembled at one size, and the full-episode export tops out at 1080P — a 2K segment downloads at 2K on its own and joins the episode at 1080P.
Segment versions · Which segments are joined.

## When it cannot be done, or fails

| You see | What it means | What to do |
| This clip is already at the highest resolution (2K). | There is no step above it. | Nothing; download it as it is. |
| We can’t find a clean original of this clip, so it can’t be enhanced. Regenerate it first. | Some older clips do not have the original file enhance works from. | Regenerate the segment, then enhance the new take. |
| This option can’t process a clip of this size (on one card) | That option has size limits this clip is outside of. | Pick another card. |
| This clip’s size is outside what enhancement can process. | No option can take it. | Regenerate on the tier you want instead. |
| Couldn’t read this clip’s size right now. Please try again in a moment. | A temporary hiccup reading the file. | Wait a minute and open the tab again. |
| Not available right now (on a card) / This enhance option isn’t available right now. | That option is paused. | Use another card, or try later. |
| Enhance clarity isn’t available right now. | The whole feature is paused. | Try later. |
| Another change to this clip is still running. | A trim, AI visuals, Redub or enhance is in progress on this segment. | Wait for it to finish. |
| The enhance didn’t go through — your credits have been refunded. | It failed and nothing was delivered. | Try again, or try another option. |
| We hit a problem while processing it. It’s been passed to our team for review, and the credits are on hold for now. | The outcome is not clear yet. | Do not resubmit; the hold is resolved to a refund or a delivered version. |
An enhance has no Stop button, like any clip job. Deleting the segment while it runs does not refund it: the work is done and charged, it just has nowhere to land. Credits held for review.

## What enhance does not do

- It does not fix a wrong face, a wrong action, a wrong room or a wrong line. Those come from the storyboard and the images; fix them there and regenerate. When a shot comes out wrong.
- It does not change framing, motion, timing, voices or the episode’s music.
- It does not take videos from outside SceneMixer, and it does not raise the composite above 1080P.
- It does not change which tier the segment was generated on; regenerating later uses whichever tier you pick then.

| Clip is now | Enhance takes it to | Note |
| 480P | 720P | The usual first step for clips made on the default tier |
| 720P | 1080P | Same as a 1080P tier’s size; the take itself is unchanged |
| 1080P | up to the 2K top step | The episode export still tops out at 1080P |
| 2K | — | Already the highest; the tab says so |

## FAQ

### Can I enhance any video, or only clips made on SceneMixer?

Only segments generated on SceneMixer, from inside the episode editor. Uploaded videos, including the source clip of a remake, cannot be enhanced.

### Does enhancing change the acting, the voices or the lip movement?

No. The take on SceneMixer is the same: framing, motion, timing and sound are kept. Instant HD redraws nothing; AI Detail redraws fine detail, so a face can shift slightly; Cinema Restore cleans noise and blur.

### Can I go from 480P straight to 1080P?

Not with enhance. SceneMixer always goes one step up from the clip's original resolution, and enhancing again replaces the earlier enhanced version instead of stacking. For 1080P, regenerate on a 1080P tier, or regenerate at 720P and enhance that.

### How much does it cost?

It is priced per second of the clip, by option and by the resolution you are going to. The price for this segment is on each card and on the button, and that is exactly what is charged. SceneMixer has no free allowance for it.

### Is it only for members?

No. Every SceneMixer account sees the Enhance tab once a segment has a finished clip.

### Can I go back to the original?

Yes. The original stays in Version History; press Use on it. Switching versions on SceneMixer is free.

### Why did my AI visuals edit lose the enhancement?

SceneMixer’s AI visuals and Redub work from the original clip, so their result is not enhanced. Enhance the new version if you want it sharper, or do those fixes first and enhance last.

### Will the whole episode be 2K if I enhance every segment to 2K?

No. SceneMixer's full-episode export tops out at 1080P. Each 2K segment downloads at 2K on its own.

### Can I stop an enhance?

No. Like other SceneMixer clip jobs it runs to the end. If it fails, the credits come back automatically or are held for review; deleting the segment while it runs does not refund it.

### Which option should I pick?

Instant HD to sharpen a clip that already looks right, AI Detail when a 480P clip looks soft and you want texture, Cinema Restore for a grainy or blurred clip or a hero shot. Compare the prices on SceneMixer’s cards.

Updated November 3, 2026


---

# Enhance the Clip or Regenerate at a Higher Tier: Which to Pick

URL: https://scenemixer.com/guides/enhance-or-regenerate-at-a-higher-tier/
Two ways to a sharper segment: enhance the take you like one step up, or generate a new take on a higher tier. What changes, what it costs, which to pick.

# Enhance the Clip You Have, or Regenerate It at a Higher Tier?

Quick answer: There are two ways to get a sharper segment. Enhance (in Edit this clip) keeps the take you already have — the same framing, acting, lips and sound — and raises its resolution one step: 480P to 720P, 720P to 1080P, up to the 2K top step. Regenerating on a higher tier makes a new take at that resolution from the same storyboard and images: sharper from the start, but the performance comes out differently. Enhance when you love the take and only the softness bothers you; regenerate when something in the take is wrong, or when you need more than one step up. Both keep the old version, so you can always switch back.

On a finished 480P segment, open Edit this clip → Enhance and note the price on the cards; then pick Wan 3.0 · 720P in the tier picker beside Generate Video and note that price. Neither is charged until you press.

## The difference in one table

| | Enhance | Regenerate on a higher tier |
| What you get | The same take, at the next resolution step | A new take, rendered at the tier’s resolution |
| Framing, motion, timing | Unchanged | Comes out differently each time |
| Dialogue, voices, lips | Unchanged; the sound is not touched | Performed again; a line can land differently |
| Faces | Instant HD redraws nothing; AI Detail may shift a face slightly; Cinema Restore cleans noise and blur | Drawn again from the character images |
| How far up | One step from the clip’s original resolution, up to 2K | Whatever the tier makes: Wan 3.0 · 720P, Wan 3.0 · 1080P and others |
| Price | Per second, by option and target resolution; shown on each card | Per second at the tier’s rate; shown on Generate Video |
| Where | Edit this clip → Enhance | Tier picker beside Generate Video, then regenerate |
| Old version | Kept in Version History | Kept in Version History |

## Enhance when the take is right

A generation is partly luck. When a segment finally has the look on her face, the timing of the door and the line delivered just so, a new generation — even at a higher tier — can lose all of it. Enhance keeps the take and changes only the picture’s resolution and detail. Pick it when:

- You are happy with everything except that the clip looks soft at 480P.
- The segment took several tries to get right and you do not want to roll the dice again.
- You want a 2K version of a hero shot for a trailer or a thumbnail; none of the Wan 3.0 or Seedance tiers on SceneMixer makes 2K, enhance does (MiniMax H3 generates at 2K, but that is a new take with its own rules).
- The clip is noisy or blurred and you want it cleaned (Cinema Restore).
How to use Enhance.

## Regenerate when something in the take is wrong

Enhance does not fix content. A wrong face, a wrong action, a wrong room, a mispronounced name or a line delivered badly will be exactly as wrong, only sharper. Regenerate when:

- The shot itself is wrong. Fix the cause in the storyboard or the images first, then regenerate. When a shot comes out wrong.
- You want two steps or more: from 480P to 1080P in one go. SceneMixer’s Enhance only goes one step from the original, and enhancing again replaces the earlier enhanced version instead of stacking.
- You need something only a different tier does, such as longer segments or another tier’s strengths. Which model for which purpose.
- The clip is old enough that Enhance says We can’t find a clean original of this clip, so it can’t be enhanced. Regenerate it first.

## How the prices compare

Both are priced per second of the segment, and both show the full price before you press, so the comparison is two numbers on your screen:

- Enhance: the card you select in Edit this clip → Enhance. The rate depends on the option and on the resolution you are going to.
- Regenerate: the price on Generate Video with the higher tier picked. For SceneMixer's Wan 3.0 tiers that is 19 credits a second at 720P and 38 credits a second at 1080P, against 10 at 480P.
Two things the bare prices do not show. A regeneration is a new roll of the dice: if it comes out wrong, you pay the higher tier again for the retake, while enhance has no retake risk on the content. And enhance comes on top of what the 480P take already cost; if you know from the start that a segment must be 1080P, generating it on a 1080P tier once can be the shorter road. 480P, 720P or 1080P · Where your credits go in one episode.

## The draft-then-finish workflow, with enhance

Drafting an episode at 480P and finishing the good segments at a higher tier now has two finishing moves:

- Generate every segment at the default 480P tier on SceneMixer.
- Watch, note what is wrong, fix the cause (storyboard, images, state picks), and regenerate those segments at 480P until each one is right.
- For each segment you keep: enhance it if you want this exact take sharper; regenerate it at 720P or 1080P if you would rather have a native take at that tier and can accept a different performance.
- Composite. The episode joins the active version of every segment; segments at different resolutions are fine together, and the full-episode export tops out at 1080P.
Draft at 480P, finish at 1080P covers the arithmetic of the regenerate route.

## Do the fixes first, enhance last

Enhance is the last step for a segment, because two of the other fixes work from the original clip:

- AI visuals and Redub start from the original, so their new version is not enhanced. The tab says so: This version was edited from the original clip, so it isn’t enhanced — you can enhance it again.
- Trim keeps the enhancement: a trimmed enhanced version stays enhanced and is not charged again.
- Regenerating makes a new take; enhance that one if you want it sharper.

## Seeing the difference before you decide

Everything on SceneMixer is kept as a version, so the cheapest test is to try and compare. Enhance once, then switch between Enhanced and the original in Version History with Use — switching is free. Watch the faces in the close-ups if you used AI Detail, and judge on the screen size your audience will use: on a phone, a clean 720P is often all that shows. Segment versions.

| You want | Pick |
| This exact take, sharper | Enhance |
| A face, an action or a line fixed | Fix the storyboard or images, then regenerate |
| 480P to 1080P in one step | Regenerate on a 1080P tier |
| A 2K version of a segment | Enhance |
| A grainy or blurred clip cleaned | Enhance, Cinema Restore |
| Every segment of a new episode at 1080P from the start | Generate on a 1080P tier |

## FAQ

### Is a 720P regeneration better than an enhanced 480P clip?

It is a different thing. A regeneration renders a new take natively at 720P; an enhanced clip is your existing take raised one step. If the take matters more than native detail, enhance; if you need the native tier and can accept a new performance, regenerate.

### Does enhance change the lip movement?

SceneMixer does not touch the sound, and the motion is the same take. Instant HD redraws nothing; AI Detail redraws fine detail, so mouths and faces can shift slightly. Regeneration performs the lines again from scratch.

### Which is cheaper?

Compare the two numbers on SceneMixer: the selected card in the Enhance tab, and the Generate Video button with the higher tier picked. Remember that a regeneration can need a retake at the higher price, and enhance comes on top of what the first take cost.

### Can I enhance a clip twice to go from 480P to 1080P?

No. SceneMixer’s Enhance goes one step from the original resolution, and enhancing again replaces the previous enhanced version. For 1080P from 480P, regenerate on a 1080P tier, or regenerate at 720P and enhance that.

### Can I try both and keep the better one?

Yes. Both results are versions of the segment. Switch between them in Version History with Use; switching on SceneMixer is free, and the active one is what plays, downloads and goes into the composite.

### Do I lose the enhancement if I edit the clip afterwards?

Trim keeps it. SceneMixer’s AI visuals and Redub work from the original, so their result is not enhanced; enhance again if you want it sharper.

### Can a whole episode be 2K?

Single segments can be enhanced to 2K and downloaded at 2K, but the full-episode export tops out at 1080P.

### Does a higher tier fix a face that keeps changing?

No, and neither does enhance. A drifting face on SceneMixer comes from the storyboard or the character images; fix it there and regenerate.

Updated November 3, 2026


---

# Etsy Product Video Size: Length, File and Format, Checked Nov 2026

URL: https://scenemixer.com/guides/etsy-product-video-size/
Etsy's own help page lists the listing-video length, file size, resolution and no-audio rule. What it says today, and how to make a 15-second AI product clip that fits.

# Etsy product video size

Quick answer: Etsy's help page for listing videos (read November 2, 2026) says: up to two videos per listing, each 3 to 15 seconds, up to 100 MB, with no audio once uploaded; the minimum resolution is 500 pixels and Etsy suggests at least 1080. The same page words the aspect ratio as 2:1 or 1:2, which is neither of SceneMixer's shapes, so test one upload. The clip that fits best is one 15-second segment taken straight from the editor rather than the finished ad, because the ad adds an end card and runs a few seconds over 15, and because the sound will be stripped anyway.

One line for the box next to the upload: A hand-poured soy candle in an amber glass jar with a wooden lid; for people who like a quiet evening; shown lit and unlit. Tone: Show how it works; 15 seconds; Landscape or Portrait.

## What Etsy's help page says (read November 2, 2026)

| Item | What the page says |
| Number | Up to 2 videos per listing |
| Length | 3 to 15 seconds |
| File size | Maximum 100 MB |
| Formats | MP4 and MOV among others; the page lists several more |
| Audio | None once uploaded to your listing |
| Resolution | Minimum 500 pixels; the ideal is at least 1080 |
| Aspect ratio | Worded as 2:1 or 1:2 |
Source: How to Add Listing Videos. Etsy's Seller Handbook (a November 5, 2024 article) adds that videos longer than 15 seconds are clipped, so trim the file yourself rather than let the upload cut the end. 3 Types of Listing Videos That Online Shoppers Love.

These numbers change. Read the help page before you export, and look at the preview in Etsy's uploader. This page covers size and format only; what a listing may show is set by Etsy's own policies.

## Why a 15-second ad is not quite a listing video

A SceneMixer ad project makes one, two or four 15-second segments and then joins them with an end card of a few seconds, so even the shortest finished ad runs a little over Etsy's 15 seconds. The end card carries a brand, a tagline and Shop now, which is an ad's closing line, not a product demonstration. And because Etsy strips the audio, the two to four spoken lines the ad writes, and its score, are lost.

So take the clip, not the cut. Each 15-second segment is its own clip in the editor: open the finished ad, select the segment, press Download under the player, and you have a 15-second file with no end card. If it is longer than you want, Edit this clip then Trim cuts the start or the end without regenerating (the price shows on the Save button). The ad progress page · Trim a clip · Download the original.

## Make a clip that works without sound

With no audio, the picture has to explain itself. Ask for it in the one line and in the tone:

- Show how it works for anything that opens, pours, folds or lights; Mouth-watering close-ups for food and drink; Minimal for a design object.
- Say what the buyer wants to see: the scale (fits in one hand), the lid coming off, the item in use.
- Ignore the dialogue. It will not be heard; do not rely on a spoken line to carry a fact, put it in the listing text.
- Mind the corner label. On a free SceneMixer account every downloaded video carries a small AI label in the corner; a paid membership turns it off by default. The AI label switch.
The product is redrawn from your photo as a clean studio shot, with its shape, colour, material and labels kept, and used in every scene; people holding it are generated. An ad for a physical product · Which photos work.

## Which tone for which Etsy category

Etsy's handbook gives category tips for listing videos (end on the finished item, show scale and materials, show wear on vintage). Here is how they line up with the ad tones:

| Etsy category | What Etsy suggests, in short | Tone and photos |
| Home and living | Show how it works, scale and materials; end on the finished item | Show how it works; first photo the item on a plain surface, extra photos in use |
| Bath and beauty | Ingredients, the item in use, end on the finished product | Show how it works or Mouth-watering close-ups; a texture photo as an extra |
| Jewellery and accessories | On a model, for movement and scale | Minimal or Warm; the piece alone first, a worn photo second |
| Clothing and shoes | On-model movement, a close-up of the fabric | A day in the life; the garment flat or on a hanger first |
| Craft supplies | Close-ups of materials, supplies used in a project | Show how it works; a texture close-up as an extra photo |
| Toys and games | The item in use, the pieces that come with it | Funny or Show how it works |
Jewelry ads · Clothing ads · Skincare and beauty ads.

## Where an AI clip is the wrong tool

An AI render shows a redrawn product and a generated person; it cannot show the real wear on a vintage piece, the true size of a handmade object next to your hand, or the actual texture of one-of-a-kind work. Etsy's handbook says a vintage listing video should show signs of wear, and a render will not. For those items film the real thing; use an AI clip as a second video for scale and styling. Etsy allows two.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Etsy limit (help page, November 2, 2026) | SceneMixer clip | Fits? |
| 3 to 15 seconds | One 15-second segment, trimmed if you like | Yes, without the end card |
| Up to 100 MB | A 15-second clip is far smaller than that | Check the file size on disk |
| No audio kept | Ignore the spoken line and the score | Yes |
| Minimum 500 pixels, ideal 1080 or more | Render at the 720P or 1080P tier | Yes |
| Aspect ratio worded as 2:1 or 1:2 | 16:9 or 9:16 | Test one upload and watch the preview |

## FAQ

### What size should an Etsy product video be?

Etsy's help page says 3 to 15 seconds, up to 100 MB, minimum 500 pixels with at least 1080 suggested, and no audio once uploaded. Read the page again before you export.

### How long can an Etsy listing video be?

Etsy's help page says 3 to 15 seconds, and its Seller Handbook says longer videos are clipped.

### Does an Etsy video need sound?

No. The help page says the video will not contain any audio once uploaded to your listing.

### How many videos can I add to a listing?

Up to two per listing, according to the help page.

### What aspect ratio does Etsy want?

The help page words it as 2:1 or 1:2. SceneMixer makes 16:9 and 9:16, so test one upload and check how Etsy frames it.

### Can I make an Etsy product video with AI?

Yes: upload one to four photos of the item to SceneMixer, pick Show how it works and 15 seconds, then download one segment, which has no end card.

### Why not use the finished ad?

The finished SceneMixer ad ends on an end card of a few seconds, so it runs a little over 15 seconds, and Etsy removes the sound anyway.

### Will the product in the video match my photo?

SceneMixer redraws the product from your photo as a studio shot with its shape, colour, material and labels kept; check it against the real item before you post.

### What does it cost?

An ad is quoted as one figure before it starts and that figure is a ceiling; length and quality set it. A 15-second ad on SceneMixer's default tier is the cheapest length.

Updated November 3, 2026


---

# Every Error Message on SceneMixer and What to Do | SceneMixer

URL: https://scenemixer.com/guides/every-error-message-and-what-to-do/
The messages you can meet while writing, generating, saving and paying: what each means and the one thing to do next. None loses your work or credits without saying so.

# Every Error Message, and What to Do About It

Quick answer: Messages on SceneMixer are written to tell you the next step. Most fall into a few families: something is still running (wait), something is missing (add it), a limit was reached (wait or trim), a save was refused (your draft is kept), or a payment did not go through (start again). Find the sentence you saw in the table below. None of these lose your text or your credits without saying so in the message.

If a message is not in this list, send feedback from the sidebar with a screenshot and we will add it. How to get help.

## Something is still running

This episode’s storyboard is still generating — try again once it finishes. This segment’s video is still generating — wait for it to finish. This shot is already generating. Refresh to see its progress. This segment’s edit is still syncing — try generating again in a moment. This episode is still being prepared — try again in a moment. Some segments in this episode are still rendering. Wait for them (or stop those tasks) before regenerating the storyboard. All of these mean: wait, or stop the running job first. Statuses.

## Something is missing

This shot has no script yet — write it before generating video. Generate the base character image first. Describe how this state differs from the main image first. Give this state a name. Fill in the voice description first. Enter what you want changed. Each names the thing to add.

## A limit was reached

Not enough credits — please top up and try again, usually as the Insufficient Credits dialog. Concurrency limit reached — wait for running tasks to finish. Up to 50 at a time — please generate in smaller batches. This segment exceeds the duration cap of the selected model — pick a model that supports longer clips. The storyboard script is too long — trim it before saving. Storage is full / Project limit reached. Too many requests. Please slow down. Content moderation was triggered repeatedly … paused for 1 hour. Wait, trim, delete, or top up, as the sentence says.

## The script you pasted

This script is over the single-parse limit — split it into volumes and upload them separately. That is too short to be a story. Paste the full script or novel text. We can’t open links. Paste the script or novel text itself. Uploading a script.

## A save was refused

Save failed — your draft is kept, please try again. The header could not be reattached safely, so the save was cancelled to avoid corrupting your script. This segment’s header doesn’t match the one being edited — the save was cancelled. Close the editor and reopen it. In every case your draft is still on screen: copy it, close, reopen, paste. Editing the storyboard.

## Names and assets

This project already has a character, scene or item with that name — choose another. This asset already has a state with that name. You’ve reached the maximum number of states for this asset. That file is too large. Please upload a smaller one. The upload did not complete. Please upload the file again.

## Signing in

The verification code is wrong or expired. Request a new one. Too many codes requested. Please try again later. Too soon — wait a few seconds and try again. Your session has expired. Please sign in again. That sign-in link is invalid or has expired. Signing in.

## Paying

This order is no longer payable. Please start a new one. Your previous payment is still being processed. Please check the order status in a moment. A membership payment is already in progress. Your subscription renews automatically — no need to buy it again. Downgrading isn’t supported; you can subscribe to a lower tier after this one ends. Your membership renews through PayPal / through your card. Please use that for plan changes. Payment received, but the amount doesn’t match the order. Credits weren’t added. Please contact support with your order number. When a payment does not go through.

## A generation failed

These appear on a segment card, a sheet or the step-1 reading when a job ends without a result. Generation failed. Please retry — contact support if it keeps failing. — try it again; if the same segment fails a second time, send feedback with the segment number. The generated soundtrack was rejected by the provider’s content check. You weren’t charged — retrying usually passes; if it keeps failing, change the dialogue. The generated image was rejected by the provider’s content check. You weren’t charged — change the description or the visual style and retry. The script couldn’t be read into characters and episodes (was the file plain text?). These credits are on hold pending a support review — no need to resubmit. The generation service hit a temporary problem. You weren’t charged — feel free to try again. When a generation fails · Blocked by content review.

## The script file

That file isn’t text (it looks like a PDF, image or archive). Upload the script as TXT, DOCX or Markdown. Only TXT, DOCX and Markdown files are supported — please pick another file. Nothing is charged; save the script in one of those formats and upload again. Which script files work.

## Before a video is generated

Some shots in segment N have no linked characters. These names are only plain text there… Generate anyway? — a name in those shots was typed rather than inserted with @, so its sheet will not be used there. Cancel, edit the segment, insert the name with @, save and generate; or generate anyway if the person is not on screen. Shots with no linked characters.

Your edit could not be synced to the production copy. The previous version is still used for generation. — shown above the storyboard when you edit in a language other than the project’s. Click Restore machine translation to go back to the text that will be generated, then make the edit again; if it fails twice, send feedback.

This segment is still being prepared. Try again in a few seconds — no credits were charged. — the system is still working out the story state for the segment; wait a moment and press Generate again. We couldn’t prepare this segment yet and have been notified. Please try again later — no credits were charged. — try again later. Building the world.

These state images haven’t been generated yet: …. Generate them in the asset’s states first — no credits were charged. — the segment uses a look (a state) whose image does not exist yet. Open the character, scene or prop, generate the named states, then generate the segment. Notes under a segment · States.

Shot N of this segment shows another place or time — a segment stays in one place and one timeline. Rewrite it to happen here and now, or move it to its own segment. No credits were charged. A segment stays in one place and one timeline.

This segment may not have been read correctly by the system, so the video may come out wrong. Please check it first. — shown for a segment you added or edited. Fix what the box lists, or confirm to generate anyway. When a segment may not have been read correctly.

This project has reached its limit for characters, scenes or items — delete unused ones before adding more. — delete what you do not use, then add. Delete a character, scene or prop.

## Generic

Something went wrong. Please try again. Not found — it may have been deleted. Try refreshing. The service is temporarily unavailable. Please try again shortly. Refresh, retry; if it persists, send feedback with a screenshot.

| You saw | It means | Do |
| …still generating / still rendering / still syncing | A job is running | Wait, or stop it first |
| …no script yet / …first | Something is missing | Add what it names |
| Not enough credits / Insufficient Credits | Balance below the price | Top up, then retry |
| Concurrency limit reached | Too many of your jobs at once | Wait for one to finish |
| …exceeds the duration cap | Segment longer than the tier allows | Trim, split, or change tier |
| Save failed / header could not be reattached / header doesn’t match | Save refused; draft kept | Copy, close, reopen, paste |
| …already has … with that name | Duplicate name | Choose another name |
| Over the single-parse limit / too short / can’t open links | The pasted text itself | Split, paste the full text, or paste text instead of a link |
| Code is wrong or expired / Too many codes | Sign-in code problem | Request a new code; wait if rate-limited |
| This order is no longer payable | Order expired or cancelled | Start a new order |
| Generation failed / …rejected by the provider’s content check | The job ended without a result | Retry; change the dialogue or description if it repeats |
| That file isn’t text | Wrong script file type | Save as TXT, DOCX or Markdown |
| Some shots in segment N have no linked characters | A name was typed, not inserted with @ | Insert it with @, or generate anyway |
| Something went wrong / Not found | Transient or deleted | Refresh and retry |

## FAQ

### Did I lose my text?

No. A refused save keeps your draft on screen. Copy it before closing the editor.

### Did I lose credits?

Not from any message on this page. SceneMixer moves credits only when a generation is delivered, or as a failure message states.

### What if my message is not listed?

Send feedback with a screenshot. The reply arrives in the bell, and by email if you chose that.

### Why do I keep seeing 'concurrency limit reached'?

You have as many jobs running as your SceneMixer plan allows. Wait for one to finish, or upgrade for a higher limit.

### The message is in a language I did not expect.

Switch the interface language in Settings; messages follow it.

Updated November 3, 2026


---

# Film Treatment Template and Example, With a Filled Sample

URL: https://scenemixer.com/guides/film-treatment-template/
A film treatment template with every section, a filled example, how it differs from a logline and outline, and how to paste it into SceneMixer for a cast and outline.

# Film treatment template and example

Quick answer: A film treatment tells the story of a film in plain prose, in the present tense, from the first image to the last, with little or no dialogue. It runs from one page to about ten and sits between a logline and a script. Use the template below: title and logline, tone, a sentence per main character, then the story in three movements. SceneMixer does not write a treatment document, but it reads one: paste it as the script (prose works) and it finds the cast, the places and an outline you can render from.

Copy the template, fill the first two sections in ten minutes, then paste the finished treatment into a new project and compare the cast it finds with your own list.

## Treatment, logline, outline, script

| Document | Typical length | Written as | Use it to |
| Logline | One sentence | Who wants what, and what stands in the way | Hook a reader |
| Synopsis | A paragraph | A short summary of the whole story | Introduce the project |
| Treatment | One to ten pages | Present-tense prose, start to finish | Tell the whole story before writing the script |
| Outline | A list | Scene-by-scene or beat-by-beat | Plan the structure |
| Script | Many pages | Scenes, action lines and dialogue | Make the film |
Lengths are conventions, not rules. Logline examples · Screenplay template · Story bible template.

## The template

| Section | Write | Keep it to |
| Title and format | THE TITLE, then format and length (a short, about 8 minutes) | One line |
| Logline | Who wants what, and what stands in the way | One sentence |
| Tone and look | Three adjectives and what the pictures feel like | Two lines |
| Characters | Name, age, role and one want each | One sentence per person |
| Setting | Where and when, and how many places | Two lines |
| Setup | The world, the person, the first disturbance | A short paragraph |
| Confrontation | The want is tested; things get worse | One to two paragraphs |
| Resolution | The choice, the cost and the last image | A short paragraph |

## A filled example

THE LAST FERRY — a short film, about 8 minutes.

Logline. On the last night before the island ferry line closes for good, a retiring captain agrees to carry one final passenger, a boy with a locked suitcase and no ticket, and must decide whether to open it.

Tone and look. Quiet, wintry, a little uncanny. Cold blues and one warm cabin light; slow shots, no music until the end.

Characters. MARTA VOSS, 61, ferry captain for thirty years, who has refused to be anyone's keeper. OTTO, 24, her deckhand, practical and anxious. ELI, about 10, a boy with a locked suitcase and a story that does not quite add up.

Setting. The Kessel Island ferry, one night in late November: a pier, a wheelhouse, a deck in fog.

Setup. The town has voted to replace the ferry with a bridge, and tonight is the last crossing. Marta counts the last ropes. One passenger waits on the pier: Eli, with a suitcase, no ticket and no parents. He says someone is waiting for him on the island. Otto wants to call the harbour office. Marta casts off instead.

Confrontation. Mid-crossing, fog closes in and the radio dies. The suitcase begins to tick. Otto wants it thrown overboard. Marta tests the boy's story by asking the name of the island's lighthouse; Eli gets it wrong, then right. She opens the case: it is full of small clocks, each stopped at a different time. Eli says they belong to people who never made the last boat, and that he has carried them across for years. Otto retreats to the wheelhouse. Marta realises the boy has stood at the rail on every crossing she remembers.

Resolution. At the island pier the clocks start, one by one, as the fog lifts. Eli steps off and does not look back. Marta looks at her own watch, stopped at the minute she first took the wheel, and winds it. Final image: the ferry's lights go out, the pier lights come on and the new bridge glows in the distance.

## What to leave out

- Pages of dialogue. One line that sells a scene is enough; the script carries the rest.
- Camera directions. Say what happens, not how it is shot.
- The past tense. A treatment is told as it happens: she casts off, the fog closes in.
- Withholding the ending. A treatment gives the ending, because it is a plan, not a teaser.
- Backstory the film never shows. If it will not be on screen, it does not belong here.

## From the treatment to a first episode

SceneMixer does not output a treatment document, and it does not read one differently from a script. Paste or upload yours as the script: .txt, .docx or .md, at least about a paragraph, and up to 300,000 characters of English per project. The read costs 4 credits per 1,000 characters.

| In the treatment | In SceneMixer |
| Characters | A character card each, with a written look and a voice |
| Setting | Locations drawn once and reused |
| Setup, Confrontation, Resolution | An outline of episodes or sequences you can read before anything is drawn |
| Tone and look | Not read from the text: you choose the Visual Style, Scene Style and Costume Style in Step 1 |
| A line you want spoken | Write it as a quoted line with the speaker so it is picked up |
Upload a script and read it · Check the summary · Writing dialogue so it is picked up · Episode outline generator · AI script breakdown · Story to movie.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Nordic Minimal, Frozen Wilderness; Costume Style Outdoor Gear; Visual Style Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw Marta, Otto and Eli and the ferry as the one main location; the pier and the wheelhouse are two more. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Treatment length | Typical use |
| One page | A pitch email, a festival form or a first read |
| Three to five pages | A producer's read; enough to show the shape of the whole film |
| Six to ten pages | A detailed plan before the script, with key scenes |

## FAQ

### What is a film treatment?

A prose telling of the whole film in the present tense, from the first image to the last, with little or no dialogue.

### How long is a treatment?

From one page to about ten. Short films usually need one to three.

### How is a treatment different from an outline?

A treatment is prose that reads like the story; an outline is a list of scenes or beats.

### Does a treatment include the ending?

Yes. It is a plan, not a teaser.

### Can SceneMixer write a treatment?

No. It does not produce a treatment document. It reads yours as a script and finds the cast, places and an outline.

### Can I paste a treatment into SceneMixer?

Yes. Paste it or upload a .txt, .docx or .md file. Prose works as a script.

### Do I need dialogue in the treatment?

No, but any line you want spoken in the video is best written as a quoted line with the speaker's name.

### How much does the read cost?

Reading on SceneMixer costs 4 credits per 1,000 characters, English counted lighter; new accounts start with 50 credits.

Updated November 3, 2026


---

# First Purchase Bonus: How It Works, When It Expires, Refunds

URL: https://scenemixer.com/guides/first-purchase-bonus-how-it-works/
New accounts get extra credits on their first purchase and first subscription. Where the bonus shows, how it is counted, why it has a countdown, what a refund does.

# The First-Purchase and First-Subscription Bonus: How It Is Counted and When It Ends

Quick answer: Your first credit purchase and your first subscription on the website each carry a bonus: extra credits on top of what you buy, granted the moment payment lands. The offer runs in stages counted from your sign-up, each with its own deadline, which is why the buy window shows a countdown next to it; after the last stage there is no bonus. It is counted on the whole order — for a yearly plan, on the full year — and arrives at once, even though the plan’s own credits arrive monthly. It is a bonus on a first order only: a second purchase is at the normal rate, and a refund of the order takes the bonus back with it.

Open the buy-credits window on a new account: each package shows its bonus and the time left at that rate.

## Where the bonus shows

In the buy-credits window, next to each package; on the membership page, on each plan. The number and the countdown come from the server at the moment you look, so what you see is what you get if you pay now. When a stage’s countdown ends, the next stage shows with its own deadline, until the last one is over.

## How it is counted

Credit purchase: bonus rate × the credits in the package, granted with the purchase. Subscription: bonus rate × the whole order’s monthly allowances, granted on payment day even though the allowances themselves arrive one per month. Yearly Plan Credits Arrive Monthly.

## Why there is a countdown

The offer runs in stages, each with its own name, rate and deadline, counted from your sign-up (for accounts older than the offer, from the day it started). The window shows the current stage’s deadline; when it passes, the next stage begins. After the last stage there is no bonus. It is once per account: the stage you buy in is the one you get.

## What it does not stack with

The bonus is a first-order thing. A promo code on the same order is separate and does not change it. A second purchase, or a second subscription, has no bonus. Promo Codes.

## Refunds and chargebacks

A full refund or a chargeback of the order reclaims the bonus along with the order’s credits; if the balance no longer covers it, queued tasks that have not started are cancelled first. For a membership, any remaining shortfall is deducted from later membership grants; for a credit pack it is noted on the order. Refund Rules.

| Order | Bonus counted on | Arrives |
| First credit purchase | Credits in the package | With the purchase |
| First monthly subscription | One month's allowance | On payment day |
| First yearly subscription | Twelve months' allowance | On payment day, at once |
| Any later order | No bonus | — |

## FAQ

### The countdown ended; did I lose the bonus?

Not if another stage follows: it begins with its own deadline. After the last stage there is no bonus.

### Do I get the yearly bonus month by month?

No. It is counted on the whole year and granted at once.

### Can I use a promo code on the same order?

Yes. The code is separate and does not change the bonus.

### I was refunded; where did the bonus go?

SceneMixer reclaims it with the order.

### Is the bonus once per person?

It is once per account, counted from that account's sign-up.

Updated November 3, 2026


---

# Four Ways to Start: a Script, an Idea, Photos or a Clip | SceneMixer

URL: https://scenemixer.com/guides/four-ways-to-start/
Which home-page tab to use: Upload Script with text, AI Create with an idea, Ad Video with photos, Remake with a clip. What each takes in, costs first, and lands you in.

# Four Ways to Start a Project

Quick answer: Have a script or a novel? Upload Script: paste it, the first step is the read. Have only an idea? AI Create: type one line, a script is written, then read. Have photos of a product, a dish, a shop or a car? Ad Video: an ad is made end to end. Have a clip you want redone with your own cast? Remake a video: the clip is read into a storyboard. All four end in the same kind of project — cast, sets, props, episodes, editor — and every step after the first is priced the same way.

The home page tour shows where the tabs are.

## Upload Script

You have: text — a screenplay, a novel chapter, a treatment, in TXT, DOCX, Markdown or pasted. First step: the read, priced by length (4 credits per 1,000 characters), shown on the button. You land in: step 1 with the summary, then step 2 with the cast. This is the main road; the other three join it. Upload and parse.

## AI Create

You have: one line — a genre, a hero, a hook — or nothing (Inspire Me). First step: the writing, priced by the length you pick, shown on the button; then the read, priced like an upload. You land in: the finished script in a window with Start Creating, then step 1. Use it when the story is not written yet or you want a first draft to react to. The one-line brief.

## Ad Video

You have: up to a few photos of the thing to advertise. First step: the photos are read for free on SceneMixer; then either All-in (one price, everything done) or Expert mode (pay the first steps, then drive the wizard yourself). You land in: the ad workbench, or the wizard with ad labels. Use it for a 15, 30 or 60-second ad with an end card, not for a story. An ad from one photo.

## Remake a video

You have: an MP4 or MOV you are allowed to use, within the length shown on the upload box. First step: the read, priced by the clip’s minutes. You land in: step 1 to pick styles, with the cast and the storyboard already read from the clip. Use it to reshoot a reference with a new cast, or to turn a rough phone video into a storyboard. Remake.

## Which one

If the words exist, upload them: the read keeps your dialogue. If you have a premise, let AI write and then edit by writing again with a sharper brief. If the goal is selling one thing, the ad tab is shorter than writing a story around a product. If the pictures exist, remake reads them rather than making you describe them.

| Tab | You have | First step costs | Lands in |
| Upload Script | Text | The read, by length | Step 1, summary |
| AI Create | An idea | Writing, by length; then the read | The script window, then step 1 |
| Ad Video | Photos | Reading photos is free; then All-in or Expert | The ad workbench or wizard |
| Remake a video | A clip | The read, by minutes | Step 1, styles |

## FAQ

### I have a novel. Which tab?

Upload Script. Paste a chapter or upload the file; the read splits it into episodes.

### I only have a premise.

AI Create. One line in, a script out, then the read.

### I want to sell a product.

Ad Video. Photos in, a 15 to 60-second ad out with an end card.

### I have a reference clip.

Remake a video, if you have the rights to it. It becomes a storyboard shot with your cast.

### Do they end up in different kinds of projects?

No. All four lead to a project with cast, sets, episodes and the same editor.

Updated November 3, 2026


---

# Free AI Story Generator: What the Free Credits Buy

URL: https://scenemixer.com/guides/free-ai-story-generator/
A story is the cheap part of making a video. What AI writing, reading and the first storyboard cost, how English is counted, and how long a story the free credits cover.

# Free AI Story Generator: What a Story Costs to Read and Write

Quick answer: A free AI story generator is free in the sense that matters here: the story stage is the cheapest part of making a video. A new account starts with 50 credits. Pasting your own story costs nothing until it is read, and the read is 4 credits per 1,000 counted characters, with English counted lightly: 3,333 English characters count as 1,000. Writing with AI is priced by the length you pick and shown on each option. The storyboard of the first episode of your first drama is free. This guide puts numbers on each step, with the site’s own worked example.

## Four places a story can cost credits

| Step | What happens | What it costs | Free part |
| Writing | AI Create writes a script from your brief | Priced by the length you pick; each option in the menu shows its price | None, but the shortest option is the cheapest |
| Pasting | You bring a story you already have | Nothing until it is read | All of it |
| Reading | The story becomes a summary, a cast, places, props and an episode list | 4 credits per 1,000 counted characters | None |
| Storyboard | Each episode becomes numbered shots with seconds | 1 credit per 100 counted characters of the episode | Episode 1 of your first drama |
| Pictures | A picture for each character, place and prop | Priced per picture after the free pool | The first 15 on the account |
Editing the text of the story or of any storyboard is always free in SceneMixer. Only the machine steps are charged, and they are charged by length, so the cheapest way to spend less is to hand over less text, not worse text. Choosing the script length and cost.

## How English is counted, and why it matters here

SceneMixer counts every length in Chinese-equivalent characters, because a thousand Chinese characters carry about as much story as a few thousand letters. For English and other alphabetic text, 3,333 characters count as 1,000. The count is then rounded up to whole thousands, so a story of 3,400 English characters pays for two units, not one and a bit. The counter under the script box on the home page shows the converted number against the cap, so you never have to do the arithmetic by hand. How characters are counted.

The table shows what that means for the reading fee. Word counts are approximate, at about 5.7 characters a word including spaces. The fee is the number of units times the per-1,000 price.

| English story (characters) | About how many words | Counted as | Units | Reading fee |
| 3,333 | 600 | 1,000 | 1 | 1 × 4 credits |
| 8,333 | 1,450 | 2,500 | 3 | 3 × 4 credits |
| 16,665 | 2,900 | 5,000 | 5 | 5 × 4 credits |
| 33,330 | 5,800 | 10,000 | 10 | 10 × 4 credits: the top of the one-pass read |
| 99,990 | 17,500 | 30,000 | 30 | 30 × 4 credits: read in two passes |
One pass handles up to 10,000 counted characters. Longer stories are read in two passes, the cast first and then the episodes, up to 300,000 English characters, and the fee is still one fee for the whole length. Long scripts are read in two passes.

## The site’s own worked example

SceneMixer prices one example throughout: a 30,000-character English story, split into ten episodes of about 3,000 characters each. In the same terms as above:

- Reading it costs 36 credits (9 units at 4 credits).
- Ten storyboards cost 90 credits in all, 1 credit per 100 counted characters. Episode 1 is free on your first drama, so the first one costs nothing.
- The cast is free up to the first 15 pictures.
For comparison, a new account starts with 50 credits. Everything above is the story stage; it is the video that costs, at 10 credits a second on the default tier, plus 1 credit a second to join. A whole 60-second episode on the default tier comes to about 660 credits, around $4.11 at the lowest credit price, which is why testing the story before the video is the cheap order. Where the credits go in one episode.

## Writing the story with AI

If you do not have a story yet, SceneMixer’s AI Create writes one from a brief. The length menu has five options, sized in counted characters, and each shows its price. For English, multiply by about three to get letters.

| Option | Counted characters | In English, about | What it gives |
| Mini | 1,000 | 3,333 characters | One scene, a test |
| Brief | 2,500 | 8,333 | One or two short episodes |
| Short | 5,000 | 16,665 | A handful of episodes; the usual first choice |
| Medium | 10,000 | 33,330 | A short season |
| Long | 20,000 | 66,660 | A full series |
Options too long to read afterwards are hidden rather than sold and then refused, so the top option can disappear in English while it shows in Chinese. The writing is charged when the script is delivered, and a writing that fails is refunded. The read that follows is priced separately by the script’s length and is shown on Start Creating. The brief matters more than the length: writing the one-line brief and AI story writer cover how to ask. The finished story is yours to rewrite for free, by hand or with a one-sentence instruction.

## Keeping the story stage cheap

- Start with Mini or Brief. One scene is enough to see the whole SceneMixer path: reading, cast, storyboard and preview.
- Stop at a unit boundary. The fee goes up in whole thousands of counted characters, so a story that is a little over 3,333 English characters pays for a second unit. Trim the extra paragraph and the unit goes away.
- Cut the backstory before pasting. You pay to read every character, and a camera cannot show a paragraph of history. Keep it as one spoken line.
- Test the opening alone. A story is read once for its whole length, so reading only the first scene to check the tone costs a fraction of reading the book.
- Edit before you regenerate. Fixing the text is free; regenerating a storyboard is charged and discards the finished videos in that episode.
Spend fewer credits on an AI drama has the longer list, and AI story generator shows what a story turns into.

## FAQ

### Is there a free AI story generator?

SceneMixer is free to start: a new account has 50 credits, pasting a story is free until it is read, and the reading and first storyboard are small charges or free. AI writing is priced by length.

### How much does it cost to read a story?

SceneMixer charges 4 credits per 1,000 counted characters, rounded up. English counts 3,333 characters as 1,000, so a 33,330 character story is 10 units.

### How long a story can the free credits read?

SceneMixer’s example, a 30,000 character English story, costs 36 credits to read; a new account starts with 50 credits. Other lengths scale by the table above.

### Is the first storyboard really free?

Yes. The storyboard for episode 1 of the first drama you ever create on SceneMixer is free. Later episodes, regenerations and later projects are charged.

### Does writing a story with AI cost credits?

Yes. SceneMixer’s length menu shows the price of each option, and the charge is made when the script is delivered. A writing that fails is refunded.

### Can I use my own story instead?

Yes. Paste it into SceneMixer or upload a .txt, .docx or .md file. It costs nothing until you press the button that reads it.

### Does editing the story cost anything?

No. Editing the story or any storyboard text is always free. Only reading, writing, pictures and video are charged.

### What is the longest story it can read?

SceneMixer reads up to 300,000 English characters, in two passes. Beyond that the read is refused before you pay.

Updated November 2, 2026


---

# Free AI Text to Video: How Far Free Takes Your Text

URL: https://scenemixer.com/guides/free-ai-text-to-video/
What a piece of text gets you for free: a cast, a first storyboard and one real preview shot. Where video starts to cost, and how to make the free part count.

# Free AI Text to Video: How Far Your Text Gets

Quick answer: Free AI text to video is free up to a point, and the point is the first paid second of video. A new account has signup credits, 15 free setup images, a free storyboard for episode 1 of its first drama and one free preview of up to 5 seconds, so a piece of text can be taken to a cast, a storyboard and one real shot of your own story without paying for video. Finishing the episode is charged by the second. Below is the path rung by rung, which rung is the last free one, and how to make the free rungs count.

## The free ladder, rung by rung

| Rung | What your text becomes | In plain words |
| 1. Sign up | Nothing yet | A new account starts with 50 credits |
| 2. Get the text in | Pasted, uploaded, or written from one sentence on AI Create | Pasting is free until it is read. Writing with AI is priced by the length you pick |
| 3. Read it | A summary, a cast, locations, props and episodes | Paid from your credits: 4 credits per 1,000 characters, English counted lighter |
| 4. Draw the cast | A picture for every character, place and prop | The first 15 images on the account are free |
| 5. Storyboard | Numbered shots with seconds | Episode 1 of your first drama is free; editing the text is always free |
| 6. Preview | One real shot of your own story, up to 5 seconds | Free, once per account, at the lowest tier |
| 7. Video | The episode | Charged from the first second. No free video |

Rungs 1 to 6 are free or paid from the signup credits. Rung 7 is where free text to video ends, because every clip is a call to a model service that charges SceneMixer by the second. What you get for free lists the conditions in full; Free AI video generator explains why there is no free video allowance.

## Make the preview show the right shot

SceneMixer’s preview is one shot of your episode, made from your own character pictures, and it is once per account. It prefers a shot where someone speaks, so you hear your cast as well as see them; if the episode has no such shot, it takes the opening shot of the first segment. So it pays to get ready before you press it:

- Make sure someone speaks. The preview looks for a shot where a character starts talking; with no spoken line anywhere, it falls back to the opening shot.
- Read the cast first. A generic face on a character picture shows up in the preview first. Fix the description, which is free, and regenerate the picture.
- Read the storyboard first. A thin shot makes a thin preview. Edit the text until the shot says who is in frame and what happens.
- Do not spend it on a draft. It is not regenerated, and a new project does not get another one.
The preview is made at the lowest tier, 480P, so judge the acting and the faces, not the sharpness. How the free preview works.

## What quietly uses up the free part

- Starting a different project first. SceneMixer’s free storyboard belongs to the first drama you ever create and is decided the moment that project is created. Start it with the text you actually mean to make.
- Remaking images. The free images are counted by what you generated, not by what exists. Deleting a picture and making it again does not give the allowance back.
- Regenerating the first storyboard. The free one is the first; regenerating it is charged, and it discards any finished video in that episode.
- Ads and remakes. Neither pays a script-read fee, so neither gets the free setup images.
- An empty project. Free images can only be spent on a project whose script has been read and paid for.

## A small text that goes the whole way

The cheapest honest test is a short text: on SceneMixer’s AI Create the shortest length is Mini, about a thousand counted characters (roughly 3,300 in English), one scene. Or paste one page of your own. Reading a page costs a few credits, the cast and the first storyboard are free, and the preview shows one shot. If the shot is what you imagined, you have found out for the price of reading a page. If it is not, edit the text and look again before any video is paid for. Choosing script length and cost.

## Where free ends, and what comes next

On SceneMixer’s default tier video is 10 credits a second, and joining the segments into an episode is 1 credit a second. A 60-second episode is about 660 credits including the join, about $4.11 at the lowest credit price, before retakes. The price is on the button before you press it. To stretch the credits you have: draft at the default tier and finish on a higher one, fix the storyboard text before regenerating, and generate one segment before the rest. Spend fewer credits on an AI drama · AI video cost calculator.

Free accounts’ videos carry a small AI badge in the corner; removing it is a membership benefit. AI video without a watermark.

## FAQ

### Is there a free AI text to video generator?

SceneMixer is free to start. Text can be read into a cast and a storyboard and previewed as one shot of up to 5 seconds for free, but video is charged from the first second.

### Can I make a full video from text for free?

No. Outside the one preview shot, SceneMixer has no free video allowance. Signup credits can be spent on video like any credits.

### How long is the free preview?

SceneMixer’s preview runs up to 5 seconds, once per account, at the lowest tier, from one shot of your own episode.

### What does my text cost to read?

SceneMixer charges 4 credits per 1,000 characters, with English counted lighter. A page is a few credits.

### Which shot does the preview use?

A shot where someone speaks if there is one; otherwise the opening shot of the first segment.

### Do I keep the free images if I do not use them?

Yes. They belong to your SceneMixer account, not one project, so what you do not use on your first drama is there for the next.

### Does editing my text cost anything?

No. Editing the storyboard text is always free; only generating video is charged.

### What does it cost after the free part?

SceneMixer charges 10 credits a second for video on the default tier plus 1 credit a second to join; the price shows on the button before you press it.

Updated November 2, 2026


---

# Free AI Video Creator: Six Checks to Pass Before You Spend on Video

URL: https://scenemixer.com/guides/free-ai-video-creator/
A free AI video creator can make a cast, a storyboard and one short preview. Six pass or fail checks on that free output, so your first paid clip is not wasted.

# Free AI Video Creator: What to Check Before You Spend a Credit

Quick answer: On a free account you can get to everything except the video itself: a short script read (paid, but small enough for the welcome credits), a cast with pictures from the free image pool, a free storyboard for episode 1 and one preview of up to 5 seconds. That free output is more than a taste; it is a test you can run. This page turns it into six pass or fail checks (the story, the cast, the storyboard, the preview, the price and the label), in an order that matters because the preview is once per account, and ends with advice on where the first paid clip should go. Each check is free, and each tells you what to do when it fails, so the first credit you spend on video is spent on something you already trust.

Try it: start a project from a short script, stop at the cast, and before you press anything that costs credits, read the six checks below once. Then run them in order.

## The order, and why it matters

SceneMixer’s early work comes in a fixed order: the read (paid, by length, but a short script costs only a few of the welcome credits), the setup images (a free pool), the storyboard (free for episode 1 of your first drama) and the preview (free). Only the preview is a one-off. It is once per account, not once per project, so a preview wasted on a cast you later change is gone for good. The editor offers it each time you open a read episode until you use it or choose Don’t show this again, and Maybe later closes the card without using it. So run checks 1 to 3 first, then the preview, then the price. The free preview · What you get for free.

## Check 1: does the read understand your story?

Look at: the summary in SceneMixer’s Step 1. Pass if the protagonist is the right person, the era and setting are what you meant and the episode count matches what you expected. A story you meant as three episodes that comes back as ten has been split differently from how you pictured it. If it fails: a name or description is a free edit in Step 2; a wrong type, region, dialogue language or split is best read again in a new project, and the read is priced by length (4 credits per 3,333 English characters), so keep the first script short. Checking the summary.

## Check 2: would you cast these faces?

Look at: every character, location and prop in SceneMixer’s Step 2. Pass if the leads look like who you meant, nothing in the sheets contradicts the script, and the places show their layout. A face or costume you do not like will be in every shot. If it fails: edit the description and regenerate; this is free while the pool lasts. The pool is 15 images for the whole account, shared by characters, scenes and props, and it applies only to projects whose script has been read and paid for. Spend it on the leads and on the rooms that appear most; minor props can wait, because the storyboard falls back to text for anything without a sheet. After the pool, an image is 7 credits on the default tier. Free sheets are per account.

## Check 3: does the storyboard read like a film?

Look at: episode 1’s storyboard, which is free on your first SceneMixer drama. Pass if every visible character is a linked name, each shot has one action, each spoken line has enough seconds to be said, the room and the positions are stated and the dialogue is in the project’s language. If it fails: edit by hand, which is free as often as you like, or ask for an AI edit in plain words at a small fee; text history restores any earlier version. Regenerating this storyboard is charged, so edit before you regenerate. A checklist before you generate · Who is in frame.

## Check 4: does the preview sound and look like your cast?

Look at: the preview, one shot of up to 5 seconds at the lowest tier, made from your own sheets. It prefers a shot where someone speaks, so you hear the voice as well as see the face. Pass if the face matches the sheet, the voice suits the character, the room looks like the scene and nothing is written in the picture. If it fails: the preview is made from the sheets and that shot as they are, so a poor sheet or a thin shot shows up here first. Fix the sheet or the shot text and generate a real segment; the preview itself cannot be repeated. When a shot comes out wrong.

## Check 5: do you know the price of the second you want?

Look at: the price on a segment’s button and your balance in SceneMixer’s top bar. Pass if you can say how many seconds you are about to buy: the balance divided by the per-second price (the default tier is 10 credits a second). A new account starts with 50 credits, and the same sum, 50 divided by 10 credits a second, shows how little video that is by itself. If it fails: there is no free video beyond the preview, so decide between stopping here, which costs nothing and keeps your cast and storyboard, or buying the smallest pack, which starts at $1.49. Credits do not expire, so a small pack can sit until you need it. Draft on the lowest tier even after you buy. Buy credits step by step · Do credits expire?.

## Check 6: do you know what you will be handing to the world?

Look at: the terms on output and the badge setting. Pass if you know that a free SceneMixer account’s downloads carry a small badge in the corner, that the output from free credits may be used commercially under the terms to the extent the law allows, that it is not exclusive to you, and that the platform you post on has its own AI declaration to tick. If it fails: read the two short guides below before you publish anything. What you get to keep · The AI badge switch.

## Where the first paid clip should go

Not on the easiest shot. The preview already tested one speaking shot, so use the first paid segment on the hardest one: the segment with two people in a room, or a costume state, or the line you are most worried about. It is the shot most likely to expose a problem with the cast or the text, and a miss on the lowest tier is cheap and its lesson reaches every later shot. If it comes out right, generate the rest on the same tier; if not, change the text and regenerate only that segment. Edit one shot and regenerate. For the text-first route see free AI text to video, and for what the free tier is made of, the free AI video generator page.

| Check | Free? | Pass if | If it fails |
| 1 Story | Edits yes; a new read is charged by length | Right lead, era, episode count | Edit in Step 2 or read again in a new project |
| 2 Cast | 15 images per account | Faces and rooms as meant | Edit the description and regenerate |
| 3 Storyboard | Episode 1 of your first drama; edits free | One action a shot, linked names, enough seconds | Edit by hand or AI edit; restore from history |
| 4 Preview | Once, up to 5 seconds | Face, voice, room right | Fix sheet or text, then a real segment |
| 5 Price | Reading prices is free | You know the seconds you are buying | Stop with everything kept, or buy a small pack |
| 6 Label and rights | Free | You know the badge, rights and platform setting | Read the two guides |

## FAQ

### What can a free AI video creator make for free?

On SceneMixer: 50 credits on sign-up, 15 setup images per account, the first storyboard of your first drama, one batch of cover posters and one preview of up to 5 seconds. Video is paid from the first second.

### Can I make a whole video for free?

No. SceneMixer’s only free video is one short preview. The free part is the cast images, the first storyboard and text edits at no charge, which let you check a project before you pay for video.

### Why use the preview last?

It is once per account. Use it after the cast and the storyboard are right, so it tests the version you will actually shoot.

### What if I stop after the free steps?

Nothing is lost. The cast and the storyboard stay in your SceneMixer account, and credits and plan credits never expire.

### What should my first paid clip be?

The hardest shot on the lowest tier: two people in a room, a costume change or a key line. A miss there is cheap and teaches you the most.

### Do I need a card to start?

No. SceneMixer sign-up uses Google, Apple or an email code. A card or PayPal is only asked for when you buy.

### Will my free videos have a watermark?

SceneMixer puts a small badge in the bottom-right corner on downloads, with a metadata line in the file. Paid plans are badge-free by default, and shared links always carry it.

### Is it free for commercial use?

Output made with free credits is treated like any other under SceneMixer’s terms, commercial use included to the extent the law allows. It is not exclusive to you.

Updated November 2, 2026


---

# Free AI Video Editor: Which Edits Cost Nothing and Which Do

URL: https://scenemixer.com/guides/free-ai-video-editor/
An itemised list of which edits to a generated video are free and which cost credits: text, takes and history against trims, redubs, picture changes and exports.

# Free AI Video Editor: The Free Edits and the Paid Ones

Quick answer: Editing a generated video here is half free. Anything that only changes words or chooses between takes costs nothing: editing the storyboard text by hand, switching to an earlier take, restoring older text, inserting or deleting a segment, downloading and sharing. Anything that makes new picture or sound is charged: a trim is 5 credits, an AI revision of the text 1 credit, lip sync 1 credit a second, an AI picture change 2 times the per-second price, and regenerating a segment its tier’s per-second price. This guide puts every edit on one table, works a three-edit example and says what to do to stay in the free column.

## Every edit, free ones first

| Edit | Cost | What it touches | Good to know |
| Edit storyboard text by hand | Free | One segment’s text | No model is called. The video you already have does not change until you regenerate the segment |
| Switch to an earlier take | Free | Which take plays, downloads, shares and joins | Every generation is kept; takes that are not active are cleaned up after 90 days |
| Restore earlier text (History) | Free | Text only | The original is always there, plus the text from before your last five changes; a restore can itself be undone |
| Insert or delete a segment | Free | The list of segments | A new segment starts empty. A segment whose video is still generating cannot be deleted |
| Subtitles, Realign to new lines, Download SRT | Free | Shown over the clip, burned into the download or saved as a text file | Only on accounts that have bought credits or a plan at least once. A daily limit per account applies |
| Download, share | Free | The active take | A free account’s video carries the small AI badge in the corner |
| AI edit of the text | 1 credit per revision | Text only | Charged only when a revision comes back and is shown to you. Failures, refusals and unusable replies cost nothing |
| Trim | 5 credits per trim, whatever the length | Cuts the start, the end or a moment from a finished clip | Nothing new is generated, so the picture you approved is the picture you keep |
| Redub with Lip sync | 1 credit per second of clip | Voices every line again; mouths follow | The picture is kept; the clip’s own ambience is not |
| AI visuals, or Redub with Re-perform | 2 times the tier’s per-second price, per second of clip | Re-makes the clip from itself with one change | Re-perform is offered for clips of 15 seconds or less |
| Enhance | Per second, by option and resolution | Same take, one resolution step up | The price is on each card before you press |
| Regenerate a segment | Tier price per second of the segment | A fresh take | The old takes stay |
| Export episode | 1 credit a second of the finished length | Joins the segments into one file | A trimmed segment is shorter, so it costs less to join |
SceneMixer holds the credits for every paid edit when you press the button, and they come back if the edit fails. None of them has a free allowance of its own, so what you can do for free is the first six rows, plus whatever the welcome credits (50 on a new account) pay for. What each clip fix costs.

## Why the line falls where it does

SceneMixer’s free edits are the ones that do not need a model. Changing text is re-read on your side and nothing is generated; switching takes only moves a pointer; restoring text puts older text back. The paid edits all start a media process or a model call: a trim re-cuts the file exactly to the frame, a redub generates new audio, an AI picture change makes the clip again, an export joins files. That is also why the free column is the right place to do your thinking. Decide what you want in the text, where it is free to be wrong, and spend credits only when you press the button that makes something.

The one trap in the free column is the difference between changing the text and changing the video. Editing a line in the storyboard changes the plan, not the clip, so the segment keeps playing the old words until you regenerate it. Fixing one shot without redoing everything · AI edit or edit by hand.

## A worked example: three problems in a 45-second episode

Take an episode of three 15-second segments on SceneMixer’s default tier, and suppose three things need fixing. This is a different job from fixing a single clip, because here the choices interact.

| Problem | Cheapest fix that works | What it costs | What it keeps |
| Segment 2: a character says “Tuesday” and the plot needs “Friday” | Redub with Lip sync | 15 seconds at 1 credit a second | Every frame. All the segment’s lines are voiced again, so re-read the others |
| Segment 2 alternative: the whole delivery is wrong, not just the word | Edit the line by hand (free), then regenerate segment 2 | 15 seconds at 10 credits a second | Nothing; it is a new take, and the old one stays in the takes list |
| Segment 3 starts a beat early | Trim with Keep | 5 credits | The whole picture; the segment is shorter by what you cut |
| Segment 1 should be at night | AI visuals, the time-of-day change | 15 seconds at 20 credits a second on the default tier (2 times a regeneration) | Acting, framing and timing; small details can differ |
| Put the episode together | Export episode | The finished length at 1 credit a second | The trim in segment 3 makes it slightly cheaper |
The pattern in the table is the useful part. A voice-only fix is a small fraction of any picture fix, a trim is a flat fee that does not grow with length, and a picture fix costs 2 times a regeneration of the same length, so of the three it is the dearest. Regenerating is worth it when the plan is wrong, because only a regeneration can change the framing, the camera or where people stand. A fix is worth it when the clip is right apart from one thing and you do not want to re-roll everything else. Lip sync or re-perform · AI visuals: what it can change.

## Staying in the free column

- Collect your notes, then regenerate once. Edit everything wrong in a segment’s text first. One regeneration pays for all of them.
- Draft on the cheapest tier. The default draft tier is 480P and is for finding out whether a shot reads; finish the keepers on 720P or 1080P. Draft at low resolution, finish at high.
- Use takes as a free A/B. Once a segment has two takes, switching between them is free, and the composite joins whichever is active.
- Treat History as undo. Try an AI edit, save it, and restore the text from before it if you change your mind. The credit for a revision you discarded is not returned, but nothing else is lost.
- Do not regenerate the whole storyboard to fix one shot. It rebuilds every segment and discards the finished videos with them, with no refund. What you lose.

## What a free account can and cannot use

On SceneMixer, credits, not plans, gate the clip fixes: trim, redub, AI visuals and enhance are priced in credits and are not on the list of paid-plan features. The paid-plan features are the ones that bring your own material in or change how the output is labelled: uploading your own pictures onto a character, place or prop, a custom visual style, the storage quota for uploads, and removing the AI badge from your videos. Subtitles are a separate gate: they need at least one purchase of credits or a plan on the account. What a paid plan changes.

If you want the same list without the free angle, video editor with AI orders the repairs from cheapest to dearest, and AI video editor explains what the editor is and is not.

## FAQ

### Is there a free AI video editor?

Part of editing is free in SceneMixer: changing storyboard text, switching takes, restoring older text, adding or deleting segments, downloading and sharing. Anything that makes new picture or sound is charged in credits.

### Is trimming a video free?

No. A SceneMixer trim is 5 credits, whatever the length, and nothing new is generated. If it fails, the credits come back.

### What does an AI edit of the text cost?

SceneMixer charges 1 credit for each revision that comes back and is shown to you. Failed, refused or unusable attempts cost nothing.

### Can I undo an edit for free?

Yes. Every SceneMixer video fix is a new take and switching back is free, and the text history lets you restore earlier wording at no cost.

### Do I need a paid plan to use the fixes?

SceneMixer needs no plan for trim, redub, AI visuals or enhance, only credits. Uploading your own pictures, a custom style and badge removal are paid-plan features.

### Are subtitles free?

They cost no credits, but they are switched on only for accounts that have bought credits or a plan at least once.

### What if an edit fails?

SceneMixer holds the credits when you press and returns them if the edit fails. An AI picture change that is blocked by content review is refunded too.

### What is the cheapest way to fix one wrong word?

A SceneMixer lip-sync redub, at 1 credit a second of the clip, keeps the picture. Editing the line by hand and regenerating the segment costs the segment’s full per-second price.

Updated November 2, 2026


---

# Free AI Video Generator From Text: Shape Your Text for the Free Part

URL: https://scenemixer.com/guides/free-ai-video-generator-from-text/
The free part of text to video is a fixed set of allowances. How the shape of your text, its cast, its opening and its length decide whether they are spent well.

# Free AI Video Generator From Text: Making the Free Part Count

Quick answer: The free part of turning text into video is a fixed allowance, not a percentage of your text: 50 sign-up credits, 15 setup images, one storyboard, one 5 second preview. That means the shape of your text matters more than its quality. A text with a small cast, a spoken line early and a length just under a pricing boundary stretches the allowance; a long text with a dozen characters and no dialogue spends it before the preview. This guide lists the fixed quantities, then shows how five shapes of text use them.

## The free part is a list of fixed quantities

| Allowance | Size | Spent by | Gone when |
| Sign-up credits | 50 credits | Reading, AI writing, pictures after the pool, video | The balance reaches zero |
| Setup images | 15 for the whole account | One per character, place or prop picture, on a project whose reading has been paid for | The pool is used up by generations, not by pictures you keep, so remaking a picture spends it again |
| Storyboard | Episode 1 of the first drama you create | That one episode’s storyboard | You create the first project, even if you delete it later |
| Preview | One clip of up to 5 seconds | Pressing the preview button once | The clip is made, even if you dislike it; it is not regenerated |
Notice that none of these SceneMixer allowances grows with the length of your text. A one-page text and a ten-page text have the same fifteen pictures, the same first storyboard and the same single preview. The text itself is the only variable, and it changes three things: how much the reading costs, how many pictures you need and how good the preview shot is. What you get for free gives the conditions on each line.

## Five shapes of text and what each does to the allowance

| Shape of text | Reading | Pictures needed | The risk to the free part |
| One sentence, written out by AI to Mini length | About 1,000 counted characters plus the writing price | A few: the two or three people, one or two places | Little. This is the cheap way to see the whole path |
| One scene of about 3,000 English characters, two characters and one place | 1 unit of 4 credits | 3, leaving most of the pool for retakes and later scenes | Little. This is the ideal first test |
| A scene with eight named people | Same, by length | 8 or more, plus places and props | The pool is half gone before any shot exists |
| A chapter of 30,000 English characters | 36 credits | As many as the chapter names, often more than the pool covers | The pictures use the pool, and every storyboard after the first is charged |
| Narration with no spoken lines | Same, by length | As the cast needs | The preview looks for a shot where someone speaks, and falls back to the opening shot |
The pattern on SceneMixer is that cast size is the thing to watch. Every named character, place and prop is a picture, and the pool of 15 is shared across every project on your account, so an ensemble story eats it quickly. Leads and the rooms that appear in most segments should come first; a minor prop can be generated later, only if a shot needs it. Free sheets are per account.

## Four ways to shape the text

1. Keep the cast small and named once. Introduce a nickname together with the full name (“Elizabeth, called Liz”) so SceneMixer’s reading treats them as one person. Two or three people carry a short story, and each extra person costs a picture and a voice.

2. Put the strongest scene in episode 1. The free storyboard belongs to the first episode of your first drama, and it is decided when that project is created. A text that opens in the middle of the conflict gives the free storyboard the best material; a text that opens on backstory spends it on setting up.

3. Write a spoken line early. The free preview prefers a shot where a character starts talking, so you hear your cast as well as see them. Write dialogue as Name (delivery): words rather than folding it into narration, so it is picked up as a line. Write dialogue so it is picked up.

4. Mind the reading boundary. The reading is billed in whole units of 3,333 English characters. A text a little over a boundary pays for the next unit, so trimming an unneeded paragraph is a real saving, not tidiness. Text below the floor of roughly a paragraph, about 330 English characters, has nothing to split into episodes.

## Where the free part ends

On SceneMixer, free ends at the first second of video. Beyond the preview there is no free video allowance, and a segment costs its tier’s price per second: 10 credits a second on the default tier, with 1 credit a second to join the episode. The price is on the button before you press it. Everything before that, the reading, the cast, the first storyboard and the preview shot, is a small or zero charge, which is why the sensible order is to finish the free steps first and judge the preview before you spend on a segment. The free preview explains how the shot is chosen.

If the preview looks wrong, it is made from the cast pictures and one shot as they stand. Fix the picture description in step 2 or the shot text, both free, and then pay for the real segment. The preview is not regenerated, and a new project does not get another one.

## A text to test with

If you want a first text that uses the allowance well, write about 3,000 English characters: two named people, one place, one object that matters, three scenes at most, and a line of dialogue inside the first few shots. That reads for one unit, needs about four pictures, gets its storyboard free and gives the preview something to say. If you do not want to write it, let the site draft it with a free AI story generator run at Mini or Brief length. For the whole ladder, including what sits between this and a finished episode, see Free AI text to video, and for the generator without the free angle, Free AI video generator.

## FAQ

### Is there a free AI video generator from text?

SceneMixer is free to start. Reading, a cast, a first storyboard and one preview shot of up to 5 seconds cost nothing or very little; video is charged from the first second.

### What text length uses the free part best?

About 3,000 English characters: one SceneMixer reading unit, a small cast, and one scene with a spoken line. Longer texts scale the reading fee and the number of pictures.

### How many characters can I draw for free?

SceneMixer’s 15 free setup images are shared by characters, places and props across the whole account, so the cast size decides how far they go.

### Does a longer text get a longer free video?

No. SceneMixer’s free preview is one clip of up to 5 seconds whatever the text length. There is no free allowance for the full video.

### Which shot does the free preview use?

A shot where someone speaks, if the episode has one; otherwise the opening shot of the first segment.

### What if I do not like the preview?

SceneMixer does not regenerate it and there is only one per account. Fix the picture description or the shot text, both free, and generate the real segment.

### Does the free storyboard apply to every project?

No. SceneMixer ties it to episode 1 of the first drama you create, and it is decided when that project is created. Deleting the project does not give it back.

### Does editing my text cost anything?

No. Editing text is always free. Only the machine steps are charged.

Updated November 2, 2026


---

# Free AI Video Generator, No Sign Up: What Works Without One

URL: https://scenemixer.com/guides/free-ai-video-generator-no-sign-up/
Generating video here needs an account, and this page says why. What you can do signed out, the six reasons generation needs an account, and how sign-up works.

# Free AI Video Generator, No Sign Up: What Works Without an Account

Quick answer: Generating video on SceneMixer needs an account, and a no-sign-up claim should be tested before you trust it. Without one you can still do a good deal: watch finished sample dramas, open their storyboards and character sheets, read the price list and the guides, play a shared clip, and type or paste a script into the create box. Pressing generate, or anything that spends credits, opens the sign-in window with what you typed kept. Sign-up is Google, Apple or a six-digit emailed code: no password and no card, and a new account starts with 50 credits.

## What works without an account

Signed out, SceneMixer is a working showroom: you can judge the output and the price before you give anything.

| Where | Signed out | What you can do |
| Landing page | Open | Watch sample videos, read what the product does, press Start creating |
| Pricing and membership pages | Open | Every unit price, each membership tier and the credit packs |
| Guides and answers | Open | How each step works, what it costs and what each message means |
| Featured Projects (finished sample dramas) | Open, read-only | Read the summary, open every character sheet, read the storyboard, play the clips, go full screen, share a clip |
| A share link someone sent you | Open | Play the clip without logging in |
| The create box | Open to type in | Paste a script, switch between Upload Script, AI Create, Ad Video and Remake; the first button that spends opens the sign-in window and keeps your text |

Samples are the part worth your five minutes. Open one, click a character to read the description that holds the face, open an episode, click a segment and play it against its text. Nothing in a sample can be changed or generated, so browsing costs nothing. Learning from a sample project is a short walk-through.

## What needs an account, and the reason for each

None of the following is a gate for its own sake. Each is something an anonymous visitor could not be given.

Credits need a balance to be held against. When you press generate on SceneMixer, the credits move from available to held, are charged when the clip is delivered and come back if it fails. That cycle needs an account to hold them. Free allowances are counted per account. The 15 setup images, the storyboard of the first drama, the cover batch and the one preview are all per account; without an account there is nothing to count them against, and the terms say sign-up benefits are not for multiple accounts. A result needs somewhere to live. A clip takes minutes to generate and the generation continues if you close the page; the clip, the sheets and the storyboard are delivered to your projects, where you can download them later. Queues are per account. A free account runs 3 generations at once and the rest queue; a queue needs an owner. Content checks are counted per account. More than 12 blocked video attempts in a day pauses video submission for 1 hour, which only works if the count belongs to someone. Share links, uploads and refunds belong to someone. Only a project's owner can create a share link, and a credit history line is how a refund or a support reply finds the task it is about.

## Sign-up in under a minute

There is no password and no registration form. Press Sign In / Register, or any button that needs an account, and the sign-in window opens over the page. Continue with Google or Apple signs you in with one tap. Continue with email asks for your address, sends a six-digit code, and you type it in; a new address gets an account on the spot and an existing one is signed in. By continuing you agree to the terms and the privacy policy. No card is asked for. If the code does not arrive, check spam, wait for the Resend countdown, or use Google. Signing up and logging in has the details.

What you get at once: 50 credits that do not expire, 15 free setup images that belong to the account, the storyboard of your first drama, one free cover batch and one free 5-second preview. Video is charged from the first second, at 10 credits a second on the default tier, with the price on the button first. What SceneMixer gives you for free.

## How to test any “no sign up” claim

Four questions will do. Where does the result go? With no account to hold it, the file has to be handed to you at once or lost; ask how long you have to download it. Who pays for each click? A tool with no account has no balance, so the cost sits in a short clip, a queue or a mark. What is the longest clip, and at what quality? Look for the number. What do they keep about you? A page that does not ask who you are can still log your IP and device; read what it says. If a page answers all four plainly, it is easy to judge. If it answers none, you have your answer. What can you still make with AI video for free? goes further on what free can reasonably include.

## FAQ

### Can I generate a video without signing up?

No. Generating, reading a script and anything else that spends credits needs a SceneMixer account. You can browse samples, pricing and guides, and type in the create box without one.

### Do I need a credit card to sign up?

No. SceneMixer sign-up takes Google, Apple or an emailed code and asks for no card. A card or PayPal is only asked for when you buy credits or a plan.

### Can I use the site without giving an email address?

Yes, with Google or Apple sign-in. The email code route needs an address to send the code to.

### Is there a guest mode?

Not for generating. Signed out on SceneMixer you can open Featured Projects read-only and play shared clips, but nothing can be created, changed or generated.

### Is there a password to remember?

No. Each SceneMixer sign-in is Google, Apple or a fresh six-digit code. You can link Google to an email account later in Settings.

### What happens to the script I typed before signing in?

It stays in the box. The sign-in window opens over the page; sign in and press the button again.

### Does the Ask AI help window work signed out?

No. It is for signed-in accounts; the guides are open to everyone and answer the same questions.

### What do I get for signing up?

SceneMixer gives 50 credits, 15 free setup images, the first storyboard of your first drama, one cover batch and one 5-second preview. Video and script reading are charged.

Updated November 2, 2026


---

# Free Sheets Are Per Account, Not Per Project: How the Count Works

URL: https://scenemixer.com/guides/free-sheets-are-per-account/
The free sheet allowance is one pool for the whole account, spent across projects that paid the parse fee. What counts, what does not, what the first project gets on top.

# Your Free Character, Scene and Prop Sheets Are Per Account, Not Per Project

Quick answer: You get 15 free sheets for the whole account — characters, scenes and props together — and they carry over from project to project until used. They do not reset when you start a second project, and they do not expire. They can only be spent in projects that paid the parse fee (so not in ad projects or remakes). Separately, your first project gets its first episode’s storyboard free. Everything else in Step 2 is paid per sheet.

Open Step 2 of any parsed project: the banner above the grid shows how many free sheets are left on the account right now.

## How the count works

Every generated sheet is one task; the free pool is charged one per successful or in-progress task, whichever type it is. A sheet that fails, or is cancelled before it starts, is not counted. The pool is account-wide: sheets used in project A are no longer available in project B. There is no per-type quota any more and no per-project reset. What you see in Step 2 and how the grid is organised: Reading Step 2.

## Where free sheets can and cannot be spent

Only on the default image model. Since October 24, 2026 SceneMixer draws free sheets on Qwen-Image 3.0, the default in Settings → Model Preferences → Image Model. Pick another image tier and those sheets are charged at that tier’s price without touching the pool; switch back and whatever is left is still there. See which image tier for your sheets.

Only in projects that paid the parse fee: a script you uploaded or wrote with AI Create and then parsed. Ad projects and remakes generate sheets too, but they never draw from the free pool; they are priced per sheet from the start. Regenerating a sheet draws from the pool while any is left, exactly like a first generation. AI edits, regeneration from a reference image and state images never use it: they are priced on their own.

## What does not give the credits back

Deleting a sheet, deleting a character, deleting the whole project, or re-parsing the script: the tasks already ran, so the pool stays where it was. If you want to try a different look for a character, edit the description and regenerate — free while the pool lasts, then a paid sheet at the default tier’s price of 7 credits. Regenerate One Sheet and Write a Description That Works.

## What the first project gets on top

Your first project also gets the storyboard for episode 1 generated free the first time; later episodes, regenerations and every later project pay the storyboard fee. New accounts start with 50 credits as well. The complete list of what is free: What SceneMixer Gives You for Free.

## Making the free pool go further

Spend it on the sheets that matter for the video: the lead characters and the rooms that appear in most segments. Minor props and one-scene locations can be generated later, only if a shot needs them; the storyboard falls back to text description for anything without a sheet. Order and price of Generate All: Generate All in Step 2.

| Action | Draws from the free pool? | Notes |
| Generate a character, scene or prop sheet in a parsed project | Yes, one per sheet | Until the pool is empty; then per-sheet price |
| Sheets in an ad project or a remake | No | Always paid |
| Sheets on an image tier other than the default | No | Charged at that tier’s price; the pool is kept for later |
| Regenerate a sheet | Yes, while the pool lasts | Same as a first generation |
| AI edit, regenerate from a reference, state image | No | Priced on their own |
| Delete and recreate a sheet or project | No refund to the pool | Tasks already ran |
| First episode's storyboard in the first project | Separate allowance | Once |

## FAQ

### Do my free sheets expire when I start a second project?

No. The pool is per account and carries over until it is used up.

### Is it five per type?

Not any more. SceneMixer has one pool of 15 for all types together.

### Why does my remake show no free sheets?

Free sheets only apply to projects that paid the parse fee; remakes and ads never draw from the pool.

### I deleted a project by mistake. Do I get the sheets back?

No; the generations already ran. The pool only counts tasks, not what you keep.

### What is the price after the pool is used?

The default image tier's price per sheet, shown on the button before you generate.

Updated November 3, 2026


---

# Generating a Video Segment: Button, Price, Preview | SceneMixer

URL: https://scenemixer.com/guides/generate-a-segment/
Where the generate button is, what the price means, what the confirmation shows, how long a segment takes, what you can do meanwhile, and the free preview.

# Generating a Segment: What Happens When You Press the Button

Quick answer: Select a segment on the timeline and press Generate Video in the preview pane. The price on the button is that segment’s length times the tier’s per-second rate. A confirmation dialog shows what the shot will use — the character, scene and prop images, the voices — and the credits that will be deducted; press Confirm Generate. The credits are held, the card on the timeline says Generating, and you can keep working or leave. When it is done the clip appears on the right and the credits are charged. Every account also gets one free preview: a short opening shot from your own script, at no cost.

Open a finished episode and click a segment. The preview pane on the right is where the button lives.

## Before you press it

Two things decide whether the shot will come out well: the storyboard text and the images. Read the segment once — is everyone visible named with a chip, do the shots have their seconds, does every line have a voice tag? Editing the storyboard. If a character or scene in the segment has no image yet, the editor warns you that the generation will lack it; generate the image in step 2 first.

## The button and the price

Generate Video sits in the preview pane on the right, with the price on it. The price is the segment’s seconds times the per-second rate of the tier in the picker beside it. Change the tier, or the seconds, and the price changes with them. Choosing a tier.

## The confirmation dialog

Pressing the button opens Generation Preview. It lists the reference images the shot will use — the characters, the scene, the props named in the text — and the reference audio for the voices, so you can see at a glance whether the right people and things are attached. Below is the exact text that will be used, visible only to you. At the bottom: N credits will be deducted and Confirm Generate. Nothing has been charged yet; Cancel closes it.

## While it runs

The credits are held, not yet spent. The timeline card turns Generating and the preview pane says roughly how long it usually takes — typically ten to twenty minutes, depending on the tier and the queue. Queued, starting shortly means it is waiting its turn. You can edit other segments, generate others (there is a limit to how many run at once; paid plans raise it), or close the tab and come back; the clip lands whether or not you are watching. If it says taking longer than usual, still running, it is still running. Why a generation takes time.

A clip cannot be stopped once you press generate; it lands as a new version you can keep or ignore. What can be stopped.

## When it is done

The clip appears in the preview pane with play, download and share controls, and the card shows its thumbnail. Only now are the credits charged. If the result is wrong, edit the text and press Regenerate; the old version is kept in Version History. If it failed, the card says so and the message tells you whether the credits came back. When a generation fails.

## The free preview

SceneMixer gives every account one free preview: about 5 seconds of the opening shot of an episode, generated from your own characters and script, at no cost — no credits held or charged. The offer appears in the editor toolbar as Generate a free preview; it takes a few minutes and shows up in the toolbar when ready. One per account, on whichever project you like.

## Not enough credits

If the segment costs more than you have, an Insufficient Credits dialog shows what the action requires, your balance, and the shortfall, with Buy Credits and Upgrade. Nothing is generated until you have the credits. Buying credits.

| Moment | What you see | Your credits |
| Before pressing | Price on the button | Untouched |
| Confirmation dialog | References, audio, text, credits to deduct | Untouched |
| Generating | Card says Generating; time estimate | Held |
| Done | Clip in the preview pane | Charged |
| Failed | Card says Failed; message explains | Returned, or on hold for review — the message says which |

## FAQ

### When are the credits actually taken?

SceneMixer holds them when you confirm and charges them when the segment is delivered. A failure returns them unless the message says they are on hold for review.

### Can I close the page while it generates?

Yes. The clip lands whether or not you are watching. Open the episode later and it is there.

### How long does one segment take?

On SceneMixer, usually ten to twenty minutes, depending on the tier and how busy the queue is. The preview pane shows the usual time.

### What is the free preview?

One short opening shot per account, from your own script, at no cost. Look for Generate a free preview in the editor toolbar.

### Why does the dialog show images I did not expect?

It lists the entries named in the segment's text. If the wrong ones are there, fix the chips in the storyboard before confirming.

Updated November 3, 2026


---

# Generate AI Video: What Happens After the Button, and What You Pay

URL: https://scenemixer.com/guides/generate-ai-video/
One segment or all of them, the queue, the states a clip passes through and what you are charged when a generation is busy, blocked, stopped or fails.

# Generate AI Video: What Happens After You Press the Button

Quick answer: Generating an AI video is a purchase with a delay. You press once, the price is held, and typically ten to twenty minutes later the credits either become a charge for a clip or come back. In between, a clip can wait in your queue, run, fail, or be blocked, and each of those settles the credits differently. This page follows the button from both sides: what you can press (one segment or Generate all), what the card says at each state, how many run at once on each plan, and a table of what you are charged in nine situations, so that a failed card does not turn into a support ticket.

Try it: open an episode with two or three finished segments and read the status on each timeline card. Then look at the price on a segment with no clip, and at the total on Generate all. The first number is one purchase, the second is a batch of them.

## Two ways to press

One segment. SceneMixer’s Generate button on a segment shows its price. Pressing it holds that amount and starts a clip at the tier selected in the top bar. Generate all. In the editor’s top bar, next to Composite, it appears while some segments have no clip. It shows a total; a dialog lists how many segments will be generated, their seconds in total and the model.

Generate all submits every segment that has storyboard text and no clip, in timeline order, one at a time, each charged at its own price. It leaves alone segments that already have a clip, so it never pays twice for the same segment, and skips those already generating or waiting. Segments longer than the selected tier allows are left out and named. Before submitting anything it checks your balance against the total; if it does not cover the batch, the buy-credits window opens and nothing is submitted. Stop here halts submission; segments already sent carry on. Generate all segments of an episode at once.

## The states a clip passes through

| Card says | What it means | Your credits |
| Pending | Text but no video yet | Untouched |
| Waiting | Pressed, but behind your other jobs | Held |
| Queued, starting shortly | Submitted, about to start | Held |
| Generating | Running; the pane shows the usual time for the tier | Held |
| Taking longer than usual, still running | Past the usual time, not failed | Held |
| Done | The clip is in the preview, a new entry in Version History | Charged |
| Failed | Not delivered; the message says which of three things happened | Returned, or on hold for review |

You can leave the page; SceneMixer runs the job on the server. Only Failed ever needs anything from you. What each status means.

## How many run at once

Each SceneMixer account runs a fixed number of generations at the same time. The rest wait with their credits held and start by themselves, in the order you pressed. The number counts every kind of generation, not just video: segments, sheets, storyboards and joins.

| Plan | Generations at once |
| Free | 3 |
| Pro | 5 |
| Studio | 8 |
| Enterprise | 10 |
| Larger plans | More, up to the top plan; the pricing page lists them |

A plan changes how many run together, never how long one takes. Queued is your own line, and clears by itself. Busy is the provider’s line: the request was refused and its credits returned, and it will not restart by itself. You press Retry. Generating many segments at once.

## What you are charged when it does not go to plan

| Situation | Charged? | What happens |
| It is waiting in your queue | Held, not charged | Starts by itself when a slot frees |
| The request never reached the model (a bad parameter, a reference that could not be fetched) | Refunded | Fix the cause and retry |
| The provider was busy or rate-limited | Refunded if it still fails | The site retries first: text and some image tiers for up to about a quarter of an hour, then once more on another account or after a short pause; video only within its first 30 minutes |
| The clip was blocked by content rules | Refunded | Change the shot. Up to 12 blocks a day per account are refunded automatically; past that they are held for review, and repeated blocks pause video for about 1 hour |
| A timeout, server error or network failure in the middle of a call | Held for review | Neither charged nor returned until a person settles it; do not generate the same segment again meanwhile |
| The clip was made but delivery failed on our side | Held for review | You are not billed for something you did not receive |
| You stop a text job before it was sent to the model | Refunded | After it was sent, the credits are spent |
| You want to stop a clip that is generating | Charged | There is no Stop; it runs to the end and lands as a new version |
| You delete the episode or project while it generates | Mixed | Work not yet started is refunded; clips already sent are settled by what the provider reports; other work already sent is held for review |

SceneMixer charges a task once, on delivery, whichever attempt delivered it, and every outcome is a row in your credit history tagged In progress, Refunded or Under review. When a generation fails · Busy, not charged · Stopping a generation · Refund rules.

## A safe order for a batch

1. Use the free preview if you still have it: one shot of up to 5 seconds at the lowest tier, with your own cast. 2. Generate one segment on the lowest tier and watch it with sound, checking the face, the room and the voice. 3. Generate all on the same tier once that one is right; six segments run blind from a wrong first shot are six regenerations. 4. Fix single shots by editing their text and regenerating only those. 5. Finish the keepers on a higher tier, then join. Draft low, finish high · A checklist before you generate.

While a clip renders you can edit the storyboard text of that segment; the edit is saved and the running generation is unaffected, so the next regeneration uses the new text. A segment that is still generating cannot be deleted, and an episode whose storyboard is being written refuses video until the storyboard lands.

## What generation does not promise

A generation is probabilistic: two identical requests can give two different clips, and the terms do not guarantee any single result. What you are promised is the money side. The price on the button is the amount held, the amount taken comes from the same calculation, and a clip you did not receive is not billed.

| Moment | What you can do | What it costs |
| Before pressing | Edit text, pick a tier, read the price | Nothing |
| After pressing | Wait, leave the page, edit the text for next time | Held until the clip lands |
| Clip lands | Watch, regenerate, fix, or keep | Charged on delivery |
| Clip fails | Read the message and retry or edit | Returned, held for review or returned after a block |

## FAQ

### How long does it take to generate an AI video?

A SceneMixer segment typically takes ten to twenty minutes on every plan. A 60-second episode is at least four segments, so how long the whole episode takes depends on how many run at once.

### What does Generate all do?

SceneMixer’s Generate all submits every segment that has storyboard text and no clip, in timeline order, at the tier in the top bar. It skips segments that already have a clip and checks your balance before sending anything.

### Am I charged if generation fails?

Not for a request the model never received, a busy provider that stayed busy, or a clip blocked by content rules. A timeout or delivery failure is held for review rather than charged or returned.

### Can I stop a video that is generating?

No. A SceneMixer clip runs to the end and lands as a new version of the segment. Text jobs such as the read and the storyboard have a Stop button.

### Why does my clip say Queued?

You started more generations than your plan runs at once. The extras hold their credits and start by themselves as others finish.

### Does a paid plan generate faster?

No. A SceneMixer plan runs more of your generations together. One segment takes the same time on every plan.

### What if I close the page?

SceneMixer continues the job on the server. Reopen the episode and the card shows the current state.

### What should I do when a batch all fails?

Read one card. If every card shows the same message, the cause is shared, such as a busy tier or an unusable reference, and fixing that one thing fixes the batch.

Updated November 2, 2026


---

# Generate All in Step 2: Order, Price and the Free Sheets | SceneMixer

URL: https://scenemixer.com/guides/generate-all-order-price-free-sheets/
What the Generate All button does on each tab, what Generate everything adds, which cards it skips, how images are priced, and how the free allowance is counted.

# Generate All: Order, Price and the Free Sheets

Quick answer: On each tab of step 2, Generate All makes an image for every entry of that tab that has none yet — and queues their state images to follow. It skips entries that already have an image, uploaded ones, and voice-only characters. Generate everything does all three tabs at once and starts the episode-1 storyboard as well. Images are priced per image on the button; the first 15 character, scene and prop images on an account are free, counted across all three tabs.

Open a finished episode and look at the left panel: that is a fully generated cast.

## The two buttons

Generate All Characters (or Scenes, or Items) sits above the grid and applies to the tab you are on. Its tooltip says what it does: generate reference images for every asset of this type that has none yet, together with their state images, which start automatically once the main image is done.

Generate everything is the one-click version: all characters, scenes and props, plus the storyboard for episode 1. When you press it the page confirms: started — all reference images and the episode-1 storyboard are generating.

## What it skips

Entries that already have an image. Entries whose image you uploaded. Characters marked Off-screen (voice only). Entries currently generating. So pressing it twice does not pay twice, and it never overwrites something you chose.

## The order and the pace

Several images generate at once; each card shows Generating and the line above the grid counts them: Characters generating — 12/20 done. State images wait for their main image and then start on their own (Queued N state images). A failed card says Failed — click to retry. You can move to step 3 while they finish.

## What it costs

Images are priced per image, and the price is on the button. SceneMixer gives every account 15 free character, scene and prop images, counted together across the three tabs, usable on any project whose script has been read; the line above the grid shows how many are left. After that each image is 7 credits on the default image tier. State images have their own small free allowance on your first project and are priced the same way afterwards. What is free.

## After it finishes

Look at the cast. A face that is not who you had in mind is cheaper to fix now than after ten videos: open the card, adjust the description, Regenerate. Regenerating one sheet. A character whose clothes change during the story gets a state. States.

## Ads and remakes

An ad made by AI generates its images on its own. A remake’s cast and places come from your video and are generated here the same way, after you have chosen a style in step 1.

| Button | Generates | Skips |
| Generate All (tab) | Every entry of this tab without an image, plus their states | Existing, uploaded, off-screen, generating |
| Generate everything | All three tabs, plus the episode-1 storyboard | The same |
| Generate (one card) | That entry only | — |

## FAQ

### Does Generate All regenerate images I already have?

No. It only fills in entries that have no image. Use Regenerate on a card to redo one.

### How many images are free?

The first 15 character, scene and prop images on your account, counted across the three tabs. The line above the grid shows what is left.

### What does Generate everything add?

All three tabs at once, and it starts the storyboard for episode 1.

### Do I have to wait for it to finish?

No. Move on; the images land in the background. Give the cast images before generating video, though.

### Why was an entry skipped?

It already has an image, was uploaded, is voice-only, or is generating.

Updated November 3, 2026


---

# Generate All Segments of an AI Episode at Once | SceneMixer

URL: https://scenemixer.com/guides/generate-all-segments-at-once/
Generate all submits every segment that has no clip yet, after one dialog with the total price. What it skips, how it queues, and what happens if credits run out.

# Generate All Segments of an Episode at Once

Quick answer: In the episode editor’s top bar, next to Composite, Generate all appears while some segments have no clip. It shows the total price; press it and a dialog lists how many segments will be generated, their total seconds and the model, plus what is left out. Press Generate and the segments are submitted one by one, each charged at its own price tag and refunded if it fails. Segments beyond your concurrency limit wait in line and start by themselves.

Before pressing, pick the model in the top bar: every segment in the batch uses it. On Wan 3.0 · 480P a segment costs 10 credits per second.

## What Generate all does

- Generates every segment with storyboard text and no clip yet, in timeline order, on the model selected in the top bar.
- Leaves alone segments that already have a clip (the dialog says how many), so it never pays twice for the same segment. To redo one, use its own Regenerate.
- Skips segments that are already generating or waiting in line.
- Leaves out segments longer than the selected model allows, and says so: …are longer than this model allows and are left out. Switch models to generate them.

## The dialog

Titled Generate all segments, it shows the count, the total seconds and the model — for example “8 segments · 96 s in total · Wan 3.0 · 480P” — and the total in credits, which is the sum of the price tags on those segments’ own buttons. The note underneath: Each segment is charged on its own and refunded if it fails. Segments beyond your concurrency limit wait in line and start automatically.

## While it submits

After Generate, the dialog counts Submitting 3/8…. Stop here stops submitting the rest; the ones already submitted carry on. When it finishes you see 8 segments submitted; if some could not be submitted, the message gives how many and the first reason only, so look along the timeline for the cards that still have no clip.

On the website, a segment you wrote or edited yourself is checked for shots with no linked characters. When one has them, you are asked about that segment alone, with the same question as a single Generate: confirm and it is submitted; cancel and only that segment is skipped (Segment N was skipped: some shots have no linked characters.) while the rest carry on. Shots with no linked characters.

If a segment’s dialogue is not in the project’s dialogue language, you are asked once; your answer applies to the rest of the batch. Dialogue language check.

## Credits

Before submitting anything, the total is checked against your balance. If it does not cover the total, the buy-credits window opens and nothing is submitted — no half-finished batch to sort out. If the balance runs out part-way (for example because something else was charged meanwhile), submitting stops at that segment and the buy-credits window opens. Each submitted segment holds its own credits and is charged when its clip arrives, refunded if it fails. What one segment costs.

## Queue and time

Your account runs a limited number of generations at once, set by your plan. The rest wait in line with their credits held and start by themselves as slots free up; you can close the page. A long episode takes a while — each clip usually takes several minutes. Generating many segments at once.

## On a phone

Generate all sits in the bottom bar, above Composite, with its price. The same dialog opens.

| Segment | Generate all |
| No clip yet, has storyboard text | Generated |
| Already has a clip | Left as it is |
| Generating or waiting in line | Not submitted again |
| Longer than the model allows | Left out; switch models |
| No storyboard text | Skipped |

## FAQ

### Does it regenerate segments I already have?

No. Only segments without a clip. Use a segment's Regenerate button to redo one.

### Can I use different models for different segments?

Not in one batch; it uses the model selected in the top bar. Generate the odd one out with its own button after switching models.

### What if I close the page?

Submitted segments keep going. Submitting stops if you leave while the dialog is still counting.

### Is it cheaper than one by one?

No, and not more expensive: the total is exactly the sum of the segments' own prices.

Updated October 12, 2026


---

# Generate All With Low Balance: What Happens and What to Do

URL: https://scenemixer.com/guides/generate-all-with-low-balance/
Generate All holds the credits for every sheet it queues the moment you click. What happens when the balance cannot cover the total, and what to do next.

# What Happens When You Press Generate All Without Enough Credits

Quick answer: Generate All holds the credits for every sheet it queues the moment you click. If the total on the button is more than your balance, the buy-credits window opens first and shows how much you are short; nothing is queued and nothing is held. If the balance runs out part-way (for example because another tab spent some credits), the sheets that fit are queued and the rest stay Pending with nothing held. Queued sheets start as your plan’s generation slots free up.

Before pressing, compare the price on the button with the balance in the top bar; the button shows the total for everything it will queue.

## What you will see

With enough credits, every pending or failed card in the tab turns Generating or Queued at once, and the credits for all of them are held immediately. Without enough, the buy-credits window opens instead of the run, with the amount you are short; the cards stay as they were. No card ever ends up Failed because of the balance, and there is nothing waiting for review.

## How many run at once

Your account runs a limited number of generations at the same time, set by your plan (3 on SceneMixer's free plan, 5 on Pro, more on higher plans), shared with anything else you have running. Sheets beyond that wait in your queue with their credits already held and start by themselves. Generating Many at Once.

## What to do next

Top up from the credits panel in the top bar, then press Generate All again. Only cards that are pending or failed are queued; cards with an image are left alone and not charged again. If you would rather not generate everything, click individual cards instead — each has its own button and its own price. How to buy credits.

## Using the free pool first

SceneMixer uses the account’s 15 free sheets before any paid one. If you want the free ones to land on the lead characters, generate those cards individually first and press Generate All afterwards. How the free pool works.

## Reading your credit history afterwards

In the credit history each queued sheet is one line, In progress while it is made and spent once it is delivered. A sheet that failed for another reason shows Refunded, or Under review while it waits for a check. Sheets that were never queued have no line at all. Reading Your Credit History.

| What happens | Credits | Next step |
| Balance covers the total | Held for every queued sheet at once | Wait; queued ones start by themselves |
| Balance short of the total | Nothing held; the buy-credits window opens | Top up, press Generate All again |
| Balance ran out part-way | Held for the queued ones only | The rest stay Pending; press again after topping up |
| A sheet failed for another reason | Refunded or held for review | Read the card; see the failure guide |

## FAQ

### Did I lose credits on the sheets that did not start?

No. Sheets that were not queued hold nothing and have no line in your history.

### If I press Generate All again, will finished sheets be regenerated?

No. Only pending and failed cards are queued.

### Why are some cards Queued rather than Generating?

Your plan runs a limited number of generations at once; the rest wait with their credits held.

### Can I generate just one card?

Yes. Each card has its own button and price.

### Does the same happen with video segments?

Yes: a segment you cannot afford is not submitted; the buy-credits window opens instead.

Updated November 3, 2026


---

# Generate the Storyboard for an Episode (First One Free) | SceneMixer

URL: https://scenemixer.com/guides/generate-the-storyboard-for-an-episode/
Step 3 lists your episodes as cards. What Parse Storyboard does, what it costs, why the first episode of a first drama is free, and what the progress states mean.

# Generating an Episode’s Storyboard

Quick answer: Step 3 is the episode list: one card per episode with its title and the counts of characters, scenes and segments. Press Parse Storyboard on a card and that episode’s outline is turned into a storyboard: segments of a few shots each, with headers, seconds, chips and lines. It is billed by the length of that episode’s script — first and repeat generations cost the same — and the first episode of your first drama is free. While it runs the card says Parsing; when it lands, Edit opens the episode editor.

Open a finished episode to see what a storyboard looks like once it has been generated and shot.

## The episode cards

Each card carries the episode number and title, a thumbnail once there is video, and how many characters, scenes and segments it has. Under the list, Segment length chooses between Standard 15s and Long take 30s for episodes you parse from now on. Segment length.

## Parse Storyboard

The button on a card that has no storyboard yet. Press it and the episode’s outline — the summary the reading wrote in step 1 — becomes a storyboard: the episode cut into segments, each segment a header plus three to five numbered shots with their seconds, the characters and props as @ chips, the lines with voice tags. It is the document you will edit and shoot. What a storyboard contains.

Before it can run, the cast should have images and the episode must be ready. A long script keeps preparing its later episodes after the reading; a card that says Preparing is not ready yet, and the button tells you to try again in a moment.

## What it costs

SceneMixer bills storyboards by the length of the episode’s script: 1 credit per 333 characters of English, shown on the button. First and repeat generations cost the same. The first 1 episode of your first drama is free — the card says Free — so you can see a storyboard before paying for one. Where credits go.

## While it runs

On a new project the card can first show Building the world, four blocks that fill in while the scenes the storyboard uses are prepared; that waiting costs nothing extra (what the four blocks are). Then the card says Parsing, then Refining shots as the storyboard is checked and tidied. Show process opens a live view; Hide closes it. It takes a few minutes. You can parse other episodes meanwhile, or leave; it lands on its own. If your interface language differs from the project’s production language, the card then shows Translating N/M while the storyboard is put into your language; you can open it before that finishes.

## When it lands

Edit opens the episode editor with the storyboard in the middle pane and every segment on the timeline, ready to generate. The editor. Read it before you shoot: a wrong name or a missing seconds value is a free edit now and a paid regeneration later.

## If it fails

The card says Storyboard parsing failed with the reason, and offers Retry. SceneMixer does not charge for a failed storyboard. Failures.

## Re-parse

Once an episode has a storyboard, the button becomes Re-parse. It rewrites the whole episode from the outline and discards its segments and videos — a confirmation says so. Use it when the breakdown is wrong throughout; when one shot is wrong, edit that shot. What re-parsing loses.

| Card says | Means | Do |
| Parse Storyboard | No storyboard yet | Press it (first episode of a first drama: Free) |
| Preparing | Episode not ready yet (long script) | Wait a moment |
| Building the world | The scenes the storyboard uses are being prepared | Wait; four blocks fill in order |
| Parsing / Refining shots | Running | Wait, or Show process |
| Translating N/M | Putting the storyboard into your language | Open it, or wait |
| Edit | Storyboard ready | Open the editor |
| Re-parse | Storyboard exists | Only to rewrite the whole episode |

## FAQ

### How much does a storyboard cost?

It is billed by the episode's script length; the price is on the button. The first episode of your first drama is free.

### Do I have to generate images before the storyboard?

The storyboard does not need them, but the video does. Give the cast images before you shoot.

### Can I parse several episodes at once?

Yes. Press Parse Storyboard on each; they run alongside each other.

### The card says Preparing.

A long script readies its later episodes in the background. Try again in a moment.

### What is the difference between Parse Storyboard and Re-parse?

Parse Storyboard writes the first storyboard. Re-parse rewrites it and discards that episode's segments and videos.

Updated November 3, 2026


---

# Generate Videos With AI: What the Read Takes In and Hands Back

URL: https://scenemixer.com/guides/generate-videos-with-ai/
You write a script, not a prompt. What the AI read takes in, what it hands back, how to write dialogue that survives, and where to steer next, with limits and costs.

# Generate Videos With AI: What Happens to Your Script

Quick answer: To generate videos with AI on SceneMixer you write a script, not a prompt, and the first thing that happens to it is a read. The read takes your text in, finds the characters, scenes and props, splits the story into episodes, and hands back a summary, an episode list and a cast, all as words you can edit before a single picture or video is paid for. Everything after that is built from what the read returned, which makes the script the place where most of the result is decided. This page says exactly what goes in, what comes back, which ways of writing survive the read and which do not, and where you steer next.

Try it: paste one scene written twice, once as lines (Mara (quietly): I saw him.) and once as prose (She told him quietly that she had seen him.). Compare the two episode outlines. The lines survive word for word; the prose becomes a summary.

## You write a script, and the read is the first thing that happens

SceneMixer has no single prompt box that produces a video. Text goes in, a read runs, and three layers of text come out for you to steer: the summary and the cast, then each episode’s storyboard, then each segment’s shots. The video model only ever sees the last layer and the pictures made from the cast. So the useful question is not “what do I type to get this shot” but “what do I write so that the read, the cast and the storyboard all agree on this shot”. Uploading a script and parsing it follows the buttons.

## What the read takes in

| Input | The rule |
| Format | Pasted text, or TXT, DOCX or Markdown. A DOCX is unpacked in your browser first. PDFs and links are refused: export to text or Word, and paste the text itself |
| Minimum | About a paragraph, roughly 330 characters of English. Below that there is nothing to split into episodes |
| Maximum | A hard stop, not a silent cut. Under about 10,000 counted characters the script is read in one pass; above that it is read in two, up to 100,000 counted characters, or 300,000 raw characters of English |
| Counting | Every 3,333 characters of an alphabetic language count as 1,000 |
| Price | 4 credits per 1,000 counted characters, charged once however the script is read, and refunded if the read fails |
| Language | Any. Dialogue in another language than the project’s dialogue language is translated into it at the read |

SceneMixer’s ceiling refuses rather than truncates for a reason: the failure measured on over-long input was a script that came back looking complete and was not, such as a fifty-episode book returned as ten. What you can feed in.

## What the read hands back

The reading panel names its stages: finding characters, finding scenes, finding props, splitting episodes, shaping the visual style, preparing episodes. When it ends, Script Parsed shows:

- the title, the protagonist, the era and how the story was adapted;
- the episode list, cut according to the project’s Type (short episodes with hooks for a Short Drama, longer ones for a TV Series);
- the counts of characters, scenes and items, each with a description and no picture yet.
For a long script SceneMixer’s first pass makes the cast, the second makes the episode list with a short outline each. The first three episodes get their full outline in the background straight away; the others are prepared when you press Generate storyboard on them, and the button waits and continues by itself. If the split is wrong, this is the moment to fix the text and read again. Long scripts are read in two passes · Checking the summary.

## Write dialogue so it survives

| You write | The storyboard gets |
| Mei (quietly): I saw him. | The line, spoken quietly, on Mei’s shot |
| She told him quietly that she had seen him. | A summary. Nothing is spoken |
| Father (on the phone, hoarse): Come home. | The line, heard, without showing the speaker |
| (caption: Three years later) | Kept as a caption line; not drawn into the clip at the moment |
| Crowd: Run! | A group line, spoken together |

One sentence per line, speaker first, delivery in brackets before the colon. Off-screen speech is a note before the line: (on the phone), (voice-over), (over the radio). A standard screenplay layout also works. Write dialogue so it is picked up · Using a standard screenplay as input.

## One spelling for every name and place

SceneMixer’s read matches names literally. A character called MARA in one scene and MRS. CHEN in the next becomes two characters. A place written as MARA’S KITCHEN in one scene and KITCHEN in the next can become two locations, and each one gets its own picture. Pick one spelling, keep scene headings such as INT. KITCHEN – NIGHT, and the read will find the places and the time of day from them.

## What to leave out of the script

- Camera directions and shot lists. SceneMixer’s storyboard plans its own shots, and you can edit them afterwards.
- Tag lists such as “masterpiece, 8K, cinematic”. The models read sentences; the tags add nothing and can be painted literally. No magic words.
- Negatives. “No crowd” puts a crowd on the list. Write the empty street. Say what is there.
- Several stories at once, and long backstory. One script, one story; the read invents what it needs.
- Real people and brands. Keep to invented names.

## The second place to steer: the storyboard

After the read SceneMixer turns each episode into a storyboard on demand: segments of three to five numbered shots, each with its seconds, the people and props as @ chips, and the lines with a voice. This is the document the video model actually sees, and editing it is free. A shot has the shape below, which is an example of the form, not a recording of output.
[Shot 2] 3s: Close-up on @Mara, lit from a window on the left. She turns the key over in her palm. @Mara (quietly): It still fits.
One action per shot, the light by its source, who is in frame by name. Generate the storyboard for an episode · How to write one shot.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Parisian Streets; Costume Style Smart Casual; Visual Style Filmic Live-Action; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the translator and the diplomat as the two leads; the bookshop and the quay as locations; the key as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Layer | Takes in | Hands back | Where you steer |
| The read | Your script or story | Summary, episode list, cast, places and props as words | The text; one name per person; dialogue as lines |
| The cast | Those descriptions | A picture for each character, location and prop | Edit a description; replace a picture |
| The storyboard | An episode outline and the cast | Segments of numbered shots with seconds and lines | Edit the shot text; free |
| The segment | One segment’s text and pictures | One clip | The tier, the seconds, a regeneration |

## FAQ

### Is there a prompt box for generating videos with AI?

No single one. You write or paste a script into SceneMixer, a read turns it into a cast and episodes, and each episode becomes a storyboard whose shots are what the video model sees.

### What should I write to generate a video with AI?

A script: scenes with places, characters named the same way each time, and dialogue as Name: line. Leave out camera directions, tag lists and negatives.

### How long can the script be?

From about a paragraph up to a hard ceiling. Above about 10,000 counted characters it is read in two passes, up to 100,000 counted characters.

### What does the read cost?

SceneMixer charges 4 credits per 1,000 counted characters, once, and refunded if the read fails. The first drama's episode-1 storyboard is free.

### Why was my dialogue turned into a summary?

Speech folded into narration is paraphrased. Write each spoken sentence on its own line as Name: words.

### Why did one character become two?

The read matches names literally. Use one spelling for each character and each place.

### Can I paste a link or a PDF?

No. SceneMixer does not fetch links and refuses PDFs. Paste the text, or export to text or Word.

### Can I edit what the read returns?

Yes. The summary, descriptions and storyboard text can all be edited for free; if the episode split is wrong, fix the text and read again.

Updated November 2, 2026


---

# Generating Many Segments at Once: Concurrency, Queue, Failures

URL: https://scenemixer.com/guides/generating-many-segments-at-once/
Press generate on six segments and only a few start; the rest queue. The concurrency limit per plan, how the queue drains, and why a batch stops or fails.

# Generating Many Segments at Once: How Many Run Together, What “Queued” Means, Why Some Fail

Quick answer: Your account runs a limited number of generations at the same time — 3 on the free plan, 5 on Pro, more on the higher plans. Press generate on more segments than that and the extra ones say Queued: they hold their credits and start automatically as running ones finish. When the balance runs short part-way, pressing generate on the next segment opens the buy-credits window instead and that segment is not submitted. When a whole batch fails, the provider refused the requests for one shared reason, and each card says which.

Select three segments, press generate on each, then a fourth: the fourth shows Queued until one of the first three finishes.

## What you see on the timeline

Each segment you start goes Generating until the account’s limit is reached; after that new ones show Queued. Queued segments have their credits held and start on their own, in the order you pressed, as slots free up. You can leave the page; the queue keeps draining. Meaning of every state: What Each Task Status Means.

## How many run at once

SceneMixer's limit is per account and depends on your plan: 3 on Free, 5 on Pro, 8 on Studio, 10 on Enterprise. It counts every generation — video segments, sheets, storyboards, composites — not just video. A busy provider can add wait on top; see Why a Generation Takes Time.

## When the balance runs short part-way

Each segment holds its credits the moment you press generate on it. When the balance cannot cover the next one, the buy-credits window opens instead and shows how much you are short; that segment is not submitted and holds nothing. Top up and press generate on it; the segments already running or queued are unaffected and nothing is charged twice. The same thing in Step 2.

## When the whole batch fails: the provider

If all of them show the same provider message — busy, invalid input, moderation — the cause is shared: the same tier was busy, the same reference was unusable, or the same character sheet was blocked. Read one card; fixing that one thing fixes the batch. A busy provider’s refusal is refunded automatically; a moderation block is refunded up to 12 blocks a day per account, and past that held for review. Your credit history shows which. When a Generation Fails and Invalid Input.

## A sensible batch

Generate the first segment alone, check the faces and the room, then start the rest. Six segments generated blind from a first shot that was wrong are six regenerations. Order matters too: segments that carry over from the previous one look better when the previous one already exists.

| Plan | Generations at once |
| Free | 3 |
| Pro | 5 |
| Studio | 8 |
| Enterprise | 10 |

## FAQ

### Is 'Queued' a failure?

No. It is waiting for a slot and starts by itself. Its credits are held, not lost.

### Can I skip the queue?

A higher plan raises the limit. Otherwise wait for a running one to finish.

### Do queued segments cost extra?

No. Same price as any segment, held while they wait.

### I closed the tab. Did the queue stop?

No. Queued and running generations continue on the server.

### Why did only some start immediately?

The limit was reached; the rest queued.

Updated November 3, 2026


---

# Generative AI Video: What Is Generated and Where You Hold the Controls

URL: https://scenemixer.com/guides/generative-ai-video/
What makes video generative rather than template-edited, the controls you really have (pictures, words, settings), how firmly each one holds, and what stays random.

# Generative AI Video: What Is Generated, and What You Control

Quick answer: Generative AI video means the pictures are produced by a model from a description, not cut from existing footage or dropped into a template. That is why you can ask for a scene nobody filmed, and also why you should not expect an identical result twice. The practical question is not whether a tool is generative but where it lets you hold the controls. In a pipeline like this one they come in three kinds that differ in how firmly they hold: pictures bind, words steer, settings frame. A reference picture is copied almost literally; a sentence of text suggests; a setting such as style, tier or shape sets the box the rest happens in. This page sorts what you can control into those three, and says what no control reaches.

Try it: pick one thing you care about in a shot, a face, a coat, where someone stands, the weather, and decide whether it should be set with a picture, with words or with a setting. The table below says which one holds.

## Generated or assembled: a quick way to tell

A template or stock editor assembles: it takes clips, text and music that already exist and places them. Nothing in the result was made for your sentence, so it is repeatable, cheap per change and limited to what exists. A generative tool makes: the picture, the speech and the sound are produced for your request. You can ask for something no library holds, a character with your description in a room you described, but each request is a new draw.

The test is easy. Ask for something that cannot be in stock footage, such as a named character in a specific outfit saying a specific line. If the tool gives you back a close match of existing material, it is assembling. Here, a segment is produced from your cast pictures, your shot text and your voices; nothing in it is retrieved from a library of clips. The join at the end is the one non-generative step: it stitches finished segments together.

## Pictures bind

When a shot is generated, the model receives the sheets of the characters, the scene and the props named in it as pictures, and it reproduces them closely: face, clothes, materials, and whatever light and background are in the picture. That is the strongest control there is, and it cuts both ways. A picture binds what it shows, and no sentence will undo it. So the way to change a face or a coat is to change the picture: edit the description and regenerate the sheet, replace it with your own image on a paid plan, or add a state picture for a changed look.

It is also why sheets are neutral: a character drawn under noir lighting brings that lighting into a sunny picnic. Mood is written per shot in the words, where it belongs. Why sheets look plain · Using your own reference images.

## Words steer

The shot text decides what the picture does that the pictures do not: where people stand, what they do, where the light comes from, the weather, the time of day, how the camera moves and, word for word, what is said. Words work on the layer above the pictures, so rain, dusk, a person sitting down and who holds a cup all ride on top of whatever the sheets show. They steer rather than fix. Nouns weigh more than function words, so a negative often produces the thing it names, and a list of adjectives pushes the important words down. Write what is in front of the camera, with the cause of the light, not the verdict you want on the result.

Describe the cause, not the result · No magic words · Write camera moves.

## Settings frame

SceneMixer’s settings set the box that pictures and words work inside. The visual style is the medium, from photographic to painted to 3D, and applies to everything. Scene style and costume style are applied once, when the script is read, and flow into every description. The shape, 9:16 or 16:9, is fixed after Step 1. The video tier sets the resolution, from 480P to 2K, and the longest segment, 15 or 30 seconds; the seconds in each shot add up inside that ceiling. Settings never say what happens in a shot, and the system keeps them from fighting the words: it does not tell the model where to put the subject, because that is the storyboard’s job.

Choosing your style settings · 15 or 30 second segments · Vertical or landscape.

## What you want to control, and which control holds it

| You want to set | Use | How firmly |
| A face, age, build | The character sheet | Binds: copied closely |
| An outfit | The sheet, or a state picture for a changed outfit | Binds; words cannot override it |
| The layout of a room | The scene sheet and the scene’s 3D model | Binds the space |
| Where people stand | Words, as room anchors | Steers; holds across cameras |
| Light, weather, time of day | Words in the shot | Steers; works on top of the pictures |
| What is said | A line with speaker and voice tags | Spoken word for word |
| How a voice sounds | A written voice or a preset from the library | Steers, or binds once a preset is attached |
| Length and detail | Seconds per shot, the tier | Frames: within the ceiling |
| Exact framing, small gestures, hands | Nothing reliably | Left to the model; regenerate |

## What stays random, and how to work with it

Two identical requests can give two different clips; SceneMixer’s terms say so and the honest version of any generative product is one where that is expected. The parts that vary are the exact framing, the performance, small details and anything you left unsaid. Work with that, not against it: read the storyboard before you pay for video, shoot one segment first, fix single shots by editing their text and regenerating only those, and keep earlier takes, since every generation of a segment is kept in Version History and switching between them is free. Edit one shot and regenerate · Segment versions.

Because every clip is a new draw, cost is per attempt, not per result: 10 credits a second on the default tier. The cheapest control you have is the one before generation, the text, which is free to edit as often as you like.

| Kind of control | Examples | Holds by |
| Pictures | Character, scene and prop sheets, state pictures, your own photo | Being copied closely |
| Words | Shot text, positions, light, weather, dialogue lines | Steering what the pictures leave open |
| Settings | Visual, scene and costume style, shape, tier, seconds | Setting the box everything happens in |
| None | Exact framing, gestures, hands | Regenerate and choose |

## FAQ

### What is generative AI video?

Video whose pictures, speech and sound are produced by a model from a description, rather than assembled from existing footage or a template.

### How is it different from a template video editor?

A template editor places existing clips, text and music, so it is repeatable and limited to what exists. A generative tool makes new picture and sound for each request, which allows new scenes but not identical repeats.

### What can I control in a generative video?

Pictures bind a face, outfit and room; words steer positions, light, weather and dialogue; settings frame style, shape, resolution and length. Fine framing and gestures are left to the model.

### Why does my text not change the character's clothes?

A reference picture is copied with its clothes, and text cannot override it. Change the picture, edit the description and regenerate the sheet, or add a state picture.

### Can I get exactly the same video twice?

No. Generation is probabilistic. You can keep a take you like, switch between takes for free, and regenerate only the shots that miss.

### Is any part of the video not generated?

The join that stitches finished SceneMixer segments into an episode is ordinary video processing. Voices come from a library of presets or a written description, and the score is chosen for the episode.

### Does generative mean random?

Partly. The pictures and words you supply hold the identity, room and story; what varies is the exact framing, performance and small detail.

### What is the cheapest control?

The text. Editing the storyboard is free in SceneMixer, and a sentence fixed before generation saves a regeneration at the per-second price.

Updated November 2, 2026


---

# Give a Character a Personality the Video Can Show

URL: https://scenemixer.com/guides/give-a-character-a-personality-the-video-shows/
Personality is not drawn into a character sheet. It comes from your script, the storyboard’s actions and delivery notes, and the voice description. Where to write it.

# Where a Character’s Personality Goes, and How to Make the Acting Show It

Quick answer: A character sheet shows how someone looks; it cannot show that she is proud or that he is a coward. Personality reaches the video by three other roads. The script: when it is read, each character gets a short note of their personality, taken from the story; it is not shown on the card, but it is passed on whenever a storyboard is written and when a voice is matched. The storyboard: what a character does in a shot, the delivery note in brackets, and the words of the line are what the camera actually films. The voice description: how they sound. To change how someone comes across, rewrite their actions and lines in the storyboard — one request can do it for a whole episode — and regenerate.

Open a segment with your character in it, press Edit, and read only what they do and how they say it. If the shots could belong to anyone, that is why the clip does not show who they are.

## Why the character sheet does not show personality

The sheet is a reference image: plain background, even light, a neutral pose, so that the same face can be placed into every shot. It is drawn from the card’s Description, which is about appearance. Words like brave, kind or arrogant have nothing to draw; at best they are ignored, at worst they turn into a fixed expression that follows the character into every scene, including the ones where she should be laughing. Why sheets are plain · Describe a face so it is not a generic face.

What a description can carry is a physical habit that shows who they are without a word: posture, how they stand, a resting expression in the face’s structure. Stands very straight, shoulders back, chin slightly raised reads as pride and still works in a sad scene; always smiling does not.

## Where personality actually comes from

| Where | What it carries | Can you edit it? |
| Your script | The reading writes a short personality note for each character from what the story shows. It is used when storyboards are written and when a voice is matched. | Not on the card; it comes from the script as it was read |
| The storyboard | What they do in each shot, the delivery note, the words of each line, how others react to them | Yes, freely, or with AI edit |
| Voice description | How they sound: pitch, pace, texture, an accent | Yes, on the card |
| Description | How they look, including posture and resting expression | Yes, on the card; regenerate the sheet |
The personality note is the reason a cold character is often written cold in the storyboard without you asking. It is also why a character you added by hand, with no script behind them, comes out flatter: the storyboard has only their description to go on. For those, put the personality into the storyboard yourself.

## Write it into the script

The strongest place is the story itself, before it is read. Readers learn who someone is from what they choose and how they say it, and so does the reading:

- Give them a choice that costs something. “She hands back the envelope without opening it” tells more than “she was honest”.
- Give them a way of speaking. Short sentences, never a question; or a joke at the worst moment.
- Let others react to them. The room going quiet when he walks in defines him better than an adjective.
- A character list at the end of a script helps, but the scenes are what the episodes are built from.
Write dialogue in your script so it is picked up · AI Create: writing the one-line brief.

## Make the acting show it, shot by shot

In the storyboard, personality is behaviour. The same line said by two different people:

| Personality | How the shot is written |
| Cold, in control | [Shot 2] 4s: Close-up. In frame only: @Ethan. He reads the report without looking up and turns the page with one finger. @Ethan (flat, not looking up): Again. (voice: 🔊Ethan) |
| Nervous, eager to please | [Shot 2] 4s: Close-up. In frame only: @Lily. She holds two printouts against her chest, glancing from one to the other. @Lily (too fast, swallowing the last word): I did both versions, just in case. (voice: 🔊Lily) |
| Warm, teasing | [Shot 3] 3s: Medium shot. @Nora leans on the doorframe, arms folded, half a smile. @Nora (amused, slow): You’re early. That’s new. (voice: 🔊Nora) |
Three things carry it: the action before the line, the delivery note in brackets (how it is said, not what it means), and the words themselves. Keep thoughts out of the description; a sentence of what a character thinks in a dialogue shot can be read out loud. How to write one shot · The video says a line you never wrote.

## Change a character across the whole episode

If the storyboards came out with the wrong temperament — the villain too friendly, the lead too passive — you do not have to rewrite every shot by hand. Open AI edit on a segment and use Apply to the whole episode… with one request, for example:

Make Ethan colder in every scene: fewer words, never smiles, does not look at people when he talks to them.

The same request runs on every segment you select; segments that need no change are not charged, and you review and save the results. SceneMixer prices each revision on the button (1 credit per segment revised). Then regenerate the segments you want to see acted differently. Revise the whole episode with one request.

## Give the personality a voice

The Voice Description on the card is how a character sounds in every line. Write it as sound, not as character traits: low, unhurried, clipped consonants, never raises his voice rather than confident CEO. A matched preset voice is picked from that description and the character’s note; after changing the description, Re-match picks again for free. How character voices work · Give a character an accent.

## Keep it consistent across a series

- Same habits, every episode. Pick two or three physical habits per lead — how she enters a room, what he does with his hands — and keep using them in the storyboards.
- Change on purpose. If a character changes over the series, make the change visible in a scene; a different temperament with no cause reads as a mistake.
- Reusing a character in another project brings their traits with them, the personality note included. Reuse a character across projects.

| You wrote | What happens | Write instead |
| “Brave and kind” in the description | Nothing to draw; may become a fixed expression | A posture or a habit, in the description; behaviour, in the shots |
| “She is furious but hides it” before her line | Can be read aloud | Her hands go still; she smiles a beat too long |
| (angry) as the delivery note on every line | Every line sounds the same | Vary it: (quiet, very slow), (clipped), (almost laughing) |
| “Confident CEO” as the voice description | A guess at a voice | low, unhurried, clipped consonants |

## FAQ

### Can I write a character's personality on the card?

SceneMixer has no personality field on the card. The description is for how they look, and the voice description for how they sound. Personality comes from your script when it is read, and from what the storyboard has the character do and say.

### Why does the character sheet not show the personality?

A sheet is a neutral reference image used in every shot; it shows appearance only. A trait like proud or shy has nothing to draw. Show it in the shots instead.

### Does the site know my character's personality at all?

Yes. When SceneMixer reads the script, each character gets a short personality note from the story. It is not shown on the card, but it is used when storyboards are written and when a voice is matched.

### My villain comes out too friendly. What do I change?

His actions, delivery notes and lines in the storyboard. AI edit with Apply to the whole episode can make the change across every segment with one request; then regenerate the segments.

### Should I write adjectives like brave or arrogant in the description?

Better not. They cannot be drawn and can turn into the same expression in every scene. Write a posture or habit instead, such as stands very straight with her chin raised.

### Does personality change the voice?

It helps pick a matched voice, along with the voice description. To change how someone sounds, rewrite the voice description in sound words and press Re-match.

### I added a character by hand and they feel flat. Why?

A character added by hand has no script behind them, so storyboards only have their description. Write their behaviour and delivery into the shots yourself.

### Do I have to regenerate after changing the storyboard?

Yes. Editing the text is free, but a finished clip only changes when the segment is regenerated.

Updated November 3, 2026


---

# Give a Character an Accent or Dialect in an AI Drama

URL: https://scenemixer.com/guides/give-a-character-an-accent/
Write an accent or dialect into a character's voice description and every line they speak comes out in it, plus the one thing to check first.

# Give a Character an Accent or Dialect in an AI Drama

Quick answer: Every character has a Voice description — one line of free text that says how the voice sounds. Put the accent in it. The line travels with every sentence that character speaks, and the video model performs it while it renders the shot, so a light Taiwanese Mandarin in the description is a character who speaks with a light Taiwanese accent from their first line onward. One thing decides whether this works at all: the character must have no audio attached. If a preset voice was matched or a recording uploaded, the description is not used, and an accent written in it goes nowhere. Check the badge before you generate — it takes ten seconds and it is the whole difference between working and not. How character voices work.

Open a character in step 2 and the Voice description box is at the bottom of the dialog, under the appearance description. That single box is where an accent lives.

## How an accent reaches the screen

Nothing about the voice is recorded, dubbed or stitched together afterwards. When a segment is generated, the video model renders the picture and says the lines in the same pass, and the description of each speaking character is handed to it as the brief for that voice. That is why the mouth shapes match the words: the voice is part of the generation, not a layer added to it. Dialogue and sound in one generation.

The accent words are not filtered, shortened or dropped: “soft and breathy, a light Taiwanese Mandarin, unhurried” travels with every line that character speaks, in every segment — and nothing asks whether you meant it. If you want Hong Kong-accented Mandarin, you say Hong Kong-accented Mandarin.

The one job the voice library cannot do is accents. It is organized by language — Chinese and English — and a match is chosen by gender, age band and manner. There is no Taiwanese-accented entry, no Scottish entry, no Brooklyn entry, and the accent words in your description play no part in choosing one. Accents come from the description; the library is for holding a voice steady. How character voices work.

## Add it in four steps

- Open the character. In step 2, Characters & Scenes, click the character’s card. The same field is in the character’s details in the episode editor; both are the one description that follows the character around the project.
- Read the voice row before you type. If it offers Auto-match voice, Pick from library and Upload Voice, this character has no audio — that is the state you want, and everything you write in the next step will be used. If a voice is attached instead, see the badge check below first.
- Add the accent to the Voice description. Keep what is already there — timbre, pace, manner — and put the accent at the end. The box takes a full sentence or two; there is no length limit to work around, and one sentence is usually enough.
- Save and generate. The accent applies from the next generation onwards, to every line the character speaks.

### Check the badge: is the description being used?

Two places show which of the two modes a character is in. On the character card in step 2, a small speaker icon sits next to the name when an audio voice is attached — no icon means the description is in charge. In the storyboard editor, the @ menu has a Voice tab listing the cast: each character has a speaking icon, and one who has no audio also gets a small Text only tag next to it. Hovering the icon says which of the two it is — Voice assigned, or Voice description only (no audio). The icon without a tag is the green light; the speaker means an audio sample is attached and the description, accent included, is not what gets performed.

If you find yourself in the wrong state, open the character and use Delete Voice — the confirmation says the remote file is kept, so nothing is destroyed and you can pick a voice again later. It is free, and it takes effect on the next segment you generate. Then your accent words are back in play.

## What to write, and what you get

Say which accent, and say how much of it. “An accent” is not a description; “a light Taiwanese Mandarin” is. These are the same one-line format the description already uses, with the accent appended:

| What you write | What you are asking for |
| Low and warm, unhurried, a light Taiwanese Mandarin | Mandarin with a Taiwanese accent, said gently and without hurry |
| Bright and quick, Hong Kong-accented Mandarin, clipped endings | The Hong Kong flavour on Mandarin, with the clipped rhythm that goes with it |
| Relaxed and a little husky, a Sichuan accent, conversational | A Sichuan-flavoured Mandarin, casual rather than performed |
| Neutral standard Mandarin with only a trace of a Northeastern flavour | Standard Mandarin with the faintest regional colouring |
| Measured and gravelly, a soft Scottish lilt, unhurried | English with a gentle Scottish lilt |
| Warm and slow, a soft Southern US drawl | English with the stretched vowels of the American South |
| Flat and weary, a London accent, clipped | English that sounds like London, tight and tired |
| Light and quick, English with a slight French accent | English spoken by someone whose first language is French |

Three habits make it work.

- Keep the original description. Timbre, pace and manner are what the character already is; the accent sits on top of them. Replacing the whole line throws away a description that was working.
- Grade the strength. a trace of, a light, a clear, unmistakable mean different things to the model, and they are the difference between a character who occasionally slips into the region and one who cannot be mistaken for anywhere else. If a first try comes out too faint, rewrite the strength word and generate the segment again.
- Write instructions, not comparisons. “Like a newsreader” or “sounds like a street hawker” reads as an instruction and gets amplified: the description is repeated on every segment, so a comparison turns into the character’s permanent manner instead of an accent. “Clear Taiwanese Mandarin, light and breathy” says what you want and stays where you put it.

## Accent, dialect, or a different language?

Three different things, three different places to put them. Mixing them up is the usual reason an accent “does not work”.

| What you want | Where it goes |
| The cast speaks the project’s dialogue language with a regional flavour — Taiwanese, Hong Kong, Sichuan, Northeastern, Shanghai, Beijing-style erhua, a rural English drawl | The character’s Voice description. The lines stay in the project’s dialogue language and wear that region on top |
| The lines themselves are written in that region’s idiom — actual regional expressions rather than standard ones | The dialogue itself, in the script or in the storyboard line. Wording plus accent together is what makes a dialect land |
| The cast speaks a different language altogether — English, Japanese, Korean | The project’s dialogue language setting, not the description. How to set the dialogue language |
| This one line is angry, whispered, hurried, or stressed on a word | The delivery note in brackets at the start of that line. Do not repeat the accent there — it is already on the character. Dialogue in the storyboard |

The distinction that matters: an accent changes how the language is spoken, a dialect setting changes what is spoken. Asking for both in one place fights itself — writing “speaks Japanese” in a description on a project whose dialogue language is English leaves the model with two instructions about the same sentence that pull against each other. If you want Japanese, set Japanese as the dialogue language and put any regional colouring in the description on top.

## When the accent does not come through

| What you see | What it comes from | What to do |
| Nothing at all changes after you save | The character has audio attached, so the description is not what is performed | Check the badge; on the card use Delete Voice, then generate the segment again |
| The accent is there but too faint | The strength word was vague, or the region was named only in general terms | Name it more precisely and grade it — “a clear, unmistakable Sichuan accent” — then regenerate that segment |
| Old episodes still sound the way they always did | Finished segments keep the voice they were shot with; that is what makes a delivered episode stable | Regenerate the segments where you want the accent. It costs the same as generating them the first time. What each clip fix costs |
| The accent is missing from a clip you fixed after the fact | The finished-clip voice panel synthesizes speech with preset voices from the library, so an accent written in the description does not come out of it | Regenerate the segment instead of redubbing it. Changing a voice in a clip you already have |
| The voice is right but drifts between segments | The description mode trades a little consistency for freedom — each generation performs the words afresh | Accept it, or move that character onto a preset voice from the library and give up the written accent for that character |
| The character’s voice is fine but a stranger’s line is odd | Lines spoken by someone outside the cast do not carry a character’s description | Give that speaker a character card of their own and write their accent there too, or leave it as a one-off line |

## What it costs

- Free: writing the description, editing it, matching a voice, changing a voice, deleting one and going back to the description.
- Member feature: uploading your own recording as a character’s voice. Members can upload a clip they have the right to use; the platform does not clone named people, and someone else’s voice without permission is not permitted.
- The only charge: generating a segment. Writing an accent is free; hearing it is an ordinary generation.

One thing to know before you reach for an uploaded voice to solve this: an attached recording and a written accent are alternatives, not partners. Attach audio and the description stops being performed — accent words included. Members who want a specific regional voice use the upload deliberately, in place of the description. How character voices work.

## FAQ

### Does writing an accent really change how the character speaks?

Yes. The voice description travels with every line that character speaks and is performed as written: the video model says the line and renders the shot in one pass, so there is no separate dubbing step that could drop it. What you do not get is a strength setting — if an accent comes out too light, say so more explicitly and generate that segment again.

### Why did my accent have no effect?

Almost always because the character has audio attached: a preset you matched or a recording you uploaded. With audio in place the description is not what the model performs. On the character card use Delete Voice to go back to the description, then generate the segment again.

### Can I add an accent after I have already generated episodes?

You can, and everything you generate from then on uses it. Segments you already shot keep the voice they were shot with; regenerate the ones where you want to hear the accent, at the usual cost of generating a segment.

### Can I add an accent from the voice panel of a finished clip?

No. That panel synthesizes speech with preset voices from the library, so an accent written in the description does not come out of it. An accent only arrives in a segment that is generated again.

### Does the voice library have regional accents in it?

No. SceneMixer's library is organized by language — Chinese and English — and a match is chosen by gender, age band and manner, not by the accent words in your description. The description is the only place an accent comes from.

### Is an accent the same thing as the dialogue language?

No. An accent or dialect keeps the project’s dialogue language and puts a regional flavour on it, so it goes in the voice description. To change the language itself, set the project’s dialogue language; writing speaks Japanese in the description while the dialogue language stays English asks for two things at once, and they pull against each other.

### Do I have to repeat the accent on every line?

No. It is a property of the character, written once in the voice description, and it covers every line they speak. What changes line by line — mood, pace, a word under stress — goes in the delivery note in brackets at the start of that line.

### Can two characters have different accents, and what about characters who are never seen?

Both. Each character carries their own voice description, so accents never interfere. Voice-only characters — narrators, the person on the other end of the phone — have a voice description too, and the accent works the same way for the lines they speak.

Updated November 3, 2026


---

# Glossary: Every Term This Site Uses | SceneMixer

URL: https://scenemixer.com/guides/glossary-of-terms/
Segment, shot, setup image, state, take, freeze, tier. Short definitions for every term this site uses, each pointing at the guide that explains it properly.

# Glossary

Quick answer: A few of these words mean something specific here and something else everywhere else — segment and shot in particular. This page is the definition point; each entry links to the guide that goes into it.

## The project

Project — One drama, ad or remake. Holds the settings, the cast and the episodes. What is inside a project

Reading the script — The first paid step for a drama you brought a script to — a remake starts by reading your video instead. Turns your text into a cast and a list of episodes. Art direction takes effect only here and wardrobe mostly does, which is why both are chosen beforehand. Choosing your style settings

Episode — A numbered part of the story with its own storyboard.

Cast — The characters, locations and props belonging to a project. Each has a description and a generated image.

Content type — Short drama, long drama or film. A structural switch: episode length, dialogue density, the performance grammar the storyboard is written in. Choosing your style settings

## Images

Setup image — The reference picture of a character, location or prop. Every later segment that uses that character or place points at it, which is what keeps a face the same across episodes. Keeping a character consistent

Character sheet — A character's setup image. SceneMixer always draws it at 9:16, whatever the project's aspect ratio is, because a full-length standing figure needs the height.

State image — An alternative setup image for the same subject after a change a sentence cannot fix — a full costume change, visible injury, a hall after it burns, a blade drawn. Costume changes and scene states

Override — Using a different image for one episode, or one state for one segment, without touching the cast. What is inside a project

## The storyboard

Storyboard — The written plan for one episode: its segments, and inside each the numbered shots with their durations, who is in frame, and the spoken lines.

Segment — One generation. Fifteen or thirty seconds of finished video, chosen per episode when you generate its storyboard. This is the unit you pay for and the unit the model produces in one go — not a cut. Aspect, length, resolution

Shot — A numbered part of a segment with its own duration. Shots inside one segment are cuts the model performs within a single generation; their durations add up to the segment's length.

In frame — The list of who is actually visible in a shot. It decides who gets a face and a reference image. Someone mentioned but not in frame is written as off-screen and is not given a face. When a shot comes out wrong

Background people — A line in a segment's header declaring that a crowd is present. Declared consistently or not at all — a crowd that appears in one segment and not the next usually traces back to this. When a shot comes out wrong

Voice tag — The marker on a spoken line saying which voice reads it. Picking character voices

## Video and delivery

Take — One generated video for one segment. Every take is kept and switching between them is free. Fixing one shot

Tier — A named row of configuration: which model, which resolution, the longest a segment may be, and the price. All three kinds are changeable in settings; video is the one worth changing. What runs underneath

Compositing — Joining an episode's segments into one file. The full-quality segments are copied rather than re-encoded. Compositing and sharing

AI badge — The small label burned into the corner of delivered video, plus a line of AI-generated-content metadata in the file. Rights and labels

## Credits and queueing

Credit — The unit everything is priced in. SceneMixer credits are bought in packs or granted by a plan; both land in the same balance. Where credits go

Freeze — Credits held aside when a job starts — not yet spent. Charged if it succeeds, released if the model never ran, held pending a manual check if we cannot tell. When a shot comes out wrong

Parked — A job waiting because your account already has as many running as your plan allows. It is promoted automatically when a slot frees; nothing failed. Why a generation takes time

Slot — One of the generations your account may run at the same time. A paid plan raises the count; it does not make a single generation faster. What a paid plan changes

Allowance — The credits a plan grants at the start of each period. They go into the same balance as bought credits and do not expire monthly. What a paid plan changes

## Words that mean something else here

Three worth flagging, because borrowing the wrong intuition from editing software is the most common source of confusion.

“Segment” is not a cut — It is one call to the model. A cut inside it is a shot, and cuts inside one segment are free in a way that continuity across two segments never is.

“Regenerate” means two different things — Regenerating one segment's video is safe and keeps the old take. Regenerating an episode's storyboard rebuilds the episode and discards every finished video in it. Fixing one shot

“Delete” has no undo — SceneMixer has no recycle bin and no restore button anywhere on the site. Where your files live

## FAQ

### What is the difference between a segment and a shot?

A segment is one generation, around fifteen seconds. A shot is a numbered part of a segment with its own duration — a cut the model performs inside that single generation. You pay per segment.

### What does it mean when credits are frozen?

They are held aside rather than spent. If the job succeeds they are charged; if the model never ran they are released; if we cannot tell, they stay held for a manual check.

### What is a setup image for?

It is the reference every later segment points at, which is what keeps a character's face and a location consistent across an episode.

### What is a tier?

A named configuration row: which model, which resolution, the maximum segment length and the price per second. You choose the video tier.

### Why is my job 'waiting' rather than failing?

Your account already has as many generations running as your plan allows, so it is parked and will be promoted automatically when one finishes.

Updated November 3, 2026


---

# GPT Image 2.5 on SceneMixer: Sunburst and Flare, Price per Image

URL: https://scenemixer.com/guides/gpt-image-2-5/
Use OpenAI's GPT Image 2.5 Sunburst or Flare for character, location and prop sheets and promo covers on SceneMixer: price per image, setup, which one to pick.

# GPT Image 2.5 on SceneMixer: Sunburst and Flare for AI drama sheets and covers

Quick answer: SceneMixer offers both GPT Image 2.5 models as image tiers: GPT Image 2.5 Sunburst and GPT Image 2.5 Flare, 25 credits per image each (about $0.16 at the lowest credit price). They draw the reference images an AI series is matched to — every character on white, every location as an empty plate, every prop, every second outfit or injury — and the promo cover posters. Both draw at about 2K (2,048 pixels on the long side). OpenAI positions Flare as the faster general-purpose model and Sunburst as the slower, more precise one. Pick either in Settings → Model Preferences → Image Model. The free sheets are not drawn on these tiers; they stay on the default model.

Choosing between the two: Sunburst vs Flare on SceneMixer. Against Google’s top tier: GPT Image 2.5 vs Nano Banana Pro.

## What GPT Image 2.5 is

OpenAI launched ChatGPT Images 2.5 on September 8, 2026, and with it two API models: Flare and Sunburst. OpenAI describes Flare as the general-purpose model, faster than the previous generation, and Sunburst as the slower one built for precise work, keeping earlier edits and reference subjects intact across turns. Those are OpenAI’s own claims; the sources are at the end of this page.

On SceneMixer you do not need an OpenAI account, an API key or a ChatGPT plan. The two models are tiers inside the site and are billed in SceneMixer credits, like every other tier.

## What it draws on SceneMixer

An AI drama series stays consistent because every shot is generated against the same reference images. SceneMixer reads your script, lists the cast, locations and props, and draws a sheet for each: a full-length casting photo on white for a character, an empty room with its light source readable for a location, the object alone on white for a prop. A second look — the wedding gown, the cast on a broken arm — becomes a state image drawn from the main sheet. The image tier you pick decides how those are drawn.

The same tier also paints the promo cover: three poster candidates with your title painted into the picture, built from your leads’ sheets. See make a promo cover poster for your AI drama.

It does not make the video. Shots are rendered on the video tiers; GPT Image 2.5 draws the pictures they are matched to.

## Where to switch it on

Settings → Model Preferences → Image Model, then pick GPT Image 2.5 Sunburst. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

- Open SceneMixer’s Settings from the account menu.
- Go to Model Preferences → Image Model.
- Choose GPT Image 2.5 Sunburst or GPT Image 2.5 Flare. The price per image is shown next to each tier.
- Back in your project, generate or regenerate the sheets in step 2.

## Sunburst or Flare

On this site the two cost the same, 25 credits per image, and draw at the same size, so the choice is about what each model is built for.

- Flare for drafting and for volume: the whole supporting cast, background locations, props. It is the faster of the two.
- Sunburst where fidelity to a reference matters: state images drawn from a lead’s main sheet, and the cover posters, which are drawn from your leads’ sheets. OpenAI says this is the model that holds a reference subject best.

A practical split: draw the cast once on Flare, look at the leads, switch to Sunburst and regenerate only the leads you want to improve. The full comparison is in Sunburst vs Flare.

## Writing descriptions that work on GPT Image 2.5

The rules that work on every tier work here. Describe what can be drawn, in plain sentences, in the order face → one anchor feature nobody else has → build and posture → one outfit. Describe causes, not results: a burn scar across the back of the left hand, not a hint of an old injury. Leave lighting and mood out of a sheet — sheets are plain on purpose, because whatever light is on the sheet is carried into every shot.

Write what is there rather than what is not: “no hat” is a weaker instruction than describing the hair. See say what is there, not what is not and describe a face so it is not a generic face.

## What it costs

This tier is not covered by SceneMixer’s free sheets. Every account’s 15 free sheets are drawn on the default image model, Qwen-Image 3.0, only. On GPT Image 2.5 Sunburst each sheet costs 25 credits from the first one (about $0.16 at the lowest credit price), and the price is printed on the button before you generate. Switch back to the default and whatever free sheets you have left are still there. A generation that is refused is not charged, and most failures are refunded automatically; a few kinds are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

For scale: a cast of six leads drawn once each on Sunburst is six images. Most of what a series costs is video seconds, so a better tier for the faces the audience watches most is usually worth it; the rest can stay on the default.

## Common errors

What you may meet on these tiers:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the description tripped the site’s check, or the model maker’s own filter refused the picture | Edit the description and generate again; nothing was charged |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The model was saturated; the site kept the job queued for a while before giving up | Generate again in a few minutes, or switch tier |
| A sheet that ignores part of the description | Too many details competing in one description | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |
| Credits held for review | A rare failure where the model may already have been paid for the attempt | Nothing to do; support checks it and releases or refunds — see credits held for review |

## All the image tiers, side by side

Ten image tiers can be picked in Settings on SceneMixer today. The price is per image; the free sheets are on the default only.

| Image tier | Credits per image | In short |
| Qwen-Image 3.0 (default) | 7 | The free sheets; a whole cast cheaply |
| Qwen-Image 3.0 Pro (Recommended) | 18 | Leads in close-up; Chinese descriptions |
| SeedDream 5.0 Lite | 8 | 2K locations and props on a budget |
| SeedDream 5.0 Pro | 25 | 2K, photographic skin, long descriptions read to the end |
| Nano Banana 2.1 | 19 | About 2K; Google’s faster model, the cheapest of the six new tiers |
| Nano Banana Pro | 46 | About 2K; detail kept from long descriptions; the most expensive tier |
| GPT Image 2.5 Flare | 25 | About 2K; OpenAI’s faster general-purpose model |
| GPT Image 2.5 Sunburst | 25 | About 2K; OpenAI’s slower, more precise model |
| Grok Imagine Image 2.0 | 35 | About 2K; xAI’s model, strong on lettering |
| MAI-Image-2.6 | 31 | About 1 megapixel; Microsoft’s model, strict content filter |

Which image tier for your sheets turns this into a choice by task.

## Sources

- The Rundown: OpenAI ChatGPT Images 2.5 — launch coverage, September 8, 2026; checked October 25, 2026.
- fal: GPT Image 2.5 — the two API models, Flare and Sunburst, as described at launch; checked October 25, 2026.
GPT Image and ChatGPT are trademarks of OpenAI; they are named here to identify the models offered as tiers. Prices on this page are SceneMixer’s own and are read from the live price list.

| | GPT Image 2.5 Flare | GPT Image 2.5 Sunburst |
| Shown in Settings as | GPT Image 2.5 Flare | GPT Image 2.5 Sunburst |
| Credits per image | 25 | 25 |
| Output size | About 2K (2,048 px long side) | About 2K (2,048 px long side) |
| OpenAI’s positioning | General purpose, faster | Slower, more precise; holds references |
| Use it for | Supporting cast, locations, props, drafts | Leads, state images, cover posters |
| Free sheets | No — default model only | No — default model only |

## FAQ

### Can I use GPT Image 2.5 on SceneMixer?

Yes. Both GPT Image 2.5 models are image tiers. Choose GPT Image 2.5 Sunburst or GPT Image 2.5 Flare in Settings, Model Preferences, Image Model, and every new image is drawn on it.

### How much does GPT Image 2.5 cost on SceneMixer?

25 credits per image for either model, about $0.16 at the lowest credit price. The price is shown on the button before you generate.

### Is GPT Image 2.5 free here?

No. Every account's 15 free sheets are drawn on the default image model, Qwen-Image 3.0. On SceneMixer, GPT Image 2.5 charges each image from the first one.

### What is the difference between Sunburst and Flare?

OpenAI built Flare as the faster general-purpose model and Sunburst as the slower, more precise one. On SceneMixer they cost the same and draw at the same size.

### Do I need an OpenAI API key or a ChatGPT plan?

No. The tiers are part of SceneMixer and billed in SceneMixer credits.

### What resolution does GPT Image 2.5 draw at here?

SceneMixer draws at about 2K: 2,048 pixels on the long side, for example 1152 by 2048 for a vertical image.

### Can GPT Image 2.5 make the video?

No. It draws the sheets and cover posters. SceneMixer makes the shots on the video tiers, which use those sheets as references.

### Does it keep my characters consistent?

Consistency on SceneMixer comes from drawing each character once and reusing that sheet in every shot, on any tier. A tier decides how well the sheet itself matches your description.

### Can it paint the title on my cover?

Yes. SceneMixer draws the promo cover on the image tier you have selected, and the model paints the title into the poster. Check the spelling before you pick one.

Updated November 3, 2026


---

# GPT Image 2.5 Sunburst vs Flare: Which to Pick on SceneMixer

URL: https://scenemixer.com/guides/gpt-image-2-5-sunburst-vs-flare/
Sunburst or Flare? On SceneMixer both GPT Image 2.5 tiers cost the same and draw at 2K. Here is which one to use for leads, extras, locations, states and covers.

# GPT Image 2.5 Sunburst vs Flare for character sheets, states and covers

Quick answer: On SceneMixer, GPT Image 2.5 Sunburst and GPT Image 2.5 Flare cost the same (25 credits per image, about $0.16) and both draw at about 2K, so the choice is only about the model. OpenAI built Flare as the faster, general-purpose one and Sunburst as the slower, precise one that holds a reference subject best. Use Flare for volume — supporting cast, locations, props, first drafts. Use Sunburst for the leads’ faces, for state images drawn from a main sheet, and for the promo cover, which is drawn from your leads’ sheets.

Both tiers on one page: GPT Image 2.5 on SceneMixer.

## What is the same on both

- Price. : 25 credits per image on either, from the first image — the account’s 15 free sheets are on the default model only.
- Size. About 2K, 2,048 pixels on the long side, on either.
- Where they are used. Character, location and prop sheets, regenerations, state images and the promo cover posters.
- How you switch. One setting; it applies to new images only, so you can change it between groups as often as you like.

## What is different

OpenAI released the two together on September 8, 2026 and described them differently:

- Flare — the general-purpose model, built to be fast. The faster of the two on this site as well.
- Sunburst — slower, built for precise work. OpenAI says it keeps reference subjects and earlier edits intact better than Flare.

On SceneMixer the second point is the one that matters. A state image (the same character in a hospital gown) and a cover poster are both drawn from an existing sheet. A model that holds the reference better keeps the face and build of the person you already approved.

## Which one for each job

| Job | Pick | Why |
| First pass over the whole cast | Flare | Faster; you are judging the descriptions, not the final faces |
| The two or three leads | Sunburst | Precision where the audience looks longest |
| Supporting cast and extras | Flare | Same price, quicker turnaround |
| Locations and props | Flare | No reference to hold; speed is the gain |
| State images (second outfit, injury, age) | Sunburst | Drawn from the main sheet; holding the reference is the point |
| Promo cover posters | Sunburst | Drawn from your leads’ sheets, with the title painted in |

## A cheap way to decide for your own project

Pick one lead whose face matters and generate the sheet on Flare. If the face is what you wrote, keep it and stay on Flare. If it is not, download it, switch the setting and regenerate the same sheet on Sunburst — regenerating on SceneMixer replaces the image on the card, so the download is your copy to compare against. That is two images. Then set the tier for the rest of the cast from what you saw, not from a benchmark.

If neither is clearly better for that face, the default Qwen-Image 3.0 at 7 credits per image may be enough; the free sheets are drawn there.

## How to switch between them

Settings → Model Preferences → Image Model, then pick GPT Image 2.5 Flare. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

Switching does not redraw anything. The card for each character, location and prop shows the image it has now; only the next generation uses the new tier.

## Sources

- The Rundown: OpenAI ChatGPT Images 2.5 — launch coverage of Flare and Sunburst, September 8, 2026; checked October 25, 2026.
- Kanaries: GPT Image 2.5, Flare vs Sunburst — developer write-up of the two models; checked October 25, 2026.
GPT Image is a trademark of OpenAI, named here to identify the models offered as tiers. Prices are SceneMixer’s own, read from the live price list.

| | Flare | Sunburst |
| Shown in Settings as | GPT Image 2.5 Flare | GPT Image 2.5 Sunburst |
| Credits per image | 25 | 25 |
| Speed | Faster | Slower |
| Holds a reference subject | Good | Better (OpenAI’s claim) |
| Best on this site for | Volume: extras, locations, props, drafts | Leads, state images, cover posters |

## FAQ

### Is GPT Image 2.5 Sunburst better than Flare?

For precise work that has to match a reference, OpenAI says yes. For everything else Flare is the faster choice at the same price on SceneMixer.

### Do Sunburst and Flare cost the same on SceneMixer?

Yes, 25 credits per image each.

### Which is faster?

Flare. OpenAI positions it as the fast general-purpose model; Sunburst trades speed for precision.

### Which should I use for character consistency?

Sunburst for state images and covers drawn from an existing sheet. Consistency across a series on SceneMixer comes mainly from reusing each sheet in every shot, which works on either.

### Can I use both in one project?

Yes. SceneMixer’s setting applies to new images only, so draw the leads on Sunburst, switch, and draw the rest on Flare.

### Do either of them use my free sheets?

No. The 15 free sheets per account are drawn on the default image model only.

### Is either one higher resolution?

No. On SceneMixer, both draw at about 2K, 2,048 pixels on the long side.

### Which one for a cover with the title painted in?

Sunburst, because SceneMixer draws the cover from your leads' sheets. Whichever you use, check the spelling of the title before you pick a candidate.

Updated November 3, 2026


---

# GPT Image 2.5 vs Nano Banana Pro for AI Drama Sheets on SceneMixer

URL: https://scenemixer.com/guides/gpt-image-2-5-vs-nano-banana-pro/
Both are image tiers on SceneMixer. Price per image, output size, what each is built for, and which to pick for leads, states, locations and cover posters.

# GPT Image 2.5 vs Nano Banana Pro: which to use for your series' sheets and covers

Quick answer: Both are image tiers on SceneMixer, so you can use either — or both in one project. Nano Banana Pro is the most expensive tier on the site at 46 credits per image; it keeps the most of a long, detailed English description and draws at Google’s 2K setting. GPT Image 2.5 costs 25 credits per image on either of its two models, draws at 2,048 pixels on the long side, and its Sunburst model is OpenAI’s pick for holding a reference subject. Rule of thumb: costume-heavy leads described in detail → Nano Banana Pro; state images and cover posters drawn from existing sheets → GPT Image 2.5 Sunburst; volume → GPT Image 2.5 Flare or Nano Banana 2.1.

Each family on its own page: GPT Image 2.5 and Nano Banana.

## The two side by side

| | Nano Banana Pro | GPT Image 2.5 (Sunburst / Flare) |
| Credits per image | 46 | 25 / 25 |
| Size on this site (vertical) | 1536 × 2752 | 1152 × 2048 |
| Maker | Google (Gemini 3 Pro Image) | OpenAI (released September 8, 2026) |
| Strength on this site | Detail kept from long English descriptions: many-part costumes, props with a mechanism | Sunburst: precise, holds a reference subject; Flare: fast, general |
| Free sheets | No | No |

Neither tier is covered by the free sheets: every account’s 15 free sheets are drawn on the default model, Qwen-Image 3.0.

## Pick by job

- A lead whose costume has many named parts — a period gown with a stomacher, sash and lace cuffs; a uniform with insignia — Nano Banana Pro. It is the tier that drops the fewest of those parts.
- A lead’s state images — the same person bandaged, aged, in a second outfit. These are drawn from the main sheet: GPT Image 2.5 Sunburst, which is built to hold a reference subject, at about half the price.
- The promo cover — drawn from your leads’ sheets with the title painted in: Sunburst first; Grok Imagine Image 2.0 is the other tier to try for lettering.
- Supporting cast, locations, props — neither premium tier is needed. GPT Image 2.5 Flare or Nano Banana 2.1 at 19 credits, or the default tier with the free sheets.

## Mixing them in one project

The image tier is one setting that applies to new images only, so a project can use both: draw the leads on Nano Banana Pro, switch to Sunburst for their state images and the cover, switch to Flare or the default for the rest. Nothing you already drew is redrawn when you switch.

Settings → Model Preferences → Image Model, then pick Nano Banana Pro. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

## What it costs for a typical cast

Take four leads, each with two state images, and a cover. On Nano Banana Pro for everything: 4 + 8 + 3 = 15 images at 46 credits. Split as above — leads on Nano Banana Pro, states and cover on Sunburst: 4 × 46 plus 11 × 25. The split costs much less, and the expensive tier goes where it changes what the audience sees most. If this is your first project, the first cover batch is free.

## How to compare on your own lead

Benchmarks rank single pictures; a series needs the sheet that matches your description. Pick one lead, generate the sheet on one tier, download it, switch and regenerate the same sheet on the other — regenerating replaces the image on the card, so the download is your copy for comparison. Two images will tell you which tier reads your writing better.

## All the image tiers, side by side

| Image tier | Credits per image | In short |
| Qwen-Image 3.0 (default) | 7 | The free sheets; a whole cast cheaply |
| Qwen-Image 3.0 Pro (Recommended) | 18 | Leads in close-up; Chinese descriptions |
| SeedDream 5.0 Lite | 8 | 2K locations and props on a budget |
| SeedDream 5.0 Pro | 25 | 2K, photographic skin, long descriptions read to the end |
| Nano Banana 2.1 | 19 | About 2K; Google’s faster model, the cheapest of the six new tiers |
| Nano Banana Pro | 46 | About 2K; detail kept from long descriptions; the most expensive tier |
| GPT Image 2.5 Flare | 25 | About 2K; OpenAI’s faster general-purpose model |
| GPT Image 2.5 Sunburst | 25 | About 2K; OpenAI’s slower, more precise model |
| Grok Imagine Image 2.0 | 35 | About 2K; xAI’s model, strong on lettering |
| MAI-Image-2.6 | 31 | About 1 megapixel; Microsoft’s model, strict content filter |

## Sources

- The Rundown: OpenAI ChatGPT Images 2.5 — GPT Image 2.5 Flare and Sunburst, September 8, 2026; checked October 25, 2026.
- Artlist: Nano Banana 2.1, what changed, and which Nano Banana to use — how Google’s Nano Banana models line up; checked October 25, 2026.
GPT Image is a trademark of OpenAI; Nano Banana and Gemini are trademarks of Google. They are named to identify the models offered as tiers. Prices are SceneMixer’s own, read from the live price list.

## FAQ

### Is GPT Image 2.5 better than Nano Banana Pro?

Neither is better for everything. On SceneMixer, Nano Banana Pro keeps more of a long detailed description; GPT Image 2.5 Sunburst is built to hold a reference subject and costs about half as much.

### Which is cheaper on SceneMixer?

GPT Image 2.5: 25 credits per image. Nano Banana Pro is 46, the most expensive image tier on the site.

### Which draws at a higher resolution here?

Nano Banana Pro draws more pixels (1536 by 2752 for a vertical image) than GPT Image 2.5 (1152 by 2048). Both are about 2K.

### Can I use both in the same project?

Yes. The image tier is one setting that applies to new images only; switch it between groups.

### Which is better for character consistency?

Consistency comes from reusing each sheet in every shot, on any tier. For state images drawn from a main sheet, Sunburst is the model built to hold the reference.

### What about Nano Banana 2.1?

Google's faster Nano Banana model, 19 credits per image on SceneMixer: the cheapest of the new tiers and a good fit for supporting cast and locations.

### Which one should I use for the cover poster?

Start with GPT Image 2.5 Sunburst or Grok Imagine Image 2.0, and check the spelling of the painted title before picking a candidate.

### Are the free sheets usable on either?

No. The 15 free sheets per account are drawn on the default image model only.

Updated November 3, 2026


---

# Grok Imagine Image 2.0 on SceneMixer: Price per Image and Uses

URL: https://scenemixer.com/guides/grok-imagine-image-2-tier/
xAI's Grok Imagine Image 2.0 as an image tier on SceneMixer: price per image, 2K output, prompt length, moderation, and when it is the right pick for an AI series.

# Grok Imagine Image 2.0 for character sheets and cover posters

Quick answer: On SceneMixer, Grok Imagine Image 2.0 is an image tier: it draws the character, location and prop sheets your series is matched to, their state images, and the promo cover posters, for 35 credits per image (about $0.22 at the lowest credit price). It draws at about 2K. It is xAI’s second-generation image model, released in August 2026, and coverage of it stresses lettering — which makes it a natural pick for a cover with the title painted in. xAI caps prompt length, so a very long description is shortened at a sentence end on this tier. The free sheets are not drawn here; they stay on the default model.

Every tier at a glance: which image tier for your sheets. Making the cover: promo cover posters.

## What Grok Imagine Image 2.0 is

Grok Imagine Image 2.0 is xAI’s second image model. It became the quality mode of Grok Imagine in August 2026 and reached developers through xAI’s image API shortly after. Coverage of the launch highlights sharper text rendering, region-level editing and multi-image references. The sources are at the end of this page.

On SceneMixer it is one of ten image tiers. You do not need a Grok or X account; it is billed in SceneMixer credits. It draws pictures only — this page is about the image model, not Grok Imagine’s video mode, which SceneMixer does not offer.

## What it draws here

Everything an image tier draws on SceneMixer: a full-length casting photo of each character on white, each location as an empty plate, each prop alone, a state image for every second look, and the three poster candidates of a promo cover. Video shots are made on the video tiers, which use those images as references.

## Where to switch it on

Settings → Model Preferences → Image Model, then pick Grok Imagine Image 2.0. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching on SceneMixer never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

- Settings, from the account menu.
- Model Preferences → Image Model.
- Pick Grok Imagine Image 2.0; the price per image is next to it.
- Generate or regenerate in step 2, or open Cover art in step 2 to make a cover.

## Where it fits best

- Cover posters. The model paints the title into the poster. A model known for lettering is the obvious tier to try first; check the spelling before you pick a candidate either way.
- Leads with a strong graphic look — an anime or comic-style series, a costume with lettering or insignia on it.
- Not the first pick for a large supporting cast on a budget: SceneMixer’s default tier is a fraction of the price and carries the free sheets.

## Long descriptions

xAI limits how long a prompt can be (about 8,000 characters). Character descriptions on this site can be long, and the image request also carries the site’s own framing for a sheet. When the whole request would go over, this tier shortens the description at the end of a sentence, so the last details are the ones dropped. Put identity first — face, the one anchor feature, build — and decoration last, which is good practice on every tier anyway.

## Content checks and refusals

Every description is checked by SceneMixer before anything is drawn; a flagged description is not charged. On top of that, xAI runs its own filter, and when it refuses a picture the job fails and is refunded. The site’s content rules are the same on every tier: there is no tier with looser rules. See content rules and checks.

## What it costs

This tier is not covered by the free sheets. Every account’s 15 free sheets are drawn on the default image model, Qwen-Image 3.0, only. On Grok Imagine Image 2.0 each sheet costs 35 credits from the first one (about $0.22 at the lowest credit price), and the price is printed on the button before you generate. Switch back to the default and whatever free sheets you have left are still there. A generation that is refused is not charged, and most failures are refunded automatically; a few kinds are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the description tripped the site’s check, or the model maker’s own filter refused the picture | Edit the description and generate again; nothing was charged |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The model was saturated; the site kept the job queued for a while before giving up | Generate again in a few minutes, or switch tier |
| A sheet that ignores part of the description | Too many details competing in one description | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |
| Credits held for review | A rare failure where the model may already have been paid for the attempt | Nothing to do; support checks it and releases or refunds — see credits held for review |

## All the image tiers, side by side

| Image tier | Credits per image | In short |
| Qwen-Image 3.0 (default) | 7 | The free sheets; a whole cast cheaply |
| Qwen-Image 3.0 Pro (Recommended) | 18 | Leads in close-up; Chinese descriptions |
| SeedDream 5.0 Lite | 8 | 2K locations and props on a budget |
| SeedDream 5.0 Pro | 25 | 2K, photographic skin, long descriptions read to the end |
| Nano Banana 2.1 | 19 | About 2K; Google’s faster model, the cheapest of the six new tiers |
| Nano Banana Pro | 46 | About 2K; detail kept from long descriptions; the most expensive tier |
| GPT Image 2.5 Flare | 25 | About 2K; OpenAI’s faster general-purpose model |
| GPT Image 2.5 Sunburst | 25 | About 2K; OpenAI’s slower, more precise model |
| Grok Imagine Image 2.0 | 35 | About 2K; xAI’s model, strong on lettering |
| MAI-Image-2.6 | 31 | About 1 megapixel; Microsoft’s model, strict content filter |

## Sources

- Vercel changelog: Grok Imagine Image 2.0 — API availability; checked October 25, 2026.
- Tech Dev Notes: xAI docs update, August 11, 2026 — grok-imagine-image-2.0 added to xAI’s image generation docs; checked October 25, 2026.
Grok is a trademark of xAI, named here to identify the model offered as a tier. Prices are SceneMixer’s own, read from the live price list.

| | Grok Imagine Image 2.0 on SceneMixer |
| Shown in Settings as | Grok Imagine Image 2.0 |
| Used for | Character, location, prop and state sheets; cover posters |
| Credits per image | 35 |
| Output size | About 2K |
| Very long descriptions | Shortened at a sentence end (xAI limit) |
| Free sheets | No — default model only |

## FAQ

### Can I use Grok Imagine Image 2.0 on SceneMixer?

Yes. Pick Grok Imagine Image 2.0 in Settings, Model Preferences, Image Model, and new sheets and covers are drawn on it.

### How much does Grok Imagine Image 2.0 cost here?

SceneMixer charges 35 credits per image, about $0.22 at the lowest credit price, shown on the button before you generate.

### Is Grok Imagine free on SceneMixer?

No. The 15 free sheets every account gets are drawn on the default image model; this tier is charged from the first image.

### Do I need an X or Grok subscription?

No. The tier is part of SceneMixer and billed in SceneMixer credits.

### Is Grok Imagine on SceneMixer less moderated?

No. The site's content rules apply on every tier, and xAI's own filter runs as well. A refused picture is not charged.

### Can I make Grok Imagine videos here?

No. Only the image model is offered. Shots are made on SceneMixer's video tiers using the images it draws.

### What resolution does it draw at?

About 2K on this site.

### Why did part of my long description not show up?

On this tier SceneMixer shortens a very long request at a sentence end to fit xAI's limit. Put the identity details first and the decoration last.

### Is it good for the title on a cover poster?

Lettering is one of the things coverage of this model stresses, so it is worth trying first for covers. Check the spelling before picking a candidate.

Updated November 3, 2026


---

# One-Click AI Ad or Expert Mode: Which to Choose | SceneMixer

URL: https://scenemixer.com/guides/hand-it-to-ai-or-advanced-mode/
Two cards after the reading: One-click generate at an all-in price, or Expert mode that hands you the workbench, paid per step. What each gives you and which to pick.

# One-Click Generate or Expert Mode

Quick answer: After the reading, a dialog asks how you want to make the ad. One-click generate is fully automatic at the All-in price on the button — a cap — and ready in about ten minutes; you get the finished ad and can rebuild the final cut or fix a shot afterwards. Expert mode charges a smaller price to start: AI prepares the product images and writes the ad script, then you take over in the ad workbench — the same editor dramas use — and run cast, sets, props and shots yourself, paying per step. Pick one-click for a quick, good ad; expert when you want to shape the script and every shot.

The drama editor is what expert mode opens, renamed for ads.

## The dialog

AI has read your product. How do you want to make the ad? A Product or place name field to correct, then two cards. Left: One-click generate — Fully automatic — ready in about N min, with All-in ✦ N. Right: Expert mode — The AI prepares the images and writes the ad script first; characters, scenes, props and shots are then yours to run step by step, paid per step, with ✦ N to start.

## One-click generate

Everything runs on its own: the product shot is prepared, the ad script written, the story broken down, the cast and set made, the shot list built, each shot filmed, the final cut assembled with the end card. The progress page shows each stage and an estimate. The all-in price is the most you will be charged; if the ad finishes for less, you pay less.

When it is done: Download, Make another, or open the finished ad in the editor to fix one shot (Fix this shot in the editor) and Rebuild final cut. The ad script of a one-click ad is not editable afterwards; the shots are. The progress page.

## Expert mode

Two automatic steps, then it is yours. AI prepares the images and writes the ad script; the progress page says so: when these two steps finish, you go straight to the ad workbench. There the ad is a three-step project like a drama, renamed: Ad script (read it, Edit it, or Rewrite with AI), Cast & set (generate or upload the people, places and props), then Shots in the editor, where you write and shoot each shot and Assemble the ad with its end card.

Every step after the handoff is priced on its button, exactly as in a drama. SceneMixer has no all-in cap; you spend what you generate. The script is editable until it has been broken down into cast and shots; after that, change the shots instead.

## Which to pick

One-click when you want an ad today, do not want to write, and are happy to fix a shot or two afterwards. Expert when the script matters — a specific claim, a specific order of shots, a particular cast — or you want to reuse a cast you already have. You cannot switch a one-click ad into expert mode or back; each is a fresh ad.

## Cancelling

Cancel this ad works before shooting starts. The confirmation says it: credits spent on steps that have started are not refunded, and a cancelled ad cannot be resumed. In expert mode, cancelling the two preparation steps also means the images and script cannot be used further. Once shooting has started, let it finish and decide then.

## Language

Both modes show which language the dialogue will be in (Dialogue: …), and for some languages a note that voices may occasionally mispronounce a word at the current quality setting. The end card follows the same language. The end card.

| | One-click generate | Expert mode |
| Who writes the script | AI | AI first draft; you edit or rewrite |
| Who runs cast, sets, shots | AI | You, in the workbench |
| Price | All-in cap on the button | A start price, then per step |
| Time | About ten minutes | As long as you take |
| Fix a shot afterwards | Yes, in the editor | Yes — it is your editor |
| Edit the script afterwards | No | Until it is broken down |

## FAQ

### Is the all-in price what I will pay?

It is the most you will pay for a one-click ad on SceneMixer. If it finishes for less, you pay less.

### Can I edit the script in one-click mode?

No. You can fix individual shots in the editor afterwards and rebuild the final cut.

### What is the ad workbench?

The same three-step project and editor a drama uses, with ad wording: Ad script, Cast & set, Shots.

### Can I switch modes halfway?

No. Each is a separate ad; start another with the other mode.

### What happens to credits if I cancel?

Steps that have started are not refunded; steps that have not are not charged. The confirmation says so.

Updated November 3, 2026


---

# How Hard Is It to Make an AI Movie? Skills, Time and What Goes Wrong

URL: https://scenemixer.com/guides/how-hard-is-it-to-make-an-ai-movie/
How hard is it to make an AI movie? The tool is easy; the decisions are not. What you need, what takes time, six things that go wrong, and a small first film.

# How hard is it to make an AI movie?

Quick answer: Making an AI movie is easy to start and hard to finish well. The tool needs no filmmaking skill: you type or paste a story, approve drawings and read a storyboard written in plain text. What is hard is the part the tool cannot do for you: a story worth watching, a cast small enough to stay steady, and the patience to fix shots one at a time. A first two-minute film is a weekend; a feature is a long project.

A fair first test: one page of story, two characters, one place, two minutes. If that works, scale up.

## What you need

| You need | Level | Where it helps |
| A story | One want, one turn | Everything else follows from it |
| Taste | Knowing when a shot is wrong | Reading the storyboard and the first cut |
| Patience | Fixing one shot at a time | Single-shot regeneration |
| Film vocabulary | Optional | Framing and camera move in the storyboard |
| Software skills | None | Everything runs in the browser |

## Six things that go wrong

- Too many characters. Every extra is another face to keep steady.
- A weak story hidden by pretty pictures. Read the storyboard cold.
- Fixing a face in the video. Fix it on the character sheet.
- Starting at feature length. Make two minutes first.
- Skipping the preview. The free 5-second preview shows the look before you spend.
- Ignoring rights. Real people's faces and famous characters are refused.
Why a character can drift · Spending fewer credits.

## Where the time goes

Writing and casting take the thinking time. Rendering runs on the server and you can queue many segments and close the page. Fixing takes the rest: most shots land, a few need a rewrite and a second try. A tour of the editor · Segment versions.

## Keep the first film small

Two characters, one place, 120 seconds: that is 120 seconds of video at 10 credits a second on the default tier plus 1 credit a second to join, and most of it can be tested in the free preview first. AI movie making · AI movie creator.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Keep it to two characters and one place; that is the whole difficulty setting for a first film. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Project | Difficulty | Why |
| 2-minute short, 2 characters | Easy | One place, few shots, quick to fix |
| 10-minute short, 3 characters | Moderate | More shots to review and fix |
| Season of 6 episodes | Hard to plan, easy to repeat | Planning matters; the cast is reused |
| Feature length | Long project | Hundreds of paid segments |

## FAQ

### How hard is it to make an AI movie?

The tool is easy to use; the hard parts are the story, a small steady cast and fixing shots one at a time.

### Do I need filmmaking experience?

No. Film vocabulary helps in the storyboard but is not required.

### How long does a first movie take?

A two-minute film can be a weekend: writing, casting, storyboard, then rendering on the server.

### What goes wrong most often?

Too many characters, a weak story, fixing faces in the video instead of on the sheet, and starting too long.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Can I try before paying?

Yes: SceneMixer gives you 50 credits, the first 15 setup images, a free first storyboard and a 5-second preview.

### Can I fix a bad shot?

Yes. Regenerate one shot without touching the others.

Updated November 3, 2026


---

# How Many Words for a 1-Minute Video? Script Length by Minute

URL: https://scenemixer.com/guides/how-many-words-for-a-1-minute-video/
How many words fit a 1-minute video: about 140 for narration, 100 for dialogue. A table from 15 seconds to 10 minutes, the timing formula, and script pages to minutes.

# How many words for a 1-minute video?

Quick answer: About 140 words for a minute of continuous narration, and about 100 for a scene made of short back-and-forth lines. Relaxed speech is commonly quoted at 130 to 160 words a minute. SceneMixer plans English at 3 words a second, which is 180 a minute, then adds pauses: 0.75 seconds at each sentence end, 0.3 at each comma and 0.5 for each spoken line. Silent shots lower the number again. A 10-page script is about 10 minutes by the usual page-a-minute rule.

Count the words in your script, divide narration by 140 (or a dialogue-only scene by 100) and compare with the length you want. Then paste it into a storyboard and let the planner set each shot's seconds from its line.

## The answer by minute

These figures come from the timing formula below, not from a stopwatch. Narration is sentences of about 12 words with one comma each; dialogue is lines of 6 words, each in its own shot.

| Video length | Narration (connected) | Dialogue (6-word lines) | Ceiling at 3 words a second |
| 15 seconds | About 35 words | About 25 words | 45 words |
| 30 seconds | About 70 words | About 50 words | 90 words |
| 1 minute | About 140 words | About 100 words | 180 words |
| 90 seconds | About 210 words | About 150 words | 270 words |
| 2 minutes | About 285 words | About 205 words | 360 words |
| 3 minutes | About 425 words | About 305 words | 540 words |
| 5 minutes | About 710 words | About 510 words | 900 words |
| 10 minutes | About 1,420 words | About 1,025 words | 1,800 words |
The ceiling is the bare 3 words a second with no pauses, which no real voice keeps up for long. Plan to the first two columns.

## How SceneMixer times a spoken line

When a storyboard is written, each line's shot gets the seconds its words need:

seconds = words ÷ 3 + 0.75 for each sentence end + 0.3 for each comma + 0.5 for the line

The shot length is then rounded up to the next half second and is never less than 2 seconds. An ellipsis counts as one pause, and the full stop in Mr. or Dr. does not end a sentence.

| Line | Words | Calculation | Shot |
| I thought you'd left already. | 5 | 5 ÷ 3 + 0.75 + 0.5 = 2.92 | 3 seconds |
| Wait. Don't open it. | 4 | 4 ÷ 3 + 2 × 0.75 + 0.5 = 3.33 | 3.5 seconds |
| You said the 9 o'clock train, and you lied. | 9 | 9 ÷ 3 + 0.75 + 0.3 + 0.5 = 4.55 | 5 seconds |
A segment of up to 15 seconds therefore holds at most about 45 words, and in practice 24 to 35 once the pauses are counted. Shot seconds and timing · AI dialogue generator · AI storyboard generator.

## A worked example: a 90-second explainer

You have 8 sentences of 14 words each and one comma in every sentence, spoken as one narration block.

| Part | Calculation | Seconds |
| Words | 112 words ÷ 3 | 37.3 |
| Sentence ends | 8 × 0.75 | 6.0 |
| Commas | 8 × 0.3 | 2.4 |
| The line | 0.5 | 0.5 |
| Speech in total | | 46.2 |
| Left for pictures with no voice | 90 − 46.2 | 43.8 |
If every sentence is its own line in its own shot, add 0.5 seconds for each line: about 50 seconds of speech. Half the video is pictures with no voice, which is where the shots that show what the words describe go.

## Script pages to minutes

The common rule is that a screenplay page is about a minute of screen time. It is a convention, not a measurement: dialogue-heavy pages run shorter on screen and action-heavy pages longer.

| Script pages | About | In SceneMixer |
| 1 page | 1 minute | 4 segments of 15 seconds |
| 4 pages | 4 minutes | 16 segments; the Short Drama limit of 240 seconds |
| 5 pages | 5 minutes | 20 segments; the TV Series limit of 300 seconds |
| 10 pages | 10 minutes | 40 segments of 15 seconds, or 20 of 30 |
| 30 pages | 30 minutes | 120 segments, made in sequences |
| 90 pages | 90 minutes | 360 segments, made in sequences (Film has no length limit) |
The project limit is on text, not on pages: up to 300,000 characters of English per project, about 50,000 words. On AI Create, the length menu counts characters (Mini 1,000, Brief 2,500, Short 5,000, Medium 10,000, Long 20,000), and a longer script gives more episodes, not longer ones. Choosing the script length · Script length caps · Vertical short drama script format.

## Why the number is smaller than you think

- Silent shots. Establishing shots, reactions and actions with no words take seconds the voice does not use.
- Pauses. Every sentence end and comma costs time, and a whispered or heavy line is slower than a brisk one.
- Long words and names. The count is words, but a line of long names takes longer than a line of short ones.
- Different languages. The table is for English.

Write to about 10 percent under the table: a line that fits comfortably gets a better delivery than one squeezed into its shot.

## Check it in three steps

- Read it aloud with a timer. Your own pace beats any table.
- If it runs long, cut words, not the pauses. A shorter line plays better than a faster one.
- Paste it into the storyboard and look at the seconds each shot was given. A shot much longer than its line has room for a reaction; a shot that is too short for its line rushes or cuts the line. Writing dialogue in the storyboard.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Neon Nights; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the people your lines belong to; the voices are matched to them, and each line's shot is timed to its words. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Language | Counted by | Planned at |
| English | Words | About 3 a second |
| Chinese | Characters | About 5 a second |
| Japanese | Characters | About 5 a second |
| Korean | Characters | About 5 a second |
| Every other dialogue language | Words | About 3 a second, as for English, until separately timed |

## FAQ

### How many words is a 1-minute video script?

About 140 words for continuous narration and about 100 for dialogue made of short lines. Relaxed speech is commonly quoted at 130 to 160 words a minute.

### How many words for a 30-second video?

About 70 words of narration or 50 words of dialogue.

### How many words for a 2-minute, 3-minute or 5-minute video?

About 285, 425 and 710 words of narration; about 205, 305 and 510 of dialogue.

### How many words for a 10-minute video?

About 1,420 words of narration or 1,025 of dialogue, before silent shots.

### How long is a 10-page script?

About 10 minutes by the usual page-a-minute rule, which is a convention rather than a measurement. In SceneMixer that is 40 segments of 15 seconds.

### How does SceneMixer time dialogue?

English at 3 words a second plus 0.75 seconds per sentence end, 0.3 per comma and 0.5 per line; shots are rounded up to half a second with a 2-second floor.

### Is it the same in other languages?

Chinese, Japanese and Korean are counted by characters at about 5 a second. Every other dialogue language is planned at about 3 words a second until it is timed separately.

### Does music change the word count?

No. SceneMixer mixes the score under the voices, and it takes no time of its own.

### Is there a free way to test?

Editing the storyboard on SceneMixer is free, and the first storyboard on your first drama and a 5-second preview are free.

Updated November 3, 2026


---

# How Much Does an AI Micro Drama Cost? A Worked Budget | SceneMixer

URL: https://scenemixer.com/guides/how-much-does-it-cost-to-make-a-micro-drama/
A worked budget for a ten-episode AI micro drama: what the script read, sheets, storyboards, video and assembly cost, which line dominates, and what is free.

# How Much a Micro Drama Costs to Make

Quick answer: Take a 30,000-character web novel that parses into 10 episodes of about 90 seconds each. Reading the script costs about 36 credits, the ten storyboards about 90, the video at the entry tier about 9,000, and assembling the ten episodes about 900: roughly 10,026 credits in all, about $62 at the lowest credit price. Video is most of it; the rest is small. A first drama gets its sheets and its first storyboard free, and every retake is priced only for the segment you redo.

Where credits go in one episode is the same budget at episode scale.

## The four lines

Reading the script: once per project, 4 credits per 1,000 characters (a 3,333-character English page counts as 1,000). It produces the summary, the episode split and the cast, set and prop descriptions. Sheets: 15 free per account, then 7 credits each on the default tier. Storyboards: 1 credit per 100 characters of the episode’s share of the script, once per episode; on a first drama the opening 1 episode(s) come free. Video: per second, from 10 credits on Wan 3.0 · 480P up to the higher tiers. Assembly: 1 credit per second of the episode.

## The worked example

30,000 characters, 10 episodes, 90 seconds each, video at the entry tier, no retakes:

## What moves the number

Total seconds: video is priced per second, so ten episodes of ninety seconds cost twice what ten of forty-five do. Shorter episodes are the cheapest lever. The tier: 720P and 1080P cost more per second than 480P; pick per segment, so a close-up can be sharp while wide shots stay on the entry tier. Which tier. Retakes: every regenerated segment is charged again, only for its own seconds. Budget one retake in three. Sheets beyond the free ones: a cast of thirty is fifteen paid sheets.

## What is free

Your first drama: 15 sheets across the account and the first episode’s storyboard. A new SceneMixer account starts with 50 credits. Editing text, the summary, the shot descriptions, voices: free. Watching, downloading and sharing a finished episode: free. What you get for free.

## Ways to pay less per credit

Bigger packs carry a bonus of up to 15%. A paid plan grants credits monthly and takes a percentage off packs. Credits cost between $0.0090–0.0096 each depending on the pack. What a paid plan changes.

| Line | Credits |
| Reading the script | 36 |
| Sheets | Free within 15, then 7 each |
| Storyboards (10 episodes) | 90 |
| Video (10 × 90 s, entry tier) | 9,000 |
| Assembly | 900 |
| Total | 10,026, about $62 |

## FAQ

### What is the biggest cost?

Video. It is priced per second, so total seconds and the tier decide most of the budget.

### Is the script read charged per episode?

No. Once per project, by length. Storyboards are per episode.

### Does a retake cost the whole episode?

No. Only the segment you regenerate, for its own seconds.

### What is free on a first drama?

The sheets within the account allowance and the first episode's storyboard. Editing is always free.

### Can I mix tiers?

Yes. The tier is chosen per segment; sharp close-ups and cheap wide shots can sit in one episode.

Updated November 3, 2026


---

# How to Get Help: Ask, Feedback, Replies, Support | SceneMixer

URL: https://scenemixer.com/guides/how-to-get-help/
The four ways to get unstuck: the Ask button that answers from these guides, Feedback with screenshots and a reply by email, the bell where replies land, and support.

# How to Get Help

Quick answer: For a how-to question, press Ask in the corner: it answers from these guides and links the page it used. Feedback is sent from inside that window: when the answer does not solve it, or after two answers, a line Still not solved? appears with Send feedback. Pick the category, attach screenshots, and choose how to be told — here, or also by email; your questions and the answers go with it, so we start where the AI stopped. Replies appear under the bell as Official Reply and under History in the feedback window, usually within one business day. For anything about your account or a payment, the contact page has the support address.

Every error message answers most questions before you need to ask.

## Ask

The Ask button in the corner opens Ask SceneMixer: type a question about using the site — what something costs, what a step does, what a status means — and it answers from the guides, with From the guides links under the answer. Ask about your own credits, generations or orders and it looks them up in your account, read-only (what it can check). Common questions are one tap. It only knows the site; it will say so if you ask about something else, and it says Our guides don’t spell this one out rather than guess. Do not type personal details into it.

## Feedback

Reached from inside the Ask window: after an answer that does not solve it, or after two answers, the line Still not solved? → Send feedback opens the form (if Ask AI is unavailable, the top icon opens the form directly). Your last question is filled in and the conversation goes along. What is it about? — Something broke, Billing & credits, Suggestion, Other; the hint in the box tells you what to include for each (which step, what you saw, what you expected; which order or task, when, how much). Add screenshots: paste or drop, JPG or PNG, up to the number shown. Tick Include the page I’m on so the problem is found faster. How should we let you know?: just show the reply here, email the account address, or email another address. Submit.

## Where replies land

Under the bell at the top right, as Official Reply; also under History in the feedback window, with the status Pending, Replied or Resolved. If you chose email, a copy goes there too. Still not solved? Tell us more continues the same thread.

## Support by email

Account questions, payments, refunds and anything with personal details: write to the support address on the contact page, from the address on your account. Include the order number where there is one; for a generation, name the project, episode and segment. Payment problems.

## Before you ask

Most questions are one of these: a greyed-out button, a task status, a failed generation, a credit line. The tour on each page can be replayed from Settings → Quick tour.

| You want | Use | Reply |
| A how-to answer now | Ask, in the corner | Immediate, from the guides |
| To report a bug or a charge | Feedback, with screenshots | Bell and History, usually within a business day |
| To suggest something | Feedback, Suggestion | Same |
| Account, payment, refund | Support email on the contact page | By email |

## FAQ

### Where is the help button?

Ask, in the corner of every page. It answers from these guides.

### How do I report a bug?

Ask first. When the answer does not solve it, or after two answers, Still not solved? → Send feedback: Something broke, add a screenshot, Submit.

### Where do replies show up?

Under the bell as Official Reply, and under History in the feedback window; by email too if you chose it.

### How long does a reply take?

Usually within one business day.

### Where do I write about a payment?

The support address on the contact page, with the order number.

Updated October 13, 2026


---

# How to Make a Short Movie With AI: A Two to Ten Minute Plan

URL: https://scenemixer.com/guides/how-to-make-a-short-movie-with-ai/
Make a short movie with AI in a weekend: one page of script, a cast of two or three, the free storyboard, then render and join. A plan with segment counts and costs.

# How to make a short movie with AI

Quick answer: A short movie is a two-to-ten-minute film with one place, two or three people and one turn. In SceneMixer you write or generate one page, set the type to Film, approve the cast, read the free storyboard, then render and join the segments. A 5-minute short is 10 to 20 segments, about 300 seconds of video at 10 credits a second on the default tier. The decisions are yours; the first cut can be done in a day.

Brief for AI Create: A four-minute short: two strangers share a laundromat at 2 a.m., one is leaving the city tomorrow; one room, one washing machine, one line at the end.

## Day one: the page and the cast

- Write the ending first, then the want that leads to it.
- Keep one place and two or three people.
- Read the cast and places the script finds; merge a nickname that is the same person.
- Approve the drawings. The first 15 setup images are free.

## Day one, later: the storyboard

The storyboard is plain text: numbered shots with seconds, framing, camera move and lines. Read it like a screenplay. Cut any shot that only repeats another; add a close-up where a line lands. Editing is free. What you can change · How to write one shot.

## Day two: render and fix

- Watch the free 5-second preview.
- Render the segments; queue several at once and let them run.
- Fix single shots rather than whole scenes.
- Join at 1 credit a second, watch the whole film once with the sound on.

## Where to send it

Festivals publish their own rules on AI-assisted work, runtime and labels; check them before you submit. Online, 16:9 for a channel and a 9:16 cut for Shorts. A short for a festival · AI short movie maker · Free AI short film maker.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the two strangers and the laundromat; nothing else is needed. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Length | Seconds of video | Segments (15 s / 30 s) |
| 2 minutes | 120 | 8 / 4 |
| 4 minutes | 240 | 16 / 8 |
| 5 minutes | 300 | 20 / 10 |
| 10 minutes | 600 | 40 / 20 |

## FAQ

### How do I make a short movie with AI?

Write one page in SceneMixer, set the type to Film, approve the cast, read the free storyboard, then render the segments and join them.

### How long does it take?

The writing, cast and storyboard can be done in a day; rendering runs on SceneMixer’s server and you can queue segments together.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; reading and the first storyboard are free or small.

### How many characters should I use?

Two or three carry a short movie; each extra is another drawing and voice.

### Can it have dialogue?

Yes, SceneMixer speaks it in any of 15 languages.

### Can I make it vertical?

Yes. Choose 9:16 Portrait in Step 1.

### Can I enter a festival?

Many accept AI-assisted work, but rules differ; check each festival's entry rules.

Updated November 3, 2026


---

# How to Make a TV Series With AI: From Season Plan to Episode One

URL: https://scenemixer.com/guides/how-to-make-a-tv-series-with-ai/
Make a TV series with AI: plan the season arc, set the type to TV Series, draw the cast once and make episodes of up to 300 seconds in order, paying per episode.

# How to make a TV series with AI

Quick answer: Making a TV series with AI is a production schedule more than a prompt. Plan the season on one page, set the type to TV Series, let the script be read into an episode outline, draw the cast and the main places once, and make episodes in order, each up to 300 seconds. You pay only for the episodes you render, so a six-episode plan can start as a one-episode test.

Brief for AI Create: A six-episode series: three siblings inherit a failing lighthouse hotel and have one summer to save it; each episode ends on a guest who changes the plan; ends with the sale offer.

## Plan the season on one page

- The engine. What repeats every week: a guest, a case, a deadline.
- The arc. What changes from episode one to the last.
- The cast. Three to five people, each with one thing they want.
- The places. Two or three that recur.
AI mini series maker · AI TV series creator · Episode outline generator.

## TV Series is not Short Drama

Short Drama assumes a phone and a hook to close every unit, up to 240 seconds. SceneMixer's TV Series allows a slower opening and a closing beat that can be an unfinished decision, up to 300 seconds. Set it before the script is read, because the read carries the rule. Short drama, series or film.

## Make it in order

- Read the whole season outline and fix the arc before anything is drawn.
- Approve the cast and places once; they carry across the season.
- Make episode one. Read its free storyboard first (free on your first drama).
- Watch it, adjust the cast if needed, then add episode two. Add episodes or continue.

## Keep a season steady

Faces, outfits, rooms and voices are drawn or chosen once and reused, so episode six looks like episode one. Costume changes are explicit states, not new descriptions. Why a series stays consistent · Costume changes and states · The same room in every episode.

## What a season costs

Six episodes of 90 seconds are 540 seconds of video at 10 credits a second on the default tier, plus 1 credit a second to join each episode. The worked example in the full pipeline is ten 90-second episodes from a 30,000-character novel: about 10,026 credits, about $62 at the lowest credit price. What a micro-drama costs.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday, Smart Casual; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the three siblings and the lighthouse hotel's two main rooms; they are reused in every episode. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Job | Ends on |
| 1 | Meet the family and the failing hotel | The decision to stay the summer |
| 2 | First guest, first repair | It works, but costs something |
| 3 | Complication | A secret about the hotel's past |
| 4 | Midpoint | The goal changes |
| 5 | Low point | The plan fails |
| 6 | Payoff | A final image that answers episode 1 |

## FAQ

### How do I make a TV series with AI?

Plan the season, set the type to TV Series, read the script into an outline, draw the cast once and make the episodes in order.

### How long is an episode?

On SceneMixer, TV Series episodes run up to 300 seconds, built from up to 60 segments.

### Can I make the whole season at once?

You can, but it is safer to make episode one first and continue after watching it.

### Will the characters stay the same?

Yes. Each is drawn once and reused in every episode.

### How much does a season cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; the worked example is about 10,026 credits for ten 90-second episodes.

### Can I stop after a few episodes?

Yes. You pay only for the episodes you render.

### Is episode one free?

Its storyboard is free on your first drama, with a 5-second preview.

Updated November 3, 2026


---

# How to Make AI Story Videos: A Step-by-Step Guide for YouTube

URL: https://scenemixer.com/guides/how-to-make-ai-story-videos/
Make AI story videos for YouTube: write a story with one want and one turn, draw the cast once, read the storyboard, render and join the segments into 3 to 15 minutes.

# How to make AI story videos

Quick answer: An AI story video is a story told in scenes by the same characters, not a run of unrelated clips. Write or generate a story with one person who wants one thing, draw the cast once, read the storyboard before anything is rendered, then render the segments and join them. A 5-minute story is one episode of 10 to 20 segments; longer stories are a few episodes with the same cast. Choose 16:9 for YouTube and cut a vertical version later if you want one.

Brief for AI Create: A five-minute story for a YouTube channel: a night-bus driver finds a lost violin on the last run and decides whether to return it; one bus, one bridge, three scenes, a quiet ending.

## Pick a story that fits a video

- One want, one obstacle, one turn. If you can say it in a sentence, it fits.
- Two or three characters. Each extra one is another drawing and another voice.
- One to three places. A place drawn once is steady and cheap to reuse.
- Something to hear. Dialogue or a narrator; silence only where you mean it.
AI story video generator · AI long story video generator.

## Narrator or dialogue

A story told in the third person can keep its narrator as voice-over lines, or you can turn the key moments into speech. Most good story videos use both: a voice-over to cover time and place, dialogue for the turn. Both are written in SceneMixer’s storyboard and are editable. Writing dialogue so it is picked up.

## Plan the length before you render

Each segment is up to 15 or 30 seconds and an episode holds up to 60 segments. Decide how long the story is, then how many segments that is, then what it costs: seconds at 10 credits a second on the default tier, plus 1 credit a second to join. Choosing script length and cost · How length is built.

## Make a first cut of the opening

Render the first scene only and watch it with the sound on. Check three things: do the faces match the drawings, does the place look like the place, does the voice sound like the person. Fix the cast, not the shots, if the answer is no. The free preview · Edit one shot and regenerate.

## Post it

For a channel, 16:9 Landscape. For Shorts, a 9:16 Portrait cut of the best minute. Videos on a free account carry a small AI label; a paid membership turns it off by default. Posting to TikTok, YouTube Shorts and Reels · A faceless YouTube channel.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, Urban Opulence; Costume Style Soft Everyday, Smart Casual; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the two or three people and the one or two places the story needs; the voice-over narrator needs no drawing. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Story length | Segments (15 s / 30 s) | How to build it |
| 3 minutes | 12 / 6 | One episode |
| 5 minutes | 20 / 10 | One TV Series episode |
| 10 minutes | 40 / 20 | Two TV Series episodes with the same cast |
| 15 minutes | 60 / 30 | Three episodes; post as parts |

## FAQ

### How do I make an AI story video?

Write or generate a story in SceneMixer, draw the cast once, read the storyboard, then render the segments and join them into one video.

### How long can a story video be?

On SceneMixer, one episode holds up to 60 segments of up to 15 or 30 seconds; longer stories are several episodes with the same cast.

### Can it narrate the story?

Yes. SceneMixer keeps the narrator as voice-over lines in the storyboard.

### Do I need a script?

No. Prose works, or type one line in SceneMixer’s AI Create and let it draft the story.

### Can I make it for YouTube?

Yes. On SceneMixer, choose 16:9 Landscape in Step 1; 9:16 for Shorts.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Is there a free try?

New accounts start with 50 credits; the first 15 setup images, the first storyboard and a 5-second preview are free.

Updated November 3, 2026


---

# How to Make AI Videos That Do Not Look Like AI: Six Tells, Six Fixes

URL: https://scenemixer.com/guides/how-to-make-ai-videos/
Six tells that give an AI video away (a drifting face, a rearranging room, jumping clothes, random crowds, a generic look, wrong sound) and the setup that fixes each.

# How to Make AI Videos That Do Not Look Like AI

Quick answer: An AI video looks like AI when something that should stay put does not: a face, a room, an outfit, a crowd, a voice. Most of these are not random flaws in the picture. They are facts that nobody gave the model, so each shot invents its own. The fixes are therefore setup, not prompt tricks: one approved picture per character, one measured room per scene, a separate picture for anything that changes shape, a stated crowd, a look chosen on purpose and a voice tied to its speaker. This page takes the six commonest tells in turn, says why each happens, where the fix lives in the product and what it costs, and ends with a three-minute test for your first segment.

Try it: watch one finished sample episode with the sound off and note the first thing that feels wrong. Match it to one of the six tells below; that is the one to fix in your own project first.

## Tell 1: the face changes between shots

Why: if every shot is drawn on its own, nothing says who a person is, so the model draws a plausible face each time. The fix: each character is one entry with one approved image, and every shot that includes her is made from that entry. The face in a close-up is the face in a wide shot and in episode nine, because the source is the entry, not the previous clip. A bad shot does not poison the next one: regenerate it, or move on.

Your part: approve the sheets in SceneMixer’s Step 2 before you pay for any video. Describe the face with the specific things a real face has, a crooked front tooth, a broken nose, grey stubble, a scar, not with verdicts like “beautiful”, because the model resolves a verdict to an average pretty face. The sheet is drawn with a plain side light that shows the bones, so the description does the rest. Free: the first 15 sheets on an account. The same face in every shot · Describe the cause, not the result.

## Tell 2: the room rearranges itself, and people swap sides

Why: a conversation is three or four shots, and if each shot invents its room, the window moves and the man who was on the left is on the right. A reverse shot even draws a wall that was never there. The fix: every scene gets an automatic AI 3D model, the space, the furniture that matters and a place for each person, and every shot in the scene is generated against that one model. The wide shot, the close-ups and the reverse shot describe the same room from different positions. There is nothing to set up.

Your part: write positions as relations to things in the room and to each other, “at the head of the table, across from her, back to the window”, once at the start of the segment and again only when someone moves. “By the window” holds from every camera; “on the left” is true from one. Everyone stays where they sat · Write positions the world can read · Why characters swap sides.

## Tell 3: clothes and wounds jump, or never change

Why: a model given a reference picture copies what is in it, the face, the clothes, the shape of an object, and no sentence will talk it out of that. The fix: a thing that looks different needs its own picture, called a state: a full change of outfit, a transformation, a burnt hall, a blade out of its sheath. The read proposes them within caps (up to three looks for a character, two for a location, one for a prop) and works out which segment uses which. Anything that changes only the situation, such as standing or sitting, day or night, rain, dust or who holds what, is written in words and works fine. Rain in particular is text, not a state.

The test is one question: does the thing itself look different, or does the situation around it? If both versions could be in one frame, they are two assets, not one asset with a state. Costume changes and scene states.

## Tell 4: a crowd that appears, vanishes, or stares at nothing

Why: whether a shot has people in the background is a fact, and if the text never states it, the model decides shot by shot. A wedding with a full church in one shot and an empty one in the next is the classic. The fix: state the crowd as a fact once for the whole segment: who, where, doing what, such as “a dozen guests at the tables, eating and talking”. If the place is empty, say what is there instead: “an empty street, chairs stacked against the cafe wall”. A negative like “no crowd” tends to produce the crowd, because a sentence is a list of things to consider drawing. Background people in a scene · Say what is there, not what is not.

Put an @ chip on every character who is visible in a shot, and on nobody who is only heard or elsewhere in the room: an unseen name invites a face that should not be there. Who is in frame.

## Tell 5: the generic AI look

Why: the average of everything is smooth skin, perfect symmetry, even light from nowhere. The fix: choose the look on purpose and in layers. Visual style is the medium; scene style is what the set is built from; costume style is what people wear. They are kept apart because they control different layers, and mixing them into one box is what makes things generic. Then give each shot a light with a source (a desk lamp, headlights across the ceiling, overcast noon), write actions as physical causes (the train brakes and her coffee tips), and keep two or three people per shot.

Leave out the magic words. “Masterpiece, 8K, cinematic, volumetric fog” adds nothing the model can place, pushes the words that matter further down, and sometimes gets painted literally: the fog really appears. How to make AI video look more realistic · No magic words · Choosing your style settings.

## Tell 6: the sound does not belong to the picture

Why: when speech is recorded or dubbed after the picture, lips and words drift apart. The fix: SceneMixer generates each segment with its sound already in it, every tagged line spoken in that character’s voice by the person whose lips move, plus the ambience of the place. Each character has one voice, chosen once and used for every line in every episode; match one from the built-in library and it is attached to every shot that character speaks in.

Your part: write each line with its speaker and voice tag, give anyone else in the shot lips closed, add one Ambient line of a few words, and mark a silent shot as No dialogue. If lettering appears in the picture, a shot invited it by mentioning a sign, a caption or a screen with content; describe a blank screen and regenerate. Dialogue and sound in one generation · Text in the picture, and wrong dialogue.

## A three-minute test for your first segment

| Look at | Pass if | If it fails |
| The face, in the close-up and the wide shot | It is the same person, with the features you wrote | Edit the description in Step 2, regenerate the sheet |
| The room, across two cuts | The window, the door and the seats have not moved | Add one sentence of positions to the segment’s first shot |
| The clothes, against the sheet | They match, or a state is chosen for the change | Add a state, or move the change into words if it is only the situation |
| The background | People are there or absent as you stated | State the crowd, or describe the empty place |
| The light | It has a visible source | Name the lamp, the window or the hour in the shot |
| The sound, with the picture on | Right voice, right mouth, nothing written on screen | Check the speaker and voice tags, then the character’s voice |

Run this on one SceneMixer segment of the lowest tier before you generate the rest; the cost of a retake there is small and the lessons carry to every later shot. A checklist before you generate.

## What setup cannot fix

Hands, five people talking at once and fine detail in a crowd are where AI video still gives itself away, and the answer is to ask for less: fewer people, one action per shot. Two identical requests can give two different clips, so some shots will need a second take. And a face too close to a real person is refused, which is a rule, not a flaw. The honest goal is not to hide that a video was made with AI, which the badge and platform labels exist to say, but to remove the glitches that distract from the story.

| Tell | Cause | Fix in the product |
| Face changes | No fixed identity | One approved sheet per character |
| Room rearranges | Each shot invents its room | A 3D model per scene; positions written as room anchors |
| Clothes jump | Reference copies what it shows | A state picture for a changed look; words for a changed situation |
| Crowd flickers | Never stated | State who, where, doing what, or what is there instead |
| Generic look | Average of everything | Three style layers, a light source, specific features |
| Sound mismatch | Sound made apart from picture | Sound generated with the clip; speaker and voice tags |

## FAQ

### How do I make an AI video look less like AI?

Fix what should stay put: one approved image per character, one room per scene, a state picture for any change of look, a stated crowd, a chosen look and a voice tied to its speaker. Then keep shots to two or three people and one action.

### Why does my character look different in every shot?

Because each shot was made without a fixed reference. In SceneMixer each character is one entry with one approved sheet, and every shot is made from it. Check the sheet and that the character has an @ chip in the shot.

### Can I keep the same room across cuts?

Yes. SceneMixer builds each scene against an automatic AI 3D model, and you describe positions relative to things in the room, such as the window or the head of the table.

### How do I change a character's outfit mid-story?

Add a SceneMixer state for a full change of outfit, burns or a transformation. Changes of situation, such as rain, night or sitting down, are written in words.

### Do words like cinematic or 8K help?

No. The models read sentences, not tags. Say where the light comes from and what the camera sees.

### Why does a negative like no crowd give me a crowd?

A sentence is a list of things to consider drawing, and the noun carries more weight than the word no. Describe what is there instead.

### Why is there writing in my picture?

A shot mentioned a sign, a caption or a screen with content. Remove the mention and regenerate; the site asks for a clean picture and carries dialogue by sound.

### Is it dishonest to make AI video look natural?

No, if you label it. The AI badge and the platform’s own AI declaration say what it is; the aim here is fewer glitches, not concealment.

Updated November 2, 2026


---

# How to Make an AI Short Drama (2026 Guide) | SceneMixer

URL: https://scenemixer.com/guides/how-to-make-an-ai-short-drama/
Six steps from script analysis to episode compositing, a worked cost example for a 10-episode vertical series, and the mistakes that break a series.

Guide

# How to Make an AI Short Drama, Step by Step

Updated November 3, 2026

Quick answer: An AI short drama is a serialized show of 60–120 second vertical episodes whose cast, sets and footage are generated by AI. Making one takes six steps: analyze the script into episodes, lock a consistent cast with reference art, localize for your market, write and edit storyboards, generate each 4–15 second shot, then composite and publish. With an AI short drama generator such as SceneMixer the whole chain runs from one upload; a 10-episode series costs roughly 10,026 credits (about $62), and you can test the first episode on the 50 free credits every account gets.

## Before you start: three decisions

- Source material. Web novels and serialized scripts work best because they are already written in cliffhanger rhythm. 8,000–60,000 characters is a comfortable range; longer books are analyzed in stages.

- Format. Vertical 9:16, 60–120 seconds per episode, one hook per episode. This is what ReelShort, DramaBox and the Chinese platforms monetize. SceneMixer projects start at 16:9, so set 9:16 once in the project settings; keep 16:9 if you are cutting for YouTube.

- Visual style and market. Choose one rendering style (realistic, 2D, 3D CG…) and one target region — Western, East Asian, Southeast Asian or Latin American — before generating anything. Changing either later means regenerating the cast.

## The six steps

### 1. Upload and analyze the script

Paste the text or upload .txt / .docx / .md. One analysis pass extracts characters, scenes, props and an episode-by-episode outline, each with a conflict arc and an end-of-episode hook. On SceneMixer this costs 4 credits per 3,333 characters and includes the first reference images; it takes minutes.

### 2. Lock the cast

Review the generated character sheets before you make any video. Fix faces, ages and signature looks now — every later shot is generated against these references. Each account gets 15 reference images free; after that they are 7 credits each on the default image model, and uploading your own art is allowed.

### 3. Localize for the market

Pick the target region and the pipeline adapts names, casting, scenery and on-screen text. One Chinese web novel can ship as a Western-cast English series and a Southeast Asian one from the same analysis.

### 4. Write and edit the storyboards

Each episode is broken into 4–15 second shots with shot language, dialogue and duration. SceneMixer bills storyboards per episode at 1 credit per 333 characters of that episode's script; editing the text afterwards is free, and regenerating a storyboard costs the same as the first pass.

### 5. Generate the shots

Video is billed from 10 credits per second (≈ $0.06/s at the lowest credit price) at the cheapest tier (Wan 3.0 · 480P; higher tiers multiply the rate). Generate, watch, and regenerate only the shots that fail — every segment keeps a version history, and the rest of the episode stays untouched.

### 6. Composite and publish

One click merges the segments into a finished episode at 1 credit per second. Download it and publish to TikTok, YouTube Shorts, Reels or a short-drama platform; keep the AIGC label on where the platform requires disclosure.

## Worked example: a 10-episode vertical series

Assume a 30,000-character novel, ten 90-second episodes, about 3,000 characters of script per episode, default video tier, no regenerations.

| Line item | Rate | Credits |

| Script analysis (30,000 characters) | 4 per 3,333 characters | 36 |

| Storyboards (10 episodes × 3,000 characters) | 1 per 333 characters | 90 |

| Video (10 × 90 s = 900 s) | from 10 per second (cheapest tier) | 9,000 |

| Episode compositing (900 s) | 1 per second | 900 |

| Total | | ≈ 10,026 |

At the lowest credit price (largest pack or annual membership, bonus included) that is roughly $62. Budget 20–40% extra for regenerated shots on a first series. Full rates are on the pricing page; the 50 free sign-up credits on SceneMixer cover a script analysis and a first short test shot.

## Mistakes that break a series

- Generating video before the cast is locked. Every face change later costs a regeneration; fix reference sheets first.

- Too many locations per episode. Hits use 3–4 physical sets per episode; more sets means more scene art and more continuity risk.

- Long single shots. Keep shots at 4–15 seconds with one action each; models drift on long, multi-action shots.

- Crowds and choreographed fights. Current video models render two-person dialogue and single-character emotion well and large crowds badly. Write around it.

- No hook. Serialized retention lives on the last five seconds of every episode. If the outline's hook is weak, rewrite it before storyboarding.

## Frequently asked questions

### How much does it cost to make an AI short drama?

On SceneMixer, a 10-episode series of 90-second vertical episodes from a 30,000-character novel comes to roughly 10,026 credits: about 36 for script analysis, 90 for storyboards, 9,000 for video at the cheapest tier and 900 for compositing — around $62 at the lowest credit price. Regenerated shots add to that; editing storyboard text is free.

### How many episodes should a first AI short drama have?

Plan the whole arc but produce the first 3–5 episodes first. They tell you whether the cast reads consistently and whether the hooks land before you spend on 40 more. Serialized platforms expect 40–80 episodes of 60–120 seconds, so the cast library has to hold up for the full run.

### How do I keep the same face across every episode?

Use a tool that stores a persistent character identity, not a per-prompt one. SceneMixer generates a full-body reference sheet per character during analysis and injects face and body crops into every segment featuring that character, so episode 40 matches episode 1.

### Can I use my own novel or script?

Yes — paste text into SceneMixer or upload .txt, .docx or .md in Chinese or English. If you only have a premise, the built-in AI writer can draft a serialized script from a one-line idea, and you can still edit every storyboard by hand for free.

### Do I need video-editing skills?

No. The unit of editing is the storyboard script: change the text of a shot and regenerate only that segment. When all segments are ready, full-episode compositing on SceneMixer is one click, billed at 1 credit per second.

### Do I have to label the video as AI-generated?

Many platforms and jurisdictions now require AI-generated content to be disclosed. SceneMixer has an optional AIGC watermark toggle in settings; check the rules of the platform you publish on and keep the label on where it is required.


---

# How to Make an Animated Movie With AI: Look, Cast, Sequences

URL: https://scenemixer.com/guides/how-to-make-an-animated-movie-with-ai/
Make an animated movie with AI: choose a look for the whole film, draw the cast once, storyboard and render it one sequence at a time, and finish with a trailer first.

# How to make an animated movie with AI

Quick answer: An animated movie is made in sequences: choose one look for the whole film, set the type to Film, draw the cast and the places once, then storyboard, render and join one sequence at a time. Each sequence is billed and made on its own, so you watch sequence one before paying for sequence two. Make a 60 to 120 second trailer first to test the look and the cast.

Brief for AI Create: A ten-minute animated short: a clockmaker's daughter wakes the tower clock that has been silent for a hundred years; one tower, one village, three characters, no villain.

## Choose the look first, and keep it

The look is chosen in Step 1 and fixed once you continue, and every character, place and prop is drawn in it. Choose for the mood of the film:

| Look | Suits |
| Painterly Anime | Big scenery, quiet scenes, emotional endings |
| Anime | Action, school, romance |
| Family 3D Animation / 3D Toy Cartoon | Family comedy, a world of toys |
| Claymation | Gentle comedy in a small place |
| Watercolor Storybook | Fables and bedtime stories |

Visual Style: Anime

Visual Style: Family 3D Animation

Visual Style: Claymation

Visual Style: Watercolor Storybook
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Sequences, not one render

The Film type on SceneMixer has no hard length cap and splits the script into sequences. A 10-minute short is 40 segments of 15 seconds or 20 of 30; a feature is hundreds. Each sequence is its own storyboard, render and join. AI animated movie generator · How length is built.

## Trailer, then sequence one

- Write a trailer script of five to eight moments and render it with the same cast. Make a movie trailer.
- Read what the look does to your cast; change a design now, not later.
- Make sequence one; watch it with sound.
- Decide how far to take it.

## Voices and music

Each character has a voice and speaks in the video, in any of 15 languages, and each episode is joined with a background score. AI video generator with dialogue · An animated short film · A 3D cartoon movie.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, Medieval Castle; Costume Style Soft Everyday, Smart Casual; Visual Style Watercolor Storybook, Family 3D Animation; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the clockmaker's daughter, her two friends and the tower; no villain needs drawing. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Length | Segments of 15 s | Segments of 30 s | Usually |
| 2 minutes | 8 | 4 | A trailer |
| 10 minutes | 40 | 20 | An animated short |
| 30 minutes | 120 | 60 | A TV special, in sequences |

## FAQ

### How do I make an animated movie with AI?

Choose a look in SceneMixer, set the type to Film, draw the cast and places once, then storyboard, render and join it one sequence at a time.

### How long can it be?

SceneMixer's Film type has no hard length cap; each sequence holds up to 60 segments.

### Which look should I choose?

Painterly Anime for emotional films, Anime for action, Family 3D or 3D Toy Cartoon for family stories, Claymation for small gentle worlds.

### Should I make a trailer first?

Yes. It tests the look and the cast cheaply.

### Will the characters look the same throughout?

Yes. Each is drawn once in SceneMixer and used in every shot.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; a trailer is a minute or two of that.

### Do the characters speak?

Yes, in any of 15 languages.

Updated November 3, 2026


---

# How to Write a Film Script With AI: Brief, Draft, Rewrite

URL: https://scenemixer.com/guides/how-to-write-a-film-script-with-ai/
Write a film or short drama script with AI: a five-part brief, a length option, a draft you check for one place per scene and speakable lines, then a rewrite loop.

# How to write a film script with AI

Quick answer: Write the brief, not the script: the genre, a lead who wants something, what stands in the way, the ending you want and the tone, in one or two sentences. Paste it into the AI Create tab, pick a Length and press Start Writing, and you get a script of episodes, scenes and dialogue. Read the draft for three things — one place per scene, lines short enough to say in their seconds, and a turn at the end of every episode — then rewrite by sharpening the brief or by editing the text. Every draft stays in Writing History. A script written anywhere else can be pasted in instead.

Brief for AI Create: A ten-minute short drama: a night-shift pharmacist notices the same stranger buying the same cough syrup every Tuesday and follows the thread to a secret she did not want; tense, small, no villain; it ends on her choosing not to call.

## The brief is the script's spine

- Genre. It sets the rhythm and the kind of hook each episode ends on.
- Lead and want. One person, with something they are after.
- Conflict. The person or force in the way; a story with no opponent drifts.
- Ending. Even one word (she wins, they never meet again) makes the episodes point somewhere.
- Tone. Cold, warm, absurd, slow-burn.
Before: A script about a pharmacist. After: the brief above. The second one gives the writer a lead, a want, an obstacle, an ending and a mood; the first gives it nothing to push against. Writing the one-line brief · SceneMixer’s Inspire Me fills the box with an example to edit.

## Pick the length for the film you mean

| Option | Script length | Gives |
| Mini | 1,000 | One scene, a test |
| Brief | 2,500 | One or two short episodes |
| Short | 5,000 | A handful of episodes; the usual first choice |
| Medium | 10,000 | A short season |
| Long | 20,000 | A full series |

A longer script gives more episodes, not longer ones: the read splits it into episodes of a minute or two. In English the same option is more characters (3,333 characters of English count as 1,000), and the price on each option already allows for it. Choosing the script length and cost · Why a length is missing from the menu.

## Read the draft four ways

- Places. Each scene should happen in one place. A segment stays in one place and one timeline, so a scene that hops between the pharmacy, a car and a flat is three scenes. One place, one timeline.
- Voices. Lines have to be said in their seconds, and English runs about three words a second: a 15-second segment carries about 45 spoken words at most, a 30-second one about 90, with nobody doing anything else. Cut speeches into short lines. Shot seconds and timing.
- Turns. The last beat of every episode should change something: a decision, a reveal, a door that opens. For a short drama that is the hook; a TV series can end on a decision left open. Hooks and cliffhangers.
- Thoughts. A thought nobody says aloud and nobody shows cannot be filmed. Turn it into a line, a look or an object, or cut it.

## Rewrite the brief, not the draft

If the draft is off, write again with a sharper brief. Each writing on SceneMixer is priced on its own, and every draft stays in Writing History, so you can go back to an earlier one. You can also take a draft into a project anyway: the summary, the cast and the storyboards are editable, and text edits are free. After the script is written.

| If the draft… | Say this in the next brief | Why it helps |
| wanders | Name the ending, even in one phrase | The episodes then point at it |
| has everyone talking in paragraphs | Name a tone that speaks in short lines (clipped, cold and fast) | Lines have to fit their seconds |
| is too long or too short | Name the number of episodes and the setting and era | Both are the optional parts of a brief |
| has no one to push against | Add the person or force in the way | A story with no opponent drifts |

## Already have a script? Bring it

Write it in any tool, or by hand, and paste it in or upload a .txt, .docx or .md file on Upload Script. Reading in SceneMixer costs 4 credits per 1,000 characters, English counted lighter, and a project can hold up to 300,000 characters of English. Screenplay layout (scene headings, character cues, dialogue) is read most reliably, but prose works. Upload a script and read it · A screenplay template · AI film script generator.

## Short drama, TV series or film

The type you choose in step 1 changes how the script is read: a Short Drama episode runs up to 240 seconds and is built around a hook, a TV Series episode up to 300 seconds, and Film has no length cap and is read in sequences. Choose it before you read the script, because the read carries the rule. Short drama, series or film · Vertical short drama script format.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence; Costume Style Smart Casual; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the pharmacist and the stranger, and the pharmacy counter as the one place; every shot is made from them. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Stage | You do | SceneMixer does |
| Brief | Five things in one or two sentences | Nothing yet |
| Draft | Pick a Length, press Start Writing | Writes episodes, scenes and dialogue; the draft is kept in Writing History |
| Check | Read it for places, voices, turns, thoughts | Nothing; text edits are free |
| Read | Press Start Creating | Splits it into episodes and draws up the cast, places and props |

## FAQ

### Can AI write a whole film script?

SceneMixer writes up to the Long option, a full series' worth of episodes, in one go. For a longer film, write it in parts or paste your own: a project can hold up to 300,000 characters of English.

### Which AI is best for writing a short film script?

SceneMixer does not rank writers. What decides the result is the brief and the rewrite loop, and the test is whether the draft can be filmed: one place per scene, lines that fit their seconds, a turn at the end.

### Does it write in screenplay format?

SceneMixer writes episodes, scenes and dialogue. You can paste a script in screenplay layout too; the scene headings and character cues are read most reliably.

### What does writing cost?

Each Length option shows its price, and the Start Writing button shows the one you picked. In SceneMixer, only a delivered script is charged, and a writing that fails is refunded.

### Can I edit the draft?

Yes. The summary, cast and storyboards on SceneMixer are editable and text edits are free.

### Can I write in another language?

SceneMixer writes the script in the language you use the site in, and the characters can speak another. Writing in your language.

### How do I write a micro drama or short drama?

Set the type to Short Drama in SceneMixer's step 1 and write for episodes of up to 240 seconds, each ending on a turn. AI micro drama.

### Is the first try free?

New SceneMixer accounts start with 50 credits, and there is a 5-second preview. Writing and reading are charged normally.

Updated November 3, 2026


---

# How to Write One Shot in an AI Storyboard | SceneMixer

URL: https://scenemixer.com/guides/how-to-write-one-shot/
The shape of a good shot: marker and seconds, framing, camera, who is visible, one thing happening, the line, the ambience. With examples you can copy.

# How to Write One Shot

Quick answer: A shot is one marker chip with its seconds, then a few sentences that say what the camera sees: the framing (close-up, medium, wide), whether the camera moves, who is in the picture (named with @ chips), the one thing that happens, and the line if there is one. Write what a viewer would see and hear, in the order they would notice it. Treat each new shot as a cut to a new framing. Keep to one action per shot, and give it enough seconds to happen.

Newer storyboards. In storyboards written or regenerated since late October 2026 a shot has no In frame only or Ambient line: write what the camera sees, what happens and who speaks. The system works out who is in frame (shown under the segment as In frame) and the sound of the place. Notes under a segment.

Open a finished episode and read three shots in a row. Every one of them follows the shape below.

## The shape

[Shot 3] 4s: Medium shot, fixed camera. @Mara at the head of the table, @Jules on her right. She slides the @Sealed letter across to him and holds his eye. @Mara (low, steady): Read it before you say no. (voice: 🔊Mara) @Jules lips closed, no lip movement. Rain on the window behind them.

Marker and seconds. Framing and camera. Who is where. What happens. The line, with its delivery note and voice chip. What the other person is doing. What we hear. That order is the one a viewer notices things in, and it is the order that reads best.

## Framing and camera

Say the size of the picture: close-up, medium close-up, medium shot, full shot, wide shot. Say whether the camera is still or moving: fixed, handheld, tracking, a slow push in. Say the angle if it matters: over @Jules’s shoulder, from the doorway, from the passenger window. One framing per shot; a change of framing is a new shot.

You do not have to write the camera position at all. Say who faces whom and the scene’s automatic AI 3D staging puts the camera where that is true. How positions hold.

## Who is in the picture

Name everyone visible with an @ chip. For one person, add In frame only: @Mara. Do not name someone who is only heard or only nearby. A listener gets lips closed, no lip movement. Who is in frame.

## One thing happens

A shot is one beat: she slides the letter; he picks it up; she looks away. Two beats are two shots. Write what is visible — hands, faces, bodies, objects — not what someone is thinking. “She hesitates, then slides the letter across” works; “she decides to trust him” gives the shot nothing to draw.

Start the first shot with Opening: and the first frame; start each later shot with the cut, such as Cut to close-up:. When the same people are in two shots in a row, change the framing or the angle between them, or the join looks like a skip. Say posture when it is not standing: kneeling, seated, leaning on the rail.

## The line

@Name (how it is said): the words. Then the voice chip from the @ menu. Delivery notes are short and physical — low, steady; through her teeth; calling from the stairs. A line from outside the picture starts its note with (voice-over) or (off-screen). Dialogue lines.

Give the line its seconds: about three words a second in English. Shot seconds.

## What we hear

End the shot with the sounds of the place if they matter: Ambient: rain on the glass, a clock in the hall. A shot with no speech gets No dialogue so nobody is made to talk.

## Three shots to copy

A reaction. [Shot 4] 3s: Close-up. In frame only: @Jules. He looks down at the letter, jaw tightening. No dialogue. Ambient: rain.

A line with movement. [Shot 5] 6s: Medium shot, handheld. @Mara stands and crosses to the window, back to him. @Mara (quiet, not turning): I already know what it says. (voice: 🔊Mara)

A reveal. [Shot 6] 4s: Slow push in on the @Sealed letter on the table; the seal is broken. No dialogue. Ambient: the clock.

| Part of the shot | Example |
| Marker and seconds | [Shot 3] 4s: |
| Framing and camera | Medium shot, fixed camera / Close-up, slow push in |
| Who is where | @Mara at the head of the table, @Jules on her right |
| One thing that happens | She slides the letter across and holds his eye |
| The line | @Mara (low, steady): Read it before you say no. (voice: 🔊Mara) |
| The listener | @Jules lips closed, no lip movement |
| Sound | Ambient: rain on the window |

## FAQ

### How long should a shot be?

Two or three seconds for a look or a gesture, four to six for a line and its reaction, longer for movement. The shots add up to the segment.

### Do I have to write the camera position?

No. Name who faces whom and SceneMixer places the camera. Write a position when you want a specific one.

### Can a shot have two actions?

Better not. Split it into two shots; each gets its own seconds.

### What if nobody speaks?

Write No dialogue at the end of the shot.

### Should I describe what a character is thinking?

Describe what shows: the face, the hands, the pause. A thought with no visible sign gives the shot nothing to draw.

Updated November 3, 2026


---

# Image to Video: Prepare Your Pictures Before You Upload

URL: https://scenemixer.com/guides/image-to-video/
What a picture needs before it goes into a video: formats, the 4,096 px and 10 MB limits, how many photos, which shapes work and what is refused. A short checklist.

# Image to Video: Getting Your Pictures Ready

Quick answer: Before a picture can become part of a video it has to get through three gates: your browser, the upload and the video model. Your browser reads almost any common photo format and turns it into a JPEG whose long side is at most 4,096 pixels. An ad takes up to four photos, each under 10 MB and 16 MB in all. The model then wants a sensible shape, roughly 9:16 for a person and 16:9 for a place. This guide lists each limit, what is refused and why, and ends with a checklist you can run in two minutes. What the video does with the picture afterwards is covered in AI image to video.

## Three doors a picture can go through

There is no single “add image” box. A picture goes in through one of three doors, and each has its own count, size limit and plan requirement.

| Door | Where | How many | Size and format | Plan |
| Photos for an ad | Home page, Ad Video tab, Add images | Up to 4, all of one subject | Under 10 MB each once converted, 16 MB in all | Any account. Reading the photos is free |
| A character, place or prop picture | Step 2, Upload & Replace on the card | One per card | JPG or PNG; about 9:16 for a person, 16:9 for a place or a prop | Paid plan |
| The asset library | Assets, Upload asset | As many as your storage allows | JPG or PNG, up to 20 MB each | Paid plan; counts against the plan’s storage |
The one door SceneMixer does not have is a single still that becomes the first frame of a clip. Pictures here are references for who and where, not frames to animate, which AI image to video: four methods explains. Everything below is about getting a picture through whichever of the three doors you need.

## What your browser does to every picture

Whatever you pick, the same thing happens before a single byte is sent. The browser opens the file, applies the camera’s rotation so an upright phone photo stays upright, paints any transparent area white, scales the picture down if its longer side is above 4,096 pixels, and saves it again as a high-quality JPEG. There is one path and no shortcut, even for a file that was already a JPEG. Renaming photo.webp to photo.jpg gets you nothing, because the site never trusts the name: it works from the picture the browser has just encoded.

The reason is practical. The image and video services further down the line accept JPEG and PNG only, so converting at the door means a file is never accepted at upload and then found unusable three steps later. The price is one extra round of JPEG compression, which is invisible at the sizes involved.

| What you pick | What happens |
| JPEG | Opened and saved again as a JPEG. Size may change a little. |
| PNG | Transparent areas are painted white and the result is saved as a JPEG. A cut-out product on a clear background arrives on white. |
| WebP | Any browser that can show it can convert it, so it works. |
| HEIC / HEIF (iPhone) | Works in Safari, which can open it. Most other browsers cannot, and then the file is refused the moment you pick it. |
| SVG, camera RAW | Refused when you pick them. The browser cannot turn them into a photo. |
| A PDF or a document | Refused as not an image. Export the page you want as a picture first. |
| A picture still over 20 MB after conversion | Refused as too large. Compress it or crop it and pick it again. |
SceneMixer shows refusals on the spot, before anything is uploaded or charged. What you can feed in lists the other kinds of input, such as scripts and video.

## The limits, in numbers

| Limit | Number | Applies to | What happens past it |
| Long side | 4,096 px | Every picture, in your browser | Scaled down before upload |
| File size after conversion | 20 MB | Every upload | Refused as too large |
| Photos in one ad | 4 | Ad Video tab | “You can use up to 4 images”; the extras are dropped |
| Size of one ad photo | 10 MB | Ad Video tab | “That image is too large — use a JPG or PNG under 10 MB” |
| Total size of the ad photos | 16 MB | Ad Video tab, all photos together | Refused as too large in total |
| Width ÷ height of an app or website screenshot | From 0.4 to 2.5 (2:5 tall to 5:2 wide) | Ad Video tab, screenshots only | “This screenshot is too tall or too wide — crop it” |
| Short side of a reference picture | 512 px | A picture handed to the video model | Scaled up to 512 on the short side (adds no detail) |
| Long side of a big reference picture | 2,048 px | Reference pictures over 2 MB | Scaled down before the video model sees them |
The last two rows are quiet repairs rather than rules. A small picture is enlarged so that a segment is not refused for being tiny, and a very large one is reduced because pixels beyond what the finished video can show are wasted. Neither will rescue an extreme shape: pictures much wider than about 4 to 1, or much taller, are not enlarged and are likely to be refused at generation with an “invalid input” message. Invalid input: what to check walks through that message.

## Which pictures work, and how to prepare each

| Your picture | It works as | Prepare it like this |
| A product on a plain background | The hero of an ad | Fill most of the frame, in focus, and put it first: the first photo is the featured subject and is also the picture on the end card |
| A dish, drink, garment, car or room | An ad photo | Same: clear, in frame, in focus. Extra photos show another angle, a detail or the place it lives |
| A shopfront or restaurant, customers included | An ad photo | Leave the people in. They are removed and the place is kept, and you will see a note saying so |
| An app or website screenshot | An ad photo, used as it is | One screen only. Crop a full-page capture to a phone or computer shape. A screenshot with a real face in it is refused |
| A selfie or snapshot of a person | A character picture | One person, face visible and turned toward the camera, even light. Waist-up usually beats a distant full-length shot. Avoid group photos, heavy filters, strong coloured light and sunglasses |
| Artwork you drew as a character sheet | A character picture, Use this image as is | Full figure, plain background, about 9:16 |
| A photo of a real place | A scene picture | Show the space as a whole. If it has people or a strong time of day in it, choose the optimise option so those do not follow you into every shot |
| A celebrity, a franchise character or a brand mascot | Nothing | The IP check refuses them when a clip is made, so there is no point preparing the picture |
SceneMixer refuses a blurry, tiny or ambiguous photo at the reading step with “We couldn’t work out what is being advertised here”, and nothing is charged for a refused picture. Which photos work for an ad · Using your own photo.

## One thing to remember about a snapshot

On a character, place or prop card the picture becomes the reference for every shot that shows it, which is why Optimize into a reference sheet exists. A snapshot brings its background, its lighting and its pose with it, and the video model will treat all of those as part of the character. The optimise option redraws the face, hair and clothing from your photo in the project’s own look on a clean background; Use this image as is keeps the photo untouched and is for artwork already made to be a sheet. Clips generated before you uploaded keep the old look until you regenerate them. Your own reference images: what changes.

The first time you upload anything, a window titled Before you upload asks you to confirm that you have the right to use the material and that anyone whose face or voice is in it agreed. You confirm once.

## A two-minute checklist

- Pick the door. A promo for one thing is the ad door. A person, a room or an object that must look the same in every shot is a card in step 2.
- Crop to the subject. For a screenshot, crop to a single screen; keep its width-to-height between 2:5 and 5:2.
- Order the photos. The hero goes first. Four is the most; one clear photo is enough.
- Do not resize it yourself. Anything up to a 4,096 pixel long side is fine, and the browser handles the rest. Do keep each ad photo comfortably under 10 MB.
- Clear the picture of things you do not want in every shot, such as a busy background or a colour cast.
- Check you may use it, and that anyone recognisable in it agreed.
- Write the words that go with it. A card picture has no description of its own, and the description is what the shot text reads, so use AI Describe and correct the result.
- Pick the file and read the message. If it is going to be refused, it happens now, for free.

## What the picture becomes

Once a picture is in SceneMixer, it is a reference: it tells the model what a person, a room, an object or a product looks like, and each shot is a new picture composed from the shot text. Your subject appears in new shots, not your photo set in motion. For the ad route that means photo to video AI, up to four photos into a 15, 30 or 60 second spot. The wider routes, and what a free account can do with a picture, are in Image to video AI.

## FAQ

### What image formats can I use for image to video?

Any format your browser can open, which in practice means JPEG, PNG and WebP, plus HEIC in Safari. The browser converts everything to a JPEG before it is sent. SVG and camera RAW are refused.

### What is the maximum image size?

On SceneMixer, the long side is reduced to 4,096 pixels in your browser, and a converted file over 20 MB is refused. For an ad, each photo must be under 10 MB and the photos together under 16 MB.

### How many pictures can I use?

A SceneMixer ad takes up to four photos of one subject. A character, place or prop card takes one picture. The asset library takes as many uploads as your storage allows.

### Will the video start from my exact picture?

No. SceneMixer treats the picture as a reference for who or what appears. Each shot is composed from the shot text, so you get your subject in a new shot.

### Why was my image refused?

The usual reasons are a format the browser cannot open, a file still over 20 MB, an ad photo over 10 MB, a screenshot with a real face or an extreme shape, or a picture the reading cannot make out.

### Does a small picture hurt the result?

A picture with a short side under 512 pixels is scaled up so it can be used, but scaling adds no detail, so a sharper source gives a sharper reference.

### Do I need a paid plan to upload a picture?

Not for the photos of an ad. On SceneMixer, putting your own picture on a character, place or prop card and uploading to the asset library are paid-plan features.

### Can I use a photo of a person?

Use only pictures you have the right to use. A photo of a real person may be flagged, and recognisable public figures and franchise characters are refused when a clip is made.

Updated November 2, 2026


---

# Image to Video AI Free: Where Free Stops for a Picture

URL: https://scenemixer.com/guides/image-to-video-ai-free/
How far a picture goes for free: sign-up credits, a free reading of your photos, and the free allowances an ad does not use. Step by step, with the first charge marked.

# Image to Video AI Free: Where the First Charge Appears

Quick answer: A picture goes a good way for free, but only as far as being read. You sign up, add up to four photos and the AI tells you what it sees, and nothing is charged until you choose how to make the ad. After that the ad costs credits, and an ad does not draw on the free allowances that a script project gets: not the 15 free setup images, not the free storyboard, not the free 5 second preview. A new account starts with 50 credits, and the price of each choice is on its button before you press it. This guide walks the picture path and marks the first rung that costs.

## The picture path, rung by rung

| Rung | What happens | What it costs |
| 1. Sign up | The Ad Video tab needs an account | Free. A new account starts with 50 credits |
| 2. Add photos | Up to four photos of one subject; your browser converts them | Free |
| 3. Reading | The composer says Reading your image…, then shows what the AI understood: a category and a name you can correct | Free. A photo that cannot be made out is refused and costs nothing |
| 4. Choose | Ad vibe, length, format, quality, and one line about what it is | Free to browse. The quality menu shows the all-in price beside each option |
| 5a. One-click generate | Everything runs on its own, about ten minutes | A single all-in price on the button. It is a cap; each stage charges its actual price |
| 5b. Expert mode | The AI prepares the product shot and writes the ad script, then you run the steps | A smaller price to start, then each step on its own button |
| 6. Fix a shot, Rebuild final cut | Change one shot, or re-join the shots with a new end card | Paid per action, the shot by its seconds and the join by 1 credit a second |
| 7. Download, share | The finished file, the link | Free. A free account’s video carries the small AI badge |
Rung 3 is the last one that costs nothing on SceneMixer, and it is more useful than it sounds: you find out whether the photos can be used, what category the AI thinks your subject is and whether it noticed people to remove, all before a credit moves. If the reading has misunderstood the subject, Pick different images and try again. Which photos work for an ad.

## The free allowances an ad does not use

New accounts get a handful of free things, and it is easy to assume they apply to the photo route too. They do not, and the reasons are specific.

| Free thing | On a script project | On an ad |
| Welcome credits | Yes | Yes. They are ordinary credits |
| 15 free setup images | Yes, once the script has been read and paid for | No. An ad pays no script-read fee, and the free images can only be spent on a project whose script has been read and paid for |
| Storyboard for episode 1 of your first drama | Yes | No. Ads are never the free first project |
| One free cover batch | Yes, on the first drama | No, same conditions as the free images |
| One free 5 second preview | Yes, once per account, in a parsed episode | No. Ad projects are left out of the preview offer |
| Reading the photos | Not applicable | Free, always |
The same table is the reason “free image to video” can mean two different things. For an ad on SceneMixer, free means the reading. For a story, free goes much further: the cast, a storyboard and a preview shot. What you get for free lists every allowance and its conditions.

## What lowers the price once you do pay

Only some of the choices on the composer move the price, and knowing which ones lets you start cheap without wasting a try.

| Choice | Changes the price? | Cheapest way to use it |
| Length (15, 30 or 60 seconds) | Yes | 15 seconds is one segment and one idea. 30 is two segments, 60 is four |
| Quality (480P, 720P or 1080P) | Yes, and the menu shows each price | 480P is for drafts and small screens; most people ship 720P |
| Ad vibe | No | Change it freely. It sets the look and the colours of the end card |
| Format (landscape or portrait) | No | Pick the shape you will actually post in |
| The one line | No | Write it. It saves a misread subject |
| Mode | Both are capped or itemised | Expert mode asks for less up front, then charges each step on its button |
Three further things keep the bill down. The all-in price is a cap, so an ad that needs fewer images than the bound simply costs less. A stage that fails is refunded on the usual rules and the retry is quoted on what is left. And a shot you do not like can be fixed on its own, which is far cheaper than a new ad. The one price is a cap.

SceneMixer shows the price on the button, and your balance is in the top bar. If the balance does not cover the all-in price, the ad does not start; it asks you to top up. Whether the credits a new account starts with cover a short ad at the lowest quality is something only the button can tell you, because prices are set by the site and can change.

## If you wanted your own picture in a free story instead

The photo-to-ad door is the only place a free SceneMixer account can upload a picture. Putting your own picture on a character, place or prop card is a paid-plan feature, and so is uploading to the asset library. The free route into a story is words: describe the character, let the site draw a setup picture, and spend the 15 free setup images that are shared across your account.

| You want | Free route | Where it stops |
| To see what the AI makes of your product photo | Add it and read it | Stops before the ad is made |
| Your own face on a character | Describe the face in words | A photo on the card needs a paid plan |
| A moving version of one still | Not offered | There is no one-photo animate tool |
| A free taste of video | The 5 second preview in a parsed script project | Once per account, lowest tier |
| A story with a cast and a storyboard | Paste a page, pay the small reading fee, then the cast and first storyboard are free | Video is charged from the first second |
Free AI text to video covers the story ladder in full, and free sheets are per account explains the pool of 15.

## What free output looks like

A free SceneMixer account’s video carries a small AI badge in the corner, burned into the picture, and a line of AI-generated-content information in the file. Downloading does not add or remove it, and a link shared through the site always shows it. Taking the badge off is a paid-membership setting, and it does not remove any duty you have to say a video is AI-made where you publish. Why shared videos carry the badge.

There is no free video allowance beyond the one preview. If you need to judge the look of a person before paying for an ad, the cheapest honest test is a script project with a short text: after a small reading fee, the cast (within the free pool) and the preview shot are free.

## FAQ

### Is there a free image to video AI?

SceneMixer is free up to the reading of your photos. Making the ad costs credits, and a new account starts with 50 of them. Video is charged from the first second in every kind of project.

### Can I make an ad from a photo for free?

Not entirely. Reading the photos is free, and the price of the ad is shown on the button before you press it. Whether the sign-up credits cover it depends on the length and quality you pick.

### Is reading my photos charged?

No. The reading is free on SceneMixer and nothing is charged until you choose one-click or expert mode. A photo that cannot be made out is refused at no cost.

### Do I get the free setup images for an ad?

No. SceneMixer’s 15 free setup images belong to script projects that have been read and paid for. Ads and remakes never draw from that pool.

### Is there a free preview for an ad?

No. SceneMixer offers the free 5 second preview in a parsed script project, not in an ad project.

### Does a free account put a watermark on the video?

SceneMixer burns a small AI badge into the corner of every video on a free account. Removing it is a paid-membership setting.

### Can I animate one photo for free?

SceneMixer has no tool that animates a single still, free or paid. A photo can become a reference for a character, place or prop, or one of up to four photos in an ad.

### How do I see the price before I pay?

Every SceneMixer button shows its price. The quality menu in the ad composer lists the all-in price of each option, and the balance is in the top bar.

Updated November 2, 2026


---

# Insert or Delete a Segment in an AI Episode | SceneMixer

URL: https://scenemixer.com/guides/insert-or-delete-a-segment/
Both are free. What a new segment contains, how to let AI write a blank segment, what deleting removes, and how to split a segment that is too long.

# Inserting and Deleting Segments

Quick answer: On a segment’s card, Insert adds a blank segment right after it: write its header and shots — or let AI write them — then generate. Delete segment removes the segment, its video and its version history — credits already spent on it are not refunded, and you are asked to confirm. Neither action costs anything. A segment whose video is generating cannot be deleted, and an episode always keeps at least one segment.

Newer storyboards. In storyboards written or regenerated since late October 2026 a segment stays in one place and one timeline, so an inserted segment that happens somewhere else needs its own Location (one place, one timeline). Segments you insert or edit are checked when you generate: if the system may not have read one correctly, a box asks you to look first (when a segment may not have been read correctly).

Open a finished episode and look at the timeline. Each card is a segment; the controls are on the card.

## Inserting a segment

Select the segment you want the new one to follow and choose Insert. A blank card appears after it, and the editor says: blank segment inserted, write its script, then generate. Numbering shifts to make room.

A blank segment cannot be generated yet — it needs a header and shots first, and it has neither. There are two ways in: write it yourself, the same way the segments around it are written — the header (Location, Time, Weather, Opening frame, Set dressing, Props) and the shots; or let AI write it, in the next section. Its Opening frame should describe where the previous segment left off. The header · Writing a shot.

## Let AI write the segment

Open the blank segment’s edit box and it starts you off with a one-line skeleton — a shot marker and a few seconds, like [Shot 1] 3s: — ready for you to fill in. Saving that skeleton untouched counts as no change at all. If you would rather describe the scene and let the AI do the typing, use AI edit instead: on a segment with no storyboard text, the panel becomes Let AI write this segment.

- On the blank card, open AI edit.
- Say what happens in the segment. Examples: “A two-person standoff on the rooftop, 4 to 6 shots, end on a hook.” · “Carry on from the previous segment: he catches up in the parking lot, and the rain has not stopped.” · “A quiet night scene: two lines of dialogue, distant traffic in the sound.”
- Press Write. The AI writes the segment in the format of the segments around it: a header, numbered shots with their seconds, lines, and ambience.
- Read the review before anything is saved: every changed line is marked, and the summary shows the shot count it wrote. If you want to adjust a word, press Edit by hand in the review.
- Save. The text lands on the card, and the segment generates like any other.

It reads the neighbouring segments to stay in format — the same header fields, the same shot markers, the same way dialogue and ambience are written — so a new segment slots in without bending the episode’s conventions. What it costs in SceneMixer is one AI revision, and only when a revision reaches the review: a failed attempt, or text that comes back unchanged, is not charged. And until the segment has text, there is nothing to generate — a blank card cannot produce a video.

One more thing to know: Revise the whole episode — the one-sentence request that runs over the episode’s segments — skips blank segments, because there is no text to revise. Give the blank one its text first. How AI edit works · AI edit or edit by hand.

## Splitting a segment that is too long

The usual reason to insert. When a segment’s shots add up past the per-segment length SceneMixer sets for this episode (15 or 30 seconds), insert a segment after it, cut the last shots from the long one and paste them into the new one, then give the new one a header. Both segments are now within that length, and each is generated and priced on its own. Shot seconds and the cap.

## Deleting a segment

Choose Delete segment on its card. The confirmation says exactly what goes: Delete segment N? Its generated video and version history will be removed; credits already spent are not refunded. Confirm and the card is gone, along with every version of its video. The text is gone too, so copy anything you want to keep first.

Two things stop a deletion. A segment whose video is still generating: wait for it to finish or stop it first. And the last segment of an episode: at least one segment must remain.

## What happens around it

SceneMixer leaves other segments untouched. The episode’s composite, if you made one, is untouched too; it is a separate file. Composite again to get a new file without the deleted segment. Numbering closes up.

## Inserting versus regenerating the storyboard

Inserting and deleting are surgical: one segment, free, nothing else moves. Regenerating the episode’s storyboard rewrites every segment and discards their videos. When the breakdown is mostly right, insert and delete; when it is wrong throughout, regenerate. What regenerating the storyboard costs you.

| Action | Cost | Removes | Blocked when |
| Insert segment after | Free | Nothing | — |
| Delete segment | Free (spent credits not refunded) | Text, video, version history of that segment | Its video is generating; it is the last segment |
| Let AI write a blank segment | One AI revision, free when it fails or changes nothing | Nothing | — |
| Regenerate storyboard | One storyboard fee | Every segment’s text and video | Any segment is generating |

## FAQ

### Does inserting a segment cost credits?

No. Neither does deleting. In SceneMixer, only generating video is charged.

### Why can't I delete this segment?

Either its video is still generating on SceneMixer — wait or stop it — or it is the only segment left.

### Do I get credits back when I delete a segment with a video?

No. Credits already spent on that video are not refunded by SceneMixer; the confirmation says so.

### Where does a new segment get its header?

You write it — copy the previous segment's header as a starting point and change what differs — or let SceneMixer’s AI write the whole segment: open its AI edit, describe what happens, press Write, and review the draft.

### Does deleting a segment change my composite?

No. The composite on SceneMixer is a separate file. Composite again to rebuild it without the segment.

### Can AI write the blank segment for me?

Yes. Open the blank card’s AI edit and the panel becomes Let AI write this segment: describe the scene and press Write. SceneMixer follows the format of the neighbouring segments — header, shots with seconds, lines and ambience — and nothing is saved until you approve the review.

### What does the edit box open with on a blank segment?

SceneMixer opens it with a one-line skeleton: a shot marker and a few seconds, like [Shot 1] 3s: . Fill it in, or replace it. Saving it untouched counts as no change.

### Do blank segments get included in the whole-episode request?

No. Revise the whole episode runs one sentence over the episode’s segments, but a blank segment has nothing to revise — write its text first, then the episode request can reach it.

Updated November 3, 2026


---

# Interface, Script, Dialogue: Which Language Controls What

URL: https://scenemixer.com/guides/interface-language-script-language-dialogue-language-what-is-what/
Five things here have a language, set in different places. Which changes the menus, which the script AI writes, which the characters speak, which the storyboard you read.

# The Languages on This Site: Interface, Script Writing, Working Text, Dialogue, and What You Read

Quick answer: Five things on this site have a language. Interface language: menus and buttons, chosen in the top bar. Writing language: the language AI Create writes your script in; it follows the interface. Working text: the language the project’s cast, scenes and storyboards are produced in, either Chinese or English, taken from the interface language when you confirm the Step 1 settings. Dialogue language: what the characters say in the clips, chosen in Step 1 among the site’s supported languages. What you read: if your interface is in a third language, the working text is translated for you to read and edit, automatically. Change the interface and the reading language follows; the dialogue language and the working text stay what they were when the script was read.

Switch the interface to another language on a parsed project and open an episode: the storyboard shows in that language within a minute; the clips still speak the dialogue language.

## Where each one is set

Interface: the language menu in the top bar; remembered on this browser. Writing language (AI Create): not a control, it follows the interface at the moment you write. Working text: not a control either; Chinese or English by the interface when you confirm Step 1 (the Chinese site is always Chinese). Dialogue: the Step 1 dropdown, applied at the parse. Reading language: follows the interface; when interface and working text differ, you read a translation. Details: Work in Your Own Language, Choose the Dialogue Language.

## Why the working text is only Chinese or English

The models that generate images and clips read those two best; producing the storyboard in Korean and handing it to a video model would cost accuracy. So the production documents are written in one of the two, and you read them in your language. Editing what you read writes the change back into the working text; nothing is lost in between. The only visible effect: the structure markers inside a storyboard (segment fields, shot numbers) are shown in your language but stored in a fixed form.

## Dialogue is separate from all of the above

Characters can speak Japanese in a project whose working text is English and whose interface is German. In SceneMixer, the dialogue language is written into the lines at the parse and copied into the clips as spoken audio; it is fixed once the script is read. What you can change after the read.

## Where a mismatch shows up

You wrote a Korean script but the summary is in English: expected, the working text is English; read it in Korean by switching the interface. The clips speak English but you wanted Korean: the dialogue language was left at the default when the script was read; edit the lines in the segments, or start a new project with the same script and choose Korean before it is read. The storyboard is half translated: translation is still running; a status line above the segment says so and the original is shown meanwhile.

## AI Create and price

SceneMixer's AI Create writes in your interface language and is priced by that language’s counting unit, then parsed into the working text like any pasted script. Write the Script in Your Language.

| Language of | Set where | Changes when |
| Menus and buttons | Top-bar language menu | Immediately |
| Script AI Create writes | Follows the interface | Next time you write |
| Working text (cast, scenes, storyboards) | Interface when you confirm Step 1; Chinese or English | Not for material already produced |
| What characters say | Step 1 dialogue language | Fixed once the script is read |
| What you read and edit | Follows the interface | Within a minute of switching |

## FAQ

### I switched the interface; will my clips be re-dubbed?

No. Clips speak the dialogue language; the interface only changes what you read.

### Can the working text be Korean?

No; SceneMixer's working text is Chinese or English. You read and edit in Korean through the automatic translation.

### The translation has a wrong word; can I fix it?

Yes, edit it like any text; the change is written back.

### Which language is my project's summary in?

The working text language; shown to you in your interface language.

### On the Chinese site?

Interface, working text and dialogue are all Chinese.

Updated November 3, 2026


---

# "Invalid Input: Check Reference Images and Duration" — What to Fix

URL: https://scenemixer.com/guides/invalid-input-reference-and-duration/
The one failure message about your segment rather than the service: an image the model cannot use, a crowded segment, or the seconds. What to change for each.

# "Invalid Input — Check the Reference Images and Duration": What It Means and What to Change

Quick answer: Every other failure message is about the service; this one is about the segment. The request was refused before anything was generated, so the credits normally come back — the task’s line in your credit history shows Refunded. Three things to check, in this order: an image you uploaded that has an unusual shape or size, a segment that names more people and props than its shots show, and shot seconds that do not add up to the segment’s length. Fix the one that applies and generate again.

Select the failed segment, open Edit, and read its Props line and the seconds of each shot.

## Where the message appears

The segment card says Failed and the preview pane shows Invalid input — check the reference images and duration. The credits held for this generation normally come back to your balance; the storyboard text, the seconds and the earlier versions are untouched. The general families of failure messages are in When a Generation Fails; this page is only about this one sentence.

## Check one: an image you uploaded

Sheets you uploaded yourself, and images you set for one episode, must be an ordinary JPEG or PNG with a sensible shape: very tall or very wide pictures, tiny ones and files that failed to load are refused. The message mentions image resolution when this is the cause. Re-upload the picture cropped closer to the sheet’s own shape (portrait for people, landscape for places and props), or let SceneMixer regenerate the sheet. Use Your Own Photo for a Character or Scene.

## Check two: a crowded segment

Everything a segment names comes along to the generation. A segment that lists many props on its Props line and names people who never appear on screen can be refused. Remove props the shots do not show from the Props line, and write people who are only talked about without @. Props and Locations in a Shot · Who is in frame.

## Check three: the seconds

A segment longer than its tier’s cap never gets this far: the editor refuses it when you press generate and names the cap (30 seconds on the 30-second tiers, 15 on the others; which tier accepts what). If you see this message on a segment close to its tier’s cap, shorten one or two shots by a second each, or split the segment in two. Insert or Delete a Segment.

## If none of the three applies

Generate it one more time. If the same segment fails again with this message and you cannot see why, send feedback with the segment number and a screenshot; SceneMixer support can read what the service answered for that task. How to get help.

| Check | How to tell | What to change |
| An image you uploaded | Message mentions image resolution | Re-upload it cropped to the sheet's shape, or regenerate the sheet |
| A crowded segment | Many props and names in one segment | Trim the Props line; write off-screen people without @ |
| The seconds | A segment close to its tier's cap | Shorten a shot or two, or split the segment |
| None of the above | Retry fails the same way | Feedback with segment number and screenshot |

## FAQ

### Were my credits charged?

Normally not: SceneMixer refunds a request refused before generating, and the task's line in your credit history shows Refunded. If it says Under review instead, support reviews it.

### Why did the same segment work on another tier?

SceneMixer's tiers accept different things and have different length caps. A segment one tier refuses can be fine on another.

### Does splitting the segment cost more?

In SceneMixer, you pay per second generated; two shorter segments cost about the same as one long one, plus you get a natural cut.

### I did not upload any image. Why does it mention images?

The message on SceneMixer is the same for every cause. Check the Props line and who is named in the shots first.

Updated November 3, 2026


---

# "IP Suspected": Celebrity Faces, Brands and Franchises in AI Video

URL: https://scenemixer.com/guides/ip-suspected-celebrity-brand-franchise/
A separate provider check refuses images and clips that resemble a known person, logo or franchise. What triggers it and how to design around it.

# “IP Suspected”: When a Sheet or Clip Is Refused for Resembling a Real Person, Brand or Franchise

Quick answer: Next to the content check there is an IP check: it refuses a sheet or a clip whose input or output resembles a real person, a brand mark or a character from a known film, game or comic. It is triggered by names in the text, by iconic outfits and props, and by realistic styles that land on a famous face even when you named nobody. The refused attempt is refunded automatically, within the same daily number as other video blocks. The fix is design, not rewording: an original face with its own anchor feature, a fictional brand, an outfit that is not the iconic one.

Hover the reason on the failed card: the provider’s original message shows, and it names the IP check.

## What triggers it

Names: an actor, a singer, a politician, a product or a franchise character named in a description or a shot. Likeness: a description detailed enough to land on a famous face, or a reference image of one. Iconic design: the outfit, the mask, the weapon, the vehicle that identifies a character even without a name. Marks: logos and lettering on clothing, bottles, screens. The check looks at both what goes in and what comes out, so a clean description can still be refused if the drawn face happens to resemble someone.

## Why realistic styles get it more

Live-action styles draw faces from the space of real faces, and a well-described adult in a suit can land on a well-known one. SceneMixer's stylised presets — anime, 3D, illustration — move the face away from any real person and clear the check far more often. If your project must be realistic, add one distinctive feature the famous face does not have: a scar, a broken nose, a specific hairline. Visual Style Presets, One by One.

## Designing around it

Replace the name with the face: jaw, brow, eyes, hair, one anchor feature. Replace the iconic outfit with your own: same function, different cut and colour. Replace the brand with the object: “a black smartphone”, “a red soda can, no label”. For a parody or an homage, the further you are from the original design, the more of the clip you get to keep. Rewriting a blocked description.

## Your own photos

Uploading a photo of a public figure as a character reference is refused by the same check and is against the rules in any case. Your own face, or a person who has given you permission, is fine; a stranger’s is not. Content Rules and Moderation.

## Credits, and repeated refusals

An IP refusal is refunded automatically; for video it counts like any other block, so past 12 blocks in 24 hours SceneMixer holds the credits for review instead. Repeated refusals count like other moderation blocks: too many in a short time pause video generation for a while, and deliberate attempts to get a real person or a franchise through are treated as a rules violation. Refund rules.

| Refused for | Change |
| A named actor or public figure | Describe a face; add a feature they do not have |
| A franchise character | Original design: new outfit, new colours, own anchor feature |
| A logo or product name | The object without the mark |
| A realistic face that happens to match someone | Switch to a stylised preset, or alter the description |
| A photo of a public figure | Not allowed; use a permitted face |

## FAQ

### I did not name anyone. Why IP suspected?

The drawn face resembled someone, or an outfit was iconic. Change one anchor feature or the style.

### Was I charged?

Normally not: SceneMixer refunds refused attempts. Past the daily number of video blocks, the credits are held for review.

### Can I make a fan film of a known series?

Not with the original characters or designs; the check refuses them. Original characters in your own story are fine.

### Does the check apply to scenes and props?

Yes: a famous building drawn as a landmark can pass, a branded product or a franchise vehicle will not.

### Can support whitelist my character?

No; the check runs at the provider. Design is the only way through.

Updated November 3, 2026


---

# Is AI Movie Maker Legit? Seven Checks You Can Run Yourself

URL: https://scenemixer.com/guides/is-ai-movie-maker-legit/
Is an AI movie maker legit? Seven checks to run before you pay (operator, prices, failed jobs, free trial, samples, terms, support) and where SceneMixer's answers sit.

# Is AI movie maker legit?

Quick answer: No single page, least of all a maker's own, can settle ‘legit’ for you, so this page lists seven checks you can run yourself, mostly before paying, and says where SceneMixer's answer to each is written so you can check rather than believe. Who runs it. What it costs. What happens when a generation fails. What you can try free. Whether the samples are real. What the terms say about your files. Whether you can reach a person and leave. It does not say whether making or publishing AI movies is legal where you live; that is not judged here.

The pages to open as you read: Pricing · Terms · Refund policy · Legal notice · Sample dramas.

## 1. Who runs it

Look for a legal notice that names the company behind the site, and a way to write to it. SceneMixer's Legal Notice names the operating company and where it is registered, and gives a support address; Contact lists the ways to reach it. A maker with no named operator is a maker you cannot hold to anything.

## 2. What it costs, before you sign up

Open the pricing page without an account. You should be able to see what a second of video costs and what the cheapest pack and plan are. Here: credit packs start at $1.49, memberships at $7.99 a month, video is 10 credits a second on the default tier, and every button shows its price before you press it. Pricing · Where credits go in one episode.

## 3. What happens when a generation fails

Ask what is charged when something goes wrong. The Terms say: “Tasks failing before reaching the provider are auto-refunded; ambiguous failures are held for review, never silently charged.” Credits are held when you submit, charged when the task delivers, and returned when it fails; each task is one line in your credit history with a tag that says which. In SceneMixer, a delivered generation is charged even if you dislike it, and fixing it is a new charge. What comes back, what waits, what doesn't · When a generation fails.

## 4. What you can try before paying

New accounts start with 50 credits. The first 15 setup images, the storyboard for episode 1 of your first drama, one free promo cover and one 5-second preview are free; reading a script, writing one with AI and video are charged normally, and there is no free video beyond the preview. What you get for free.

## 5. Are the samples real?

Each sample project says how it was made: its settings, its number of episodes, and that it was produced end to end from the script. Open one, read its storyboard text, and compare it with what is on screen shot by shot. If a maker's samples cannot be traced to a script and settings, ask why. The sample dramas · Learn from a sample project.

## 6. What the terms say about your files and your rights

Read the sections on ownership and data. SceneMixer's Terms say you retain rights in your uploads and own the generated results to the extent permitted by law, and: “Unless you separately consent or the law requires it, we do not use your uploads or outputs to train models; upstream providers process data under their own terms (see Privacy).” Before your first upload a window asks you to confirm you have the right to use the material and that every real person whose face or voice appears has consented. Deleting a project has a seven-day window, and you can delete your account. Privacy · The seven-day window · Delete your account · What you get to keep.

## 7. Can you reach a person, and can you leave?

Support is by email and by the Ask window; the refund guide says replies take up to 7 business days. Card and PayPal subscriptions can be cancelled any time from the Membership page, with access until the end of the paid period. SceneMixer's credit packs and memberships are non-refundable except for a duplicate charge, a purchase that could not be delivered, or a purchase by a minor without a guardian's consent. Credits have no expiry unless the purchase or campaign page states one. Refund and cancellation policy · Subscribe, upgrade, cancel · Do credits expire?.

## Red flags in any AI movie maker

- The price appears only after you have paid, or only per batch with no per-second figure.
- ‘Unlimited’ with no terms behind it. Why AI video unlimited plans disappeared.
- No refund page, or one that does not say what happens to a failed job.
- Samples with no settings, no script and no way to compare.
- No named operator and no contact other than a form.
- A maker that never says what it cannot do. Ours says it on each page; for example it takes no audio, draws no readable on-screen text and refuses recognisable people.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence; Costume Style Smart Casual; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the cast and read the storyboard before you render anything; it is the cheapest way to check a maker's claims. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Check | Where to look on SceneMixer | What you should see |
| Operator | Legal Notice, Contact | The company named, its registration, a support address |
| Price | Pricing; the button before every action | Per-second video price, pack and plan starting prices |
| Failed jobs | Refund guide; your credit history | Held, charged, refunded: one tag per task |
| Free trial | What you get for free | Welcome credits, free setup images, a free storyboard, a short preview |
| Samples | Sample dramas | Settings and the script behind each one |
| Terms | Terms, Privacy | Ownership, no training on your uploads without consent, the confirm-before-upload window |
| Support and exit | Refund policy; Membership page | Cancel any time; refunds in the listed cases |

## FAQ

### Is SceneMixer legit?

This page can't call it for you; it lists seven checks and where our answers are written. Run them, and read the Terms and the refund policy before you pay.

### Is it legal to make AI movies?

That is a question of the law where you live and the platforms where you publish, and this site does not judge it. Pages with official wording: copyright and YouTube.

### Is it free?

Part of it is: SceneMixer's account allowances are free, video is charged from the first second. What is free.

### Who owns what I make?

Under SceneMixer’s Terms you own the generated results to the extent permitted by law, with no guarantee they are copyrightable. With sources.

### Do you train on my scripts?

SceneMixer's Terms say not unless you separately consent or the law requires it; upstream providers process data under their own terms, set out in Privacy.

### What if a generation fails?

Failures that never reached the provider are refunded automatically on SceneMixer; unclear ones are held for review and not charged silently. A delivered generation is charged.

### Will my credits expire?

An expiry applies only where the purchase or campaign page states one, according to SceneMixer’s Terms.

### What if I am charged twice?

A duplicate charge on SceneMixer is refunded to the original payment method; email support with the order number.

Updated November 3, 2026


---

# Keep Characters Consistent with Kling 4.0: Sheets and Slices

URL: https://scenemixer.com/guides/keep-characters-consistent-with-kling/
How a cast stays recognisable across episodes: one reference sheet per character, attached to every segment, and what Kling's identity features officially add.

# Keeping Characters Consistent When You Finish on a Kling Tier

Quick answer: a character stays recognisable because of the reference layer, not because of the tier that renders the shot. Here the cast is drawn once — one reference image per character in step 2 — and the site attaches those images to every segment the character appears in, on any tier, including Kling 4.0 and Kling 4.0 Flash. The storyboard is the second half: it names exactly who is in frame for each shot, and the lines and voices stay with the same character across the series. Kling's own identity features — element references and Character Identity 3.0 — are the officially announced capabilities of the model itself; what a tier here exposes is what the model picker shows.

## Consistency is three layers, and a tier is none of them

The still above is from The Veil at Dawn, produced on this site; its cast kept the same faces because the sheets below were attached to every segment.

A face drifts for three reasons, and each has its own layer: a written description gets diluted sentence by sentence, a reference image carries the light it was shot in, and two shots with nothing in common between them have no anchor. The fix is not a better tier or a longer prompt — it is the layer that holds the information:

- The character sheet — what the face and build are. Drawn once, reused everywhere.
- The scene plate — where the scene is and how it is lit, so the sheet is not fighting the background.
- The storyboard — who is in each shot, where they stand, and what carries over from the shot before.

That model is worked through in detail in AI short drama character consistency: 3 layers. What follows is how it plays out when you finish the episode on a Kling tier.

## One sheet per character, drawn once

In step 2, each character gets a reference image of their own, generated from the script's description of them: face and build first, with one instantly recognisable anchor (a scar, a mole, a piece of jewellery, a haircut), against a plain background under directional light. That sheet is the cast's identity; every segment a character appears in is rendered against it, whatever tier produces the take.

Three things worth knowing about the sheets:

- They are per character, not per shot. Regenerating a sheet changes the character everywhere; you are not fixing faces segment by segment.
- You can replace or regenerate one without touching the rest of the cast. Regenerate or replace one sheet.
- Scenes and props get the same treatment — a location image and a prop image each, so a room or an object does not change between episodes either.

Your first 15 reference images on the account are free (characters, scenes and props share that pool); after that each one is charged at the image tier's rate.

## What gets attached to each segment

When a segment renders, SceneMixer looks at the storyboard and attaches the references that segment needs: the sheet of every character listed in frame, the location image, the prop images, and any state image for a costume change or a changed scene. You do not choose references by hand for every shot — that assembly is what the pipeline does.

Two consequences to keep in mind:

- The in-frame list is the switch. A character who is not named in a shot is not in it; a character who is named gets their sheet attached. Naming someone who is not visible is one of the most common ways a shot drifts from the plan.
- State images are part of consistency too. If a character changes clothes or a room burns down, that is a new reference, not a sentence. Costume changes and scene states.

## Writing shots so the face holds

The storyboard is where continuity is decided, one shot at a time:

- One beat per shot, with the seconds to carry it.
- Who is in frame, honestly. Carried-over positions and actions, not a fresh description that contradicts the shot before.
- The same character keeps the same voice from episode to episode: voices are set per character in step 2, either described in words or matched to a preset from the library that you can listen to. How character voices work.

How to write one shot is the long version of this list, and spatial consistency covers where people stand between shots.

## What the Kling tiers change, and what they do not

The tier decides the picture, not the identity. Choose it per segment in the model picker:

| Tier | Resolution | Segment cap | Price |
| Kling 4.0 | 1080P | up to 30 seconds | 38 credits per second |
| Kling 4.0 Flash | 720P | up to 20 seconds | 29 credits per second |

A higher-resolution tier gives a sharper picture — useful when a face is seen in close-up or a clip will be watched on a large screen. It does not make a face more stable: the same sheet rides along on both tiers. The cheap way to use that: draft the episode on the default tier while the story is still moving, then re-render the keepers — especially the close-ups — on Kling 4.0. Making a short drama with the Kling tiers walks the whole flow.

One cap to plan around: Kling 4.0 Flash renders segments up to 20 seconds. A long dialogue scene split into shorter segments renders on either tier; a single unbroken take past 20 seconds needs Kling 4.0, which takes up to 30.

## Kling's own identity features, as officially announced

The model itself ships with identity tools, and it is worth knowing what they are — with the label that matters attached: these are the officially announced capabilities of Kling's own model, not a description of what a tier here exposes.

- Element reference: as officially announced by Kling, a subject — a person, an object, a place — can be defined once and referenced in later generations.
- Character Identity 3.0: officially announced with Kling 3.0, a character can be defined from a reference video or several images, and Kling's own guidance says a character video of about 3–8 seconds gives the strongest consistency.
- Element Voice Control: the officially announced feature that binds a voice to a referenced subject, so the character arrives with the same sound.
- Omni Reference: officially announced with Kling 4.0 — up to 15 references per task, ten images, five videos, up to seven subjects, plus a voice.

Where you are working through this site, the equivalent job is done by the sheets, the scene plate and the storyboard, and the take renders on whichever tier you picked. The two are not in competition: they are the same idea — keep the identity in a reference and reuse it — expressed in two places. The announcement is at Kling's 4.0 model release page.

## A checklist before you call a series consistent

- Five shots, one character, side by side. Put five shots of the same character next to each other, including a close-up and a full-body, and look at them at the same size. If the face changes between them, the sheet is the place to fix it.
- One anchor feature visible in every shot where the face is readable.
- Wardrobe and props match the state the story is in at that episode.
- Every shot lists who is actually in frame, and nobody else.
- Voices are set per character, not per line, so episode 10 sounds like episode 1.
- Only the keepers are re-rendered on the higher tier; the rest stays where it was.

## FAQ

### Do Kling tiers keep a face consistent on their own?

Consistency comes from the reference layer, not the tier. Here the character sheet is attached to every segment the character appears in, on any tier; the tier decides resolution, segment length and price.

### Where do I draw the character sheets?

In step 2, after the script is read: one reference image per character, plus scene and prop images. The first 15 reference images on your SceneMixer account are free.

### Can I fix one character without redoing the others?

Yes. Regenerate or replace that one sheet; every segment featuring the character picks up the new sheet. The rest of the cast is untouched.

### Does a higher-resolution tier make faces more stable?

No. It makes the picture sharper, which helps in close-ups and on large screens; the identity still comes from the sheet, and the same sheet rides along on both tiers.

### What does Kling officially announce for character identity?

Element references for subjects, Character Identity 3.0 (define a character from a reference video or several images; a 3-8 second character video is Kling's own recommendation for the strongest consistency) and Element Voice Control. Kling 4.0 was officially announced with Omni Reference at up to 15 references per task.

### Is any of that a promise about what happens here?

No. Those are the officially announced capabilities of Kling's model. What a tier on this site exposes is what the model picker and the pricing page show, and the identity layer here is the sheet plus the storyboard.

### What about voices across episodes?

A character's voice is set per character in step 2 - described in words, or matched to a preset from the library - so the same character sounds the same in every episode.

### What is the single most common cause of a face drifting mid-episode?

A shot that describes the character again in new words, or names someone who is not actually in frame, instead of carrying over the previous shot's positions. Write one beat per shot and keep the in-frame list honest.

### Do the Kling tiers cost extra for references?

No. The reference images are generated in step 2 and charged there once (the first 15 on the account are free); attaching them to a segment is part of rendering, which is billed per second.

Updated November 3, 2026


---

# Kling 3.0 Multi-Shot: 6 Shots in One Take, and Storyboards

URL: https://scenemixer.com/guides/kling-3-0-multi-shot-storyboard/
Kling 3.0 cuts up to 6 shots inside one 3-15 second take, with native audio and character references. What was announced, and how it maps to a storyboard.

# Kling 3.0 Multi-Shot and Native Audio: What the Model Does

Quick answer: Kling 3.0 and 3.0 Omni, announced in February 2026, generate a take of 3–15 seconds in which the model can cut between up to 6 shots, each with its own duration, framing, angle, content and camera movement; Kling's own blog describes an AI Director handling shot-reverse-shot dialogue and cross-cutting. Audio comes out of the same pass — dialogue in five languages with accents and per-speaker line assignment — and the reference modes cover elements, multiple images and video elements, with Character Identity 3.0 for keeping a face. Everything below is what Kling announced; the parts about storyboards are about writing shots down, which is what this site does episode by episode.

## What Kling 3.0 announced

The still above is from Crimson Vow, produced on this site; the storyboard screenshot further down is the page that decided its shots.

Kling 3.0 and 3.0 Omni were announced in February 2026 (a press release dated February 6, 2026, and material on Kling's own blog). The headline capability is multi-shot generation: one generation produces several shots with cuts between them, rather than one unbroken camera move.

Per the announcement: a take runs 3–15 seconds; a generation can contain up to 6 shot changes; and each shot can carry its own duration, framing, angle, content and camera movement. Kling's blog describes an AI Director that handles shot-reverse-shot dialogue and cross-cutting inside the take. Third-party write-ups of the same release market the feature as Smart Storyboard and Custom Storyboard.

Access is tiered on Kling's side: the official page states that unlocking 3.0 requires a Pro, Premier or Ultra membership.

## Native audio: the second half of the announcement

Kling 3.0 generates video and audio in one pass rather than dubbing afterwards. The announced language list is English, Chinese, Japanese, Korean and Spanish, with accents (American, British and Indian English among them, plus several Chinese dialects). In scenes with several speakers, the announced behaviour is that lines are assigned to the right speaker.

The practical consequence is timing. When the line is part of the render, the mouth movement, the pause and the shot length are decided together; when a line is added later, someone has to bend one to fit the other. The reasoning behind generating dialogue natively is covered in more depth in Native dialogue vs dubbing.

## Reference modes and Character Identity 3.0

The announced modes are text-to-video, image-to-video, first-and-last-frame, multi-image reference, element reference (a subject — a person, an object, a place), and video element reference.

Character Identity 3.0 is the consistency layer: a character can be defined from a reference video or several images. Kling's guidance says the strongest consistency comes from a character video of about 3–8 seconds. An Element can also be bound to a voice (Element Voice Control), so the same subject arrives with the same sound.

On-screen text is the last announced capability worth naming: signage, subtitles and logos are described as staying legible during motion, which matters for any shot where a sign carries story information.

## Six shots in one take: what changes in the writing

A generation that can cut internally changes what a take is for. Instead of one camera movement, you are writing a short sequence: an establishing shot, the reaction, the two-shot, the insert — with the cuts already inside the clip.

Two consequences follow, and both are about editing discipline rather than models:

- The cut between shots inside the take is decided at generation time. You write it, you do not find it later in an edit. That is closer to a storyboard than to a prompt.
- Each shot inside the take still needs its own information — who is in frame, how long it lasts, what is said. A six-shot take is six small scenes with one shared take number.

## The storyboard idea, one level up

Everything Kling 3.0 asks the writer for — shot boundaries, per-shot duration, framing, who is in frame, what is said — is what a storyboard already is. The model made the paper step part of the input.

That is the same structure this site uses for a series. A script is read into episodes; each episode gets a storyboard made of segments; each segment holds numbered shots, and each shot carries its seconds, the people in frame and the lines to be spoken. Segments then render one at a time and composite into one file per episode. How to make an AI short drama walks the pipeline end to end.

So if you are coming to this from Kling 3.0, the useful translation is: what the model calls a multi-shot generation is what a storyboard page calls a segment; what it calls a shot with duration and framing is what the storyboard calls a numbered shot with its seconds. The vocabulary is different; the discipline is the same, and it is the part that transfers.

## What a storyboard looks like here

A segment in this editor is short and specific. It opens with where and when (location, time of day, weather), then lists numbered shots. Each shot line carries its duration in seconds, the characters who are actually in frame, and any spoken line, with the sound noted separately. In SceneMixer, editing any of that text is free; regenerating the segment is what costs.

Two habits make six-shot writing work on either side of the fence:

- One beat per shot. If a shot has two things happening, it is two shots. A cut inside a take is still a cut.
- Write the in-frame list honestly. Naming someone who is not visible in the shot is one of the most common causes of a frame that drifts from the plan. How to write one shot covers what belongs on the line and what does not.

## Where the two approaches differ

Multi-shot generation decides the cut at generation time, inside one take. SceneMixer cuts between segments: each segment renders as its own take and the episode composite joins them. The trade is straightforward — internal cuts buy continuity inside a scene and cost you the ability to re-cut after the fact; separate segments make every join reviewable and re-renderable on its own.

For series work the practical pattern is the same either way: fix the story in the script, fix the shot list in the storyboard, and only then spend render time. Editing one shot without redoing the episode is the version of that discipline here.

## Things to keep in mind about the announcement

- 3.0's numbers are its own. 3–15 seconds and 6 shots describe 3.0; the later 4.0 announcement moved the numbers again. Compare like with like.
- Membership-gated. The official page ties 3.0 access to paid Kling membership tiers; the free tier does not reach it.
- Third-party names are marketing. “Smart Storyboard” and “Custom Storyboard” come from coverage of the release; Kling's own blog describes the same capability as multi-shot generation with an AI Director.
- Character video range. The 3–8 second guidance for the strongest character identity is Kling's own recommendation, not a limit on the formats you can try.

## Sources

- Kling: Video 3.0 Omni multi-shot and native-audio guide (kling.ai blog, June 2026) — shot counts, durations, audio languages, reference modes and Character Identity 3.0.
- Bernama: Kling 3.0 press release (February 6, 2026) — the launch and the model list (Video 3.0, Video 3.0 Omni, Image 3.0, Image 3.0 Omni).

Kling is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Specifications are quoted as announced and may change.

| Kling 3.0, as announced | Figure |
| Take length | 3–15 seconds |
| Shots per generation | Up to 6 shot changes |
| Per-shot settings | Duration, framing, angle, content, camera movement |
| Native audio languages | English, Chinese, Japanese, Korean, Spanish |
| Character reference | Character Identity 3.0; 3–8 second character video for strongest consistency |

## FAQ

### What is multi-shot generation?

One generation that contains several shots with cuts between them, instead of a single unbroken camera move. Kling 3.0 announced up to 6 shot changes per generation, each with its own duration, framing, angle, content and camera movement.

### How long is a Kling 3.0 take?

3 to 15 seconds, per the announcement. A later announcement (Kling 4.0) raised the ceiling to 30 seconds, so check which version a page is describing.

### Does Kling 3.0 generate sound?

Yes - audio and video come out of the same pass, with dialogue announced in English, Chinese, Japanese, Korean and Spanish, accents and Chinese dialects, and lines assigned to the right speaker in multi-person scenes.

### What is Character Identity 3.0?

Kling's character-consistency feature: define a character from a reference video or several images and reuse that identity. Kling's guidance says a 3-8 second character video gives the strongest consistency.

### Can a character keep its voice?

The announcement includes Element Voice Control, binding a voice to a referenced subject so the same character arrives with the same sound.

### Do I need a paid membership to use Kling 3.0?

Kling's official page states that unlocking 3.0 requires a Pro, Premier or Ultra membership.

### How does this relate to a storyboard?

A multi-shot generation is written shot by shot: duration, framing, who is in frame, what is said. That is what a storyboard page holds, and it is how this site writes every episode - segments of numbered shots with seconds, the cast in frame and the lines.

### Do creative tools on this site render with Kling 3.0?

No. This page describes what Kling announced. The video tiers available here, with their prices and limits, are listed on the pricing page, and Kling is not among them.

### Where can I read the original material?

Kling's own blog post on Video 3.0 Omni multi-shot and native audio, and the February 2026 press release. Both are linked in the Sources section at the end of this page.

Updated November 3, 2026


---

# Kling 4.0: New Features Announced, and What Changes From 3.0

URL: https://scenemixer.com/guides/kling-4-0-new-features/
Kling 4.0 was announced on September 28, 2026: 3-30 second takes, 10 keyframes, 4K, Omni Reference. The announced specs versus 3.0, and what is missing.

# Kling 4.0: What Was Announced, and What Changes From 3.0

Quick answer: Kling 4.0 is the next generation of the Kling video model, announced on September 28, 2026 together with a lighter Kling 4.0 Flash. On paper a single generation grows from the 3–15 seconds of Kling 3.0 to 3–30 seconds, resolution is announced up to 4K, control moves to a first frame, a last frame and up to 10 keyframes placed anywhere on the timeline, and Omni Reference accepts up to 15 references per task. What has not happened is a release: as of October 15, 2026 we found no report of the full version going live, the API is not open, and no model IDs or prices have been published. Every specification on this page is the officially announced one, and the sources are listed at the end.

## What was announced, and when

The still above is from Neon Debt, an episode this site took from script to composite.

On September 28, 2026 Kling announced two models: Kling 4.0 and Kling 4.0 Flash. The announcement is published on Kling's own model-release page; the specification tables below restate it, and each figure can be traced to the sources at the end of this page.

The two models did not arrive together in the same form. Kling 4.0 Flash opened the same day as an early-access preview for a limited batch of creators (accounts differ on exactly who qualifies, so we do not name a condition here). The full Kling 4.0 was described in the announcement as arriving in October; no day was given.

That is the state we found when we checked on October 15, 2026: no release announcement for the full model, no open API, no published model IDs, and no published pricing for either version. If all of that changes, the first place it should show up is Kling's own release page — not a roundup like this one.

## The announced specifications

This is everything the announcement stated about the full Kling 4.0, in one place. Where a feature was announced as coming later, it says so.

| Announced | Kling 4.0 |
| Clip length | 3–30 seconds in one generation; extension to a maximum of 2 minutes is announced as coming soon |
| Resolution | 720p, 1080p and 4K |
| Colour | 10-bit HDR at 1080p and 4K — announced as coming soon, not available at announcement |
| Aspect ratios | 21:9, 16:9, 1:1, 9:16 and Auto |
| Frame control | First and last frame, plus up to 10 keyframes at any position on the timeline |
| References | Omni Reference: up to 15 references per task — at most 10 images, at most 5 videos (30 seconds combined), at most 7 subjects of which up to 3 may be video subjects, plus a voice |
| Audio | Stereo, with tighter lip sync; spoken lines and on-screen text in 9 languages, with accents and dialects |
| Video editing | Up to 5 input clips (one main), each 3–30 seconds, 30 seconds in total; a re-edit can change a single expression, camera, background or one spoken line and leave the rest untouched |
| Prompt length | Up to 8,000 tokens |

Those are the officially announced numbers. Nothing in the table is a measurement: nobody outside Kling has run a released model yet, so treat them as the specification the launch is measured against.

## What a keyframe is, and why ten of them matter

The frame controls are the most consequential change for anyone writing shot by shot. A first frame pins where a shot starts and a last frame pins where it ends; the model moves the action between those two points. Kling 3.0 already had first-and-last-frame mode.

Keyframes go further: they are additional pinned frames placed at chosen moments inside the clip. Up to 10 of them were announced. In practice this lets you write a shot the way a storyboard reads — a character raises the letter at second 4, drops it at second 9, turns to the door at second 14 — instead of describing the whole movement in prose and hoping the take lands on each beat.

For a series, the difference is mostly about landing beats: the moments a cut depends on. A take that has to end on a specific pose, a prop that has to change hands on cue, a reveal that has to happen at a written second. Keyframes are how the announced 4.0 gives those moments a fixed time.

## What changes from Kling 3.0

Kling 3.0 and 3.0 Omni were announced in February 2026 and are documented on Kling's own blog. The table compares what each generation announced; the 3.0 column is what that release actually stated, not what it later added.

| | Kling 3.0 (announced February 2026) | Kling 4.0 (announced September 28, 2026) |
| One generation | 3–15 seconds | 3–30 seconds, extension to 2 minutes coming soon |
| Shot control | Up to 6 shot changes in one generation; per-shot duration, framing, angle, content and camera movement | First and last frame plus up to 10 keyframes at any position on the timeline |
| References | Element reference for subjects, multi-image reference, video element reference | Omni Reference: up to 15 per task across images, videos, subjects and a voice |
| Audio | Audio and video in one pass; dialogue in English, Chinese, Japanese, Korean and Spanish; accents and Chinese dialects; lines assigned per speaker | Stereo, tighter lip sync, 9 languages for lines and on-screen text |
| Editing | Not part of the 3.0 announcement | Up to 5 input clips; single-expression, single-line, camera or background re-edits |

The pattern is that 4.0 turns dials 3.0 had already introduced. Longer takes, more pinned moments, more references, more languages, plus an editing layer that 3.0 did not announce at all.

## 4.0 and 4.0 Flash are not the same release

The second model announced on September 28 is Kling 4.0 Flash, positioned for high-volume creation: clips of 3–20 seconds, 720p output, 8-bit SDR, and the text-to-video, first-frame and Omni Reference modes. It does not get first-and-last-frame or multi-keyframe control in the announcement, and its audio and editing capabilities were not stated.

Flash went to a limited group of creators first. Which one suits a given production is a real question with a short answer and a long one — Kling 4.0 vs Kling 4.0 Flash works through it.

## What is announced but not here yet

Three items in the announcement are explicitly later work, and it is worth keeping them separate from what exists:

- Extension to 2 minutes. Announced as coming soon. Until it ships, the long-take ceiling in the announcement is 30 seconds.
- 10-bit HDR at 1080p and 4K. Also announced as coming soon. The resolution list itself (720p, 1080p, 4K) is not marked that way.
- The full model's release itself. Announced for October 2026 with no date; as of October 15 there was no report of it going live.

Two more things are simply unknown rather than promised: pricing — nothing was published for either version — and the API, which was not open at the time of writing. Anyone quoting a per-second price for Kling 4.0 today is quoting a guess.

## How to read this page, and announcements like it

Everything above is the officially announced specification. We did not run the model, because there is no released model to run, and we did not repeat measurements from other sites as if they were tests.

That is also the way to read the next announcement you find: check whether the numbers come from the maker's own release notes or from a roundup of them, and check the date on whatever you are reading. Specifications change between announcement and release, and "coming soon" items have a habit of slipping past the launch.

Side by side with this site's own pipeline, the same announcement is covered on our Kling 4.0 page, with the lighter model on Kling 4.0 Flash.

## What a 30-second, keyframed generation means for series work

Strip away the version numbers and the announcement is about length and control per take: a single generation can now cover a beat that used to need two or three, and the moments inside it can be pinned.

Series work has wanted exactly that for a while. A short-drama episode is a stack of segments, and every segment joins the next one at a cut, so the value of a longer take is not the length itself — it is fewer joins to police, and more of the performance kept in one piece.

The other half of the working method does not change with any version number: the story is still decided in the script and the shot list. SceneMixer turns a script into an episode outline, then a storyboard where every shot carries its seconds, the cast in frame and the lines to be spoken, then renders those shots segment by segment and joins them into one file per episode. How to make an AI short drama walks through that pipeline end to end, and character consistency covers the reference-image layer that keeps a cast recognisable across episodes — the same problem Omni Reference is built for.

Video tiers on this site are chosen per segment and priced per second, with the price printed on the generate button before anything is charged; the current list and rates are on the pricing page. When a model like Kling 4.0 reaches general release, it is the kind of thing that gets added as a tier; until then this page stays a specification page, not a how-to.

## Sources

- Kling: 4.0 model release page (kling.ai) — the announcement itself; the specification tables above restate it, checked October 15, 2026.
- Runware: Kling 4 specification table — a restatement of the announced specs (clip length, resolution, keyframes, Omni Reference limits, audio, editing).
- OrcaRouter: Kling 4.0 release-date roundup (October 4, 2026) — release status: full version still to come, Flash in limited preview.
- Kling: Video 3.0 Omni multi-shot and native-audio guide (kling.ai blog) — the Kling 3.0 column in the comparison.

Kling is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Specifications are quoted as announced and may change before release.

| Version | Announced | One generation | Resolution |
| Kling 3.0 / 3.0 Omni | February 2026 | 3–15 seconds, up to 6 shot changes | Not stated in the 3.0 announcement |
| Kling 4.0 Flash | September 28, 2026 | 3–20 seconds | 720p, 8-bit SDR |
| Kling 4.0 | September 28, 2026 | 3–30 seconds (2-minute extension coming soon) | 720p, 1080p, 4K (10-bit HDR coming soon) |

## FAQ

### Is Kling 4.0 out yet?

Not as far as we could establish. The full version was announced on September 28, 2026 for October, and as of October 15, 2026 we found no release report, no open API and no published pricing. Kling 4.0 Flash opened to a limited group of creators on announcement day.

### How long can a Kling 4.0 clip be?

3 to 30 seconds in one generation, per the announcement. An extension to a maximum of 2 minutes was announced as coming soon, not as available.

### What can Omni Reference take?

Up to 15 references per task, as announced: at most 10 images, at most 5 videos adding up to 30 seconds, at most 7 subjects (up to 3 of them video subjects), plus a voice.

### What is the difference between a keyframe and a first or last frame?

First and last frame pin the two ends of a clip. Keyframes are additional pinned frames at chosen moments inside the clip; Kling 4.0 was announced with up to 10 of them. They are how a shot can land on a written beat.

### Does Kling 4.0 have native audio?

The announcement describes stereo audio with tighter lip sync, and dialogue and on-screen text in 9 languages with accents and dialects. Kling 3.0 already generated audio and video in one pass in five languages.

### Can Kling 4.0 edit a video I already have?

The announcement describes video editing with up to 5 input clips (one main), each 3-30 seconds and 30 seconds in total, including re-edits that change one expression, camera, background or single line. This was in the announcement, not something we have seen running.

### How much will Kling 4.0 cost?

Unknown. No pricing was published with the announcement for either version, and any per-second figure you see elsewhere today is not an official number.

### Is Kling 4.0 available through an API?

Not at the time of writing. The model-release page covers the release itself; the API was not open when we checked on October 15, 2026.

### Where should I check the release status?

Kling's own release page, kling.ai/dev/model-release/kling-4. Roundups of it, including this one, lag the source by definition; we date our checks and list our sources so you can see exactly how current they are.

### Does this page describe how to use Kling 4.0 on this site?

No. This page is about what Kling announced. This site's video tiers, their prices and their limits are listed on the pricing page, and no Kling tier is among them.

Updated October 15, 2026


---

# Kling 4.0 vs Kling 4.0 Flash: Which One to Choose

URL: https://scenemixer.com/guides/kling-4-0-vs-kling-4-0-flash/
Both were announced on September 28, 2026. Flash: 720p, 20-second takes, built for volume. Full 4.0: 4K, 30 seconds, keyframes. Announced specs compared.

# Kling 4.0 vs Kling 4.0 Flash: How the Two Announced Versions Differ

Quick answer: Kling 4.0 and Kling 4.0 Flash were announced together on September 28, 2026 and split the work by volume. Kling 4.0 Flash is the lighter one: takes of 3–20 seconds, 720p output only, 8-bit SDR, text-to-video, first-frame and Omni Reference modes, and no first-and-last-frame or multi-keyframe control in the announcement. The full Kling 4.0 goes to 3–30 seconds and 4K, adds first and last frame plus up to 10 keyframes, up to 15 Omni References, stereo audio in 9 languages and a video-editing pass. Flash opened to a limited group of creators on announcement day; the full version was announced for October with no date, and as of October 15, 2026 neither price had been published. Everything below is the officially announced specification.

## Two models, one announcement

The still above is from The Last Envelope, a series episode produced end to end on this site.

Kling announced both models on September 28, 2026 on its model-release page. They are not two sizes of the same release: the announcement positions Kling 4.0 Flash as the version built for high-frequency creation and the full Kling 4.0 as the flagship of the generation. Flash opened the same day as an early-access preview for a limited batch of creators; the full version was announced for October with no day named.

As of October 15, 2026 we found no report of the full version going live, no open API, and no published pricing for either model. So the honest comparison today is a comparison of announced specifications — which is still enough to answer the practical question, because the two versions were announced with different jobs.

## The announced specifications, side by side

Where the announcement does not state a figure for Flash, this table says so rather than filling the gap.

| | Kling 4.0 Flash (announced) | Kling 4.0 (announced) |
| Take length | 3–20 seconds | 3–30 seconds; extension to 2 minutes coming soon |
| Resolution | 720p only | 720p, 1080p and 4K |
| Colour | 8-bit SDR | 10-bit HDR at 1080p and 4K, announced as coming soon |
| Frame control | First-frame image-to-video; no first-and-last-frame and no multi-keyframe control in the announcement | First and last frame, plus up to 10 keyframes anywhere on the timeline |
| References | Omni Reference, limits not announced | Omni Reference: up to 15 per task — 10 images, 5 videos (30 seconds combined), 7 subjects (up to 3 video subjects), plus a voice |
| Audio | Not stated in the announcement | Stereo with tighter lip sync; lines and on-screen text in 9 languages with accents and dialects |
| Editing | Not stated in the announcement | Up to 5 input clips (one main), 30 seconds in total; re-edit a single expression, camera, background or line |
| Status | Early access for a limited group of creators from September 28, 2026 | Announced for October 2026, no date; not released as of October 15, 2026 |

Read the blanks as blanks. “Not stated” means the announcement did not cover it — not that Flash cannot do it. Specifications like that are usually confirmed at general release, and that is the moment to re-check this table.

## What 20 seconds versus 30 seconds changes

The length gap looks small and is not. A take is one continuous generation: everything inside it keeps its own continuity for free, and every place you have to join two takes is a place where framing, light or a face can drift.

Take a 90-second vertical episode. On a 30-second ceiling the beat structure fits in three to four takes; on 20 seconds it is five or more, and the difference shows up as extra joins to check. For shots that are simply long — a slow push-in, a walk-and-talk, a door opening onto a reveal — the 30-second ceiling is the difference between one take and an edit.

Neither number is a promise about quality inside the take; longer generations are harder, not easier. It is a production-arithmetic difference, and it favours the full version when you are shooting dialogue scenes.

## What 720p versus 4K changes

Flash renders at 720p. That is a real, usable resolution for a phone feed, and a poor one for a television, a projector, or any edit that crops in — cropping throws pixels away, and at 720p there are few to spare.

The full version was announced at 720p, 1080p and 4K, with 10-bit HDR at the two higher settings marked as coming soon. HDR is the quieter item of the two: it changes how highlights and graded colour survive, which matters for night exteriors and stylised interiors.

The practical split follows the screen: if your destination is a vertical feed, 720p covers it and Flash's price-per-second position is the argument. If any deliverable is watched on a large screen, treat Flash as the draft and the full version as the master.

## The control gap is bigger than the length gap

The announcement gives the full version two control layers that Flash does not get: keyframes — up to 10 pinned frames at chosen moments inside the take — and a video-editing pass that takes up to five input clips and can rewrite one expression, one camera move, one background or a single line while leaving the rest untouched.

Those are the features that make a second attempt cheap. Without pinned beats, the way to fix a take is to regenerate it and hope; with them, you change the one moment that missed. If your work is iterative — you know you will dial scenes in over several passes — this layer is why the full version is the one to plan around, assuming the announced capabilities survive to release.

## Why Flash exists, and what it is for

“Built for high-frequency creation” is the announcement's own phrase, and it describes a real workflow: many takes, quickly, at a size you can review on a phone. Storyboard previews, timing checks, animatics before the expensive pass, or a steady output of social cuts that do not need a 4K master.

Sketching is not a lesser craft. Most of the cost of a series is in scenes you shoot more than once; a cheap, fast take you can look at is worth more than a beautiful one you were afraid to try. The catch is the ceiling: at 20 seconds and 720p, with no keyframe or editing layer announced, Flash cannot finish a shot that needs to land a beat or survive a crop.

## How to choose

- Phone-first, high-volume, tolerant of cuts: Flash. Shorter takes mean more of them, and the announced mode list covers text-to-video, first-frame and references.
- Dialogue scenes and long takes: the full version's 30-second ceiling halves the joins in a talking scene.
- Any large-screen deliverable: the full version. 4K, and HDR when it arrives, is the only announced path above 720p.
- Work you know you will re-edit: the full version, for keyframes and the editing pass.
- Draft, then finish: if both become available at sane prices, Flash for passes 1–3 and the full version for the keepers is the classic split — the same logic as drafting low and finishing high on the tiers here. Draft at 480P, finish at 1080P walks through that pattern on this site.

## What is still unknown

Five things were not settled by the announcement, and each of them can change the choice:

- Price. Nothing published for either model. A per-second rate is the single biggest input to a draft-versus-finish decision, and it does not exist yet.
- Release dates. Flash is in a limited creator preview; the full version was announced for October with no day, and as of October 15 had not been reported live.
- API access. Not open at the time of writing.
- Flash's reference limits. Omni Reference is listed for Flash, but the numbers were not announced.
- Flash's audio and editing. Not stated either way. Silence is not a no, but it is also not a feature you can plan around.

Until those land, the two models are a specification question, not a purchase question. Nothing on this page is a measurement; nobody outside Kling has run a released version.

Each version also has its own page next to this site's pipeline: Kling 4.0 and Kling 4.0 Flash.

## What this means for a series pipeline

Whatever happens with these two versions, the shape of the work does not move: a story is broken into episodes, episodes into shots, shots into takes; takes get joined; the joins are where attention goes.

That is how SceneMixer is built: a script becomes an episode outline, the cast is drawn once as reference sheets, every shot is written into a storyboard with its seconds, the people in frame and the line to say, and segments render one at a time and composite into one file per episode. Tiers are chosen per segment and priced per second, with the price shown on the generate button; the current list is on the pricing page.

Longer takes and pinned beats — the direction this announcement points — are exactly what episodic work asks a model to do. When a model like this reaches general release, it is the kind of thing that gets added as a tier here. Until then, this page stays a comparison of what was announced.

## Sources

- Kling: 4.0 model release page (kling.ai) — the announcement of both models; checked October 15, 2026.
- Runware: Kling 4 specification table — the announced specs of the full version and Flash, including the 20-second and 720p limits.
- Pixmax: Kling 4 release overview — a restatement of the announcement and of the Flash positioning.
- OrcaRouter: Kling 4.0 release-date roundup (October 4, 2026) — release status as of early October.

Kling is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Specifications are quoted as announced and may change before release.

| | Kling 4.0 Flash | Kling 4.0 |
| Take length | 3–20 seconds | 3–30 seconds |
| Resolution | 720p | 720p, 1080p, 4K |
| Keyframes and editing | Not in the announcement | 10 keyframes, 5-clip editing |
| Built for | High-frequency creation | Flagship quality and control |

## FAQ

### What is the main difference between Kling 4.0 and 4.0 Flash?

Take length and resolution, as announced: Flash does 3-20 seconds at 720p only, the full version 3-30 seconds at up to 4K, plus keyframes, more references, stereo audio and a video-editing pass that Flash's announcement does not include.

### Is Kling 4.0 Flash available now?

It opened on September 28, 2026 as an early-access preview for a limited group of creators. We have not confirmed general availability; check Kling's release page for the current state.

### Is the full Kling 4.0 released?

Not as far as we could establish by October 15, 2026. It was announced for October with no specific date, and no pricing or API access had been published.

### Can Flash do keyframes?

Not in the announcement. Flash lists first-frame image-to-video but no first-and-last-frame and no multi-keyframe control. Keyframes, up to 10, are announced for the full version.

### Does Flash have audio?

The announcement does not say. It may simply not have been covered; treat it as unknown until release notes state it.

### Which one should I use for phone-first vertical drama?

By the announced specs, Flash's 720p and 20-second takes cover a vertical feed, and its positioning is high-volume work. If you need takes longer than 20 seconds or any large-screen output, that is the full version's side of the line.

### Will Flash be cheaper?

It is positioned as the lighter, high-frequency model, but no prices were published for either version, so the size of the gap is unknown.

### Do announced specifications change before release?

Often. Resolution lists and limits can shift, and items marked coming soon can slip past the launch. We date our checks and link our sources so you can see how current this page is.

### Does this site have either model as a tier?

No. The video tiers here, with their per-second prices, are listed on the pricing page. Neither version of Kling is one of them; this page covers what Kling announced.

Updated October 15, 2026


---

# Kling Version History: 1.0 to 4.0, Release by Release

URL: https://scenemixer.com/guides/kling-version-history/
Kling 1.0 arrived in June 2024; 4.0 was announced on September 28, 2026. What each generation added, told from released material, with third-party dates marked.

# Kling Version History: From 1.0 in 2024 to the 4.0 Announcement

Quick answer: the Kling line runs 1.0 (June 2024) → the 2.x series, where 2.1 split into Standard, Pro and Master and 2.6 added native audio → the 3.0 generation announced in February 2026 with multi-shot generation, native audio and element references (Video 3.0, Video 3.0 Omni, Image 3.0, Image 3.0 Omni) → 4.0 and 4.0 Flash, announced on September 28, 2026, with takes announced up to 30 seconds and up to 10 keyframes. The early entries are dated by third-party timelines; 3.0 and 4.0 are dated by Kling's own releases. The whole timeline, source by source, is below.

## How to read this history

The still above is from Starfall Command, an episode produced on this site; whatever model rendered its segments, the version history of that model will keep moving.

Versions age quickly into folklore. Most “history of Kling” pages mix dates from press releases, store listings and memory, and by the time you read them the newest entry is already stale. This page separates the two kinds of evidence:

- Maker's own material for the 3.0 and 4.0 entries: a press release and Kling's own pages.
- Third-party timelines, marked as such, for 1.0 and the 2.x series, where we could not reach a first-party page.

Where a date is third-party, the text says so. Where something could not be confirmed, it is in the “not verified” section rather than in the timeline.

## The timeline

| Version | When | What it added | Evidence |
| Kling 1.0 | June 2024 | The first public Kling model | Third-party timeline (fal.ai) |
| 2.1 | 2.x series | Split into Standard, Pro and Master editions | Third-party roundup (Atlas Cloud) |
| 2.6 | 2.x series | Native audio synchronisation | Third-party roundup (Atlas Cloud) |
| 3.0 / 3.0 Omni | February 2026 | Multi-shot generation (up to 6 shots in one 3–15 second take), native audio in five languages, element references, Character Identity 3.0 | Press release (February 6, 2026) and Kling's own blog |
| 4.0 / 4.0 Flash | Announced September 28, 2026 | Takes announced up to 30 seconds, 4K, up to 10 keyframes, Omni Reference with up to 15 references; Flash as the lighter 720p, 20-second model | Kling's own model-release page |

## Kling 1.0, June 2024

The line starts with Kling 1.0 in June 2024. This date comes from third-party timelines — fal.ai's explainer among them — rather than a first-party page we could still reach, so treat the month as well-attested rather than exact to the day.

What matters about 1.0 for reading later announcements is the baseline: short clips from text or a single image, with no audio. Every capability discussed on this page — multi-shot takes, synchronised dialogue, consistent characters, keyframes — is a later addition to that starting point.

## The 2.x series

The 2.x generation is where Kling started splitting a model into editions and adding sound. Two details, both from third-party roundups (Atlas Cloud's guide):

- 2.1 arrived in Standard, Pro and Master editions — the first sign of the tiering that later turns up everywhere in AI video.
- 2.6 added native audio synchronisation — sound generated with the picture rather than added after it.

We did not find first-party pages dating either release, so this section stays at that level of detail rather than repeating exact dates we cannot trace. The audio work matters historically because it is the thread that runs into 3.0's per-speaker dialogue and 4.0's stereo.

## 3.0 and 3.0 Omni, February 2026

Kling 3.0 was announced in February 2026 — social posts on February 4 and a press release dated February 6 — and it was the release that turned Kling from a clip generator into something closer to a shot-sequence generator.

The announced generation, all four models (Video 3.0, Video 3.0 Omni, Image 3.0 and Image 3.0 Omni), included:

- Multi-shot takes: a 3–15 second generation with up to 6 shot changes, each shot carrying its own duration, framing, angle, content and camera movement, and an AI Director handling shot-reverse-shot and cross-cutting.
- Native audio: dialogue generated with the video in English, Chinese, Japanese, Korean and Spanish, with accents and per-speaker assignment.
- Reference modes: element (subject) reference, multi-image reference and video element reference, with Character Identity 3.0 for keeping a face and Element Voice Control for binding a voice to it.
- On-screen text that stays legible in motion.

The details, with sources, are in Kling 3.0 multi-shot and native audio.

## 3.0 Turbo: reported, not confirmed

Some third-party blogs describe a further 3.0 Turbo or Omni-engine release dated to mid-2026. We could not confirm that release on any first-party page, so it is deliberately not in the timeline above and carries no date here. If Kling's own release pages pick it up, it belongs in that table; until then it stays a claim from roundups.

## 4.0 and 4.0 Flash, announced September 28, 2026

The current generation was announced on September 28, 2026: Kling 4.0 and the lighter Kling 4.0 Flash. The headline numbers, all as announced:

- Longer takes: 3–30 seconds in one generation, with extension to a maximum of 2 minutes marked as coming soon. Flash: 3–20 seconds.
- More resolution: 720p, 1080p and 4K, with 10-bit HDR at the top two marked coming soon. Flash: 720p only.
- Frame control: first and last frame plus up to 10 keyframes at positions on the timeline. Flash's announcement does not include them.
- References: Omni Reference, up to 15 per task across images, videos, subjects and a voice.
- Audio and editing: stereo with tighter lip sync, 9 languages, and an editing pass over up to 5 input clips.

Flash opened to a limited group of creators on announcement day; the full version was announced for October with no date, and as of October 15, 2026 we had found no report of it going live and no published pricing for either model.

The full breakdown is in Kling 4.0: what was announced, and the two versions are compared in Kling 4.0 vs Kling 4.0 Flash.

## Why versions matter when you follow a guide

Advice written for one generation ages badly against the next. A guide that says “up to 15 seconds” is describing 3.0; one that says “up to 30 seconds” is describing the 4.0 announcement. Both were correct when written, and neither tells you what is running today.

Three habits keep you out of dated advice: check the version number in the text, check the date on the page, and when a page quotes specs, look for whether it says announced or measured. For anything past the announcement stage, the maker's own release page is the only source that settles it.

## What this page does not claim

- Exact dates for 1.0 and the 2.x releases. The third-party timelines agree on the shape (1.0 in mid-2024, 2.1 and 2.6 in the series that followed) but we did not reach first-party pages for them.
- 3.0 Turbo. Reported by third parties, not confirmed by Kling; no date given here.
- Anything about 4.0 past the announcement. No measured specs, no pricing, no release confirmation — nothing of that kind exists yet, as far as we could find.

## Sources

- Kling: 4.0 model release page (kling.ai) — the September 28, 2026 announcement.
- Kling: Video 3.0 Omni multi-shot and native-audio guide (kling.ai blog) — the 3.0 generation's capabilities.
- Bernama: Kling 3.0 press release (February 6, 2026) — the 3.0 launch and model list.
- fal.ai: What is Kling 4.0 (explainer) — dates the 1.0 release to June 2024.
- Atlas Cloud: Kling AI guide — the 2.x details (2.1 editions, 2.6 native audio), third-party.
- Runware: Kling 4 specification table — the announced 4.0 and Flash specifications.

Kling is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Where a date rests on third-party material, this page says so; specifications are quoted as announced.

| Year | Version | Milestone |
| 2024 | Kling 1.0 | First public release, June 2024 |
| 2024–2025 | 2.1 / 2.6 | Standard, Pro and Master editions; native audio synchronisation arrives |
| 2026 (February) | 3.0 / 3.0 Omni | Multi-shot generation, native audio, element reference, Character Identity 3.0 |
| 2026 (September 28) | 4.0 / 4.0 Flash | Announced: up to 30-second takes, 4K, 10 keyframes, Omni Reference |

## FAQ

### When did Kling start?

Kling 1.0 dates to June 2024 in third-party timelines such as fal.ai's explainer; we did not find a first-party page dating it, so treat the month as the reliable part.

### What did the 2.x series add?

Two things from third-party roundups: 2.1 split the model into Standard, Pro and Master editions, and 2.6 added native audio synchronisation - sound generated with the picture rather than dubbed after.

### What came with Kling 3.0?

Multi-shot generation (up to 6 shot changes in a 3-15 second take), native audio in five languages with per-speaker lines, element and multi-image references, Character Identity 3.0 and on-screen text that stays legible. It was announced in February 2026.

### Is there a Kling 3.0 Turbo?

Some third-party blogs date a 3.0 Turbo release to mid-2026. We could not confirm it on a first-party page, so this page does not date it or list it in the timeline.

### When was Kling 4.0 announced?

September 28, 2026, together with Kling 4.0 Flash. The full version was announced for October with no specific date; as of October 15, 2026 we found no report of it going live.

### What is the newest version I can actually use?

Check Kling's own release page. Announced and released are different states, and this page dates every claim to the day we checked it rather than guessing what is live today.

### Why do different sites give different numbers for Kling?

Because they are describing different generations, or repeating pre-release specifications as if they were measurements. A 15-second figure is 3.0; a 30-second figure is the 4.0 announcement.

### Does this site use Kling models?

No. This page is a version history. The video tiers available here, with prices and limits, are listed on the pricing page, and Kling is not among them.

Updated October 15, 2026


---

# Leads, Supporting Cast and Voice-Only Characters | SceneMixer

URL: https://scenemixer.com/guides/leads-supporting-and-voice-only-characters/
Who the reading casts, what the role badges mean, what an Off-screen voice-only character is, who becomes background people, and how to add or remove someone.

# Leads, Supporting Cast and Voice-Only Characters

Quick answer: The reading puts everyone who acts or speaks in the Characters tab and gives each a role badge: Lead, Supporting or Villain. A part that is heard but never seen — a narrator, a voice on the phone — is marked Off-screen: it gets a voice and no image. Crowds and passers-by are not cast at all; they are background people, declared per segment. Roles are labels you can change; anyone can be added with Add or removed from the card.

Newer storyboards. In storyboards written or regenerated since late October 2026 a crowd is described in the shot rather than declared in a header clause; see background people in a scene. Who each shot shows is worked out by the system and listed under the segment as In frame (notes under a segment).

Open a finished episode and count the characters in the left panel against the people you see on screen.

## Who gets cast

Anyone who acts, speaks, or is handed something. A courier with one line is cast; three hundred wedding guests are not. The reading makes that call from the script, and you can overrule it in step 2: Add creates a character it left out; the card’s details let you delete one it should not have made. Neither costs credits; adding a character the script did not produce needs a paid plan.

## Lead, Supporting, Villain

The badge on a character card is the role the reading assigned. Leads get the most detailed descriptions and are the faces the drama is built around; supporting cast and villains are described enough to be recognisable every time. The badge is a label: open the card and change it under Role if the reading got it wrong. It does not change the price of anything.

## Off-screen: voice only

A character whose card says Off-screen and Voice only — no image generated is heard but never seen: the narrator, the voice on the other end of the call, the announcer. It is skipped by Generate All and never appears in a shot. Give it a voice like any other character. In the storyboard its lines carry (voice-over) or (off-screen) and it is not named in the picture. Writing its lines.

If a character marked off-screen is in fact meant to be seen, edit the card: it then needs an image like everyone else.

## Background people

The crowd is not cast. It is declared per segment, at the end of the header’s Set dressing line: Background people: guests seated in the pews, murmuring. That puts them in every shot of the segment without giving any of them a face of their own. The segment header.

## How many

A short drama works best with a handful of faces. Every cast member needs an image before video, and shots with four or more named people are where faces begin to blend. If the reading cast twenty people for a two-hander, delete the ones who only pass through and let them be background people.

## Voices for each

Every character, on screen or off, gets one voice in step 2 — matched, picked from the library, or uploaded. A character with no voice description cannot be matched; fill it in first. Character voices.

| Kind | Image | Voice | Named in shots |
| Lead / Supporting / Villain | Yes | Yes | Yes, with @ |
| Off-screen (voice only) | No | Yes | No — lines carry (voice-over) |
| Background people | No | No | No — declared in Set dressing |

## FAQ

### Can I change a character's role?

Yes. Open the card and change Role. It is a label; it changes no price.

### What is an Off-screen character?

One that is heard but never seen. It gets a voice and no image, and its lines are written as voice-over.

### The reading cast someone who is only a passer-by.

Delete the character in step 2 and let them be background people in the segments that need a crowd.

### The reading missed a character.

Press Add on SceneMixer's Characters tab, describe them, and generate an image. Adding is a paid-plan feature.

### Do supporting characters cost more or less?

Neither. Every image is priced the same on SceneMixer, and roles change nothing about price.

Updated November 3, 2026


---

# Learn from a Sample Project: Open, Browse, Play, Copy | SceneMixer

URL: https://scenemixer.com/guides/learn-from-a-sample-project/
Featured Projects are finished dramas you can open read-only: browse the summary, the sheets, the storyboard and the editor, play and share, and see what good looks like.

# Learning from a Sample Project

Quick answer: A project under Featured Projects on the home page opens straight into its first episode in the editor, read-only, with a banner: This is a showcase project, view only. There you see the storyboard text, the segments with their seconds and the clips; play, go full screen, download, share. Press Back to reach the episode list in step 3, and from there the cast, sets and props with their sheets in step 2 and the summary in step 1. Nothing can be changed or generated, so browsing costs nothing. Use it to see what a parsed summary, a good sheet and a storyboard look like before you make your own.

Open a sample now: an episode in the editor.

## Where

Featured Projects below your own cards on the home page, and on the landing page. They are in the language of your interface where a version exists. Signed out, they still open; signed in, they appear the same.

## Step 1: the summary

Read the title, protagonist, era, adaptation strategy and the episode list. This is what a read produces from a script; compare it with your own after your first read. Checking a summary.

## Step 2: the sheets

Click any card to open it: the sheet, the appearance and clothing text, the voice. Notice how specific the descriptions are — that specificity is why the face holds across the drama. Writing a description.

## Step 3 and the editor

Open an episode. The storyboard text on the left, the timeline below, the player on the right. Click a segment: read how a shot is written, how long it is, who is in frame. Play the clip and match it to the text. Version history is not shown on samples; each segment shows its chosen take. The editor.

## What you cannot do

Edit text, generate, add assets, rename, delete. Any such button shows Showcase projects cannot be modified. What you can do: play, full screen, download, share a segment or the assembled episode.

## What to copy

The shape of a description; the length of a segment; how much a line of dialogue says; how a storyboard names the cast and props. Then paste your own script and compare your first read with the sample’s. Upload and parse.

| In a sample | You can | You cannot |
| Step 1 | Read the summary and episode list | Edit |
| Step 2 | Open every sheet and read its description | Generate or upload |
| Step 3 / editor | Read the storyboard, play, download, share | Edit, generate, see version history |
| The card | Open it | Rename or delete |

## FAQ

### Do samples cost credits to open?

No. Nothing can be generated in them, so browsing is free on SceneMixer.

### Can I copy a sample into my account?

No. Open it to learn from; make your own from your own script.

### Why is version history missing?

In SceneMixer, samples show one chosen take per segment; history is hidden on them.

### Can I share a sample's video?

Yes. Share works on samples like on your own videos.

### Are samples in my language?

Where a version exists, yes; otherwise the original.

Updated November 3, 2026


---

# Lip Sync or Re-perform: Which Redub to Pick for an AI Video Clip

URL: https://scenemixer.com/guides/lip-sync-or-re-perform-which-redub-to-pick/
Redub offers two ways to change a clip's dialogue: Lip sync keeps the picture, Re-perform acts it again. Which to pick, and what to do if it looks or sounds off.

# Lip Sync or Re-perform: Choosing How to Redub a Clip

Quick answer: Lip sync is the default: the picture stays frame for frame, every line is voiced again and the mouths are moved to match; the original ambience is dropped. It costs 1 credit per second of clip and works on clips of any length. Re-perform plays the clip again with the new lines, so the delivery and timing fit the words and the sound is a full new performance, but the picture may shift a little. It costs 2× the per-second price of the clip’s tier and is offered only for clips of 15 seconds or less.

Same 10-second clip on Wan 3.0 · 480P: Lip sync costs 10 × 1; Re-perform costs 10 × 20. Try Lip sync first.

## How Lip sync works

The site voices each line of the segment in the voice shown on its row, lays them out one after another from the start of the clip with short pauses in between, and uses that as the clip’s new soundtrack. The picture is kept, and the speakers’ mouths are moved to fit the new audio. Two consequences follow. The pauses of the original are not kept: lines that were spread across the clip now come one after another. And the soundtrack holds only the voices: the take’s own music and ambient sound are gone (the episode’s score is a separate layer, mixed back under the new version automatically).

It looks best when faces are turned toward the camera. On faces in profile, or when a speaker is small in the frame, the result can look or sound less natural; listen to the voice before you run it.

## How Re-perform works

Re-perform treats the new lines as a new take: the clip is made again from itself with the lines replaced and the chosen voices as the speakers’ voice samples. Because the performance is made fresh, the timing and delivery follow the new words and the soundtrack is a complete take, ambience included. The framing and positions follow your clip, but faces and small movements can differ a little. It uses the same machinery as AI visuals, which is why it has the same price and the same 15-second limit.

## Choosing

- A word or two changed, lines already back to back: Lip sync.
- A clip longer than 15 seconds: Lip sync (Re-perform is greyed out).
- The picture must not change at all: Lip sync.
- The take’s own ambience matters, or the new line needs a different rhythm: Re-perform. (The episode’s score comes back by itself either way.)
- The line is much longer than the old one: neither may fit; edit the storyboard, add seconds to the shot and regenerate.

## Price side by side

In SceneMixer, Lip sync costs 1 credit × the length of the whole clip in seconds. Re-perform: 2× the per-second price of the tier the clip was made on × the length. On any tier, Re-perform is many times the price of Lip sync; both prices are printed on the two mode buttons before you choose. Both are held when you press and refunded if the redub does not go through. Every fix price.

## Either way

Both write changed words back into the storyboard, both save a Redubbed version and keep the generated clip, and a segment keeps one redubbed version. Change a line after the video is made.

| | Lip sync | Re-perform |
| Picture | Every frame kept | Made again; small changes possible |
| Voices | Re-voiced, one after another from the start | A new performance |
| The take’s own sound | Dropped; the episode’s score returns on its own | New soundtrack, ambience included |
| Clip length | Any | Up to 15 seconds |
| Price | 1 per second | 2× tier price per second |
| Ad projects | Yes; in a newly scored ad, re-compositing mixes the ad’s music back under it. In an ad made before the new scoring, the clip’s own music goes with the replaced soundtrack — re-perform or regenerate that clip instead | Yes |

## FAQ

### Why is Re-perform greyed out?

The clip is longer than 15 seconds, or re-rendering is unavailable for this clip at the moment. Use Lip sync.

### Why does the lip-synced clip sound empty?

At first it holds only the voices: the take’s own ambience is dropped, and the episode’s score is mixed back under it automatically, usually within minutes. Re-perform keeps a full soundtrack from the start, on clips up to 15 seconds.

### Can I keep the original pauses with Lip sync?

No. The lines are placed one after another from the start. Re-perform follows the new delivery instead.

### The redubbed clip looks or sounds wrong. What now?

Switch back to the clip you started from in SceneMixer's Version History; it is always kept. A redub that looks off on a side-on or small speaker usually does better as Re-perform on a clip of up to 15 seconds.

### Which looks more natural?

For faces toward the camera, Lip sync is usually convincing and much cheaper. Side-on or very small speakers are harder for it; if the result looks off, try Re-perform on a clip up to 15 seconds, or regenerate the segment with the new line.

Updated November 3, 2026


---

# Logline Examples for Short Films and Series (and How to Write Yours)

URL: https://scenemixer.com/guides/logline-examples/
Logline examples for short films, dramas and series, a one-sentence formula, and how a logline becomes the one line you type into AI Create for a script and cast.

# Logline examples for short films and series

Quick answer: A logline is one sentence that says who wants what and what stands in the way. The formula that works: When [event], a [specific person] must [goal] before [stakes], but [obstacle]. In SceneMixer a good logline is also your best prompt: type it into AI Create and the script, the cast and the storyboard are drafted from it.

Take any logline below, paste it into AI Create, and read the script it drafts before you render anything.

## The formula

| Part | Write | Example |
| Trigger | What starts it | When a lighthouse hotel is left to three estranged siblings |
| Person | Someone specific | a retired chef and her two younger brothers |
| Goal | What they must do | must make it profitable in one summer |
| Stakes | Why now | or lose the only home they share |
| Obstacle | What stands in the way | but every guest knows a secret about the place |

## Examples by genre

| Genre | Logline |
| Drama | When a night-bus driver finds a lost violin, she has one run to decide whether to return it, but its owner is the one passenger she has avoided for years. |
| Thriller | A coroner who can hear the last ten seconds of a dead person's life must stop a killing in a town where everyone lies about their last night. |
| Romance | Two strangers who share a laundromat at 2 a.m. have one night before one of them leaves the city. |
| Fantasy | A clockmaker's daughter wakes a tower clock that has been silent for a hundred years and learns what time it has been keeping. |
| Comedy | A family of five has one hour to hide a goat from the landlord before the open day. |
| Revenge | A fired accountant returns to the firm as the auditor, with a week to find who forged her signature. |

Hooks and cliffhangers · AI hook generator.

## From the logline to a project

Type the logline into AI Create. Read the draft script; the characters in your logline are now a cast list with written looks and voices. Fix what is wrong, then approve the drawings and read the storyboard. A movie from a prompt · AI film script generator · The free preview.

## Check your logline

- Is the person specific, not “a man”?
- Is the goal something you can film?
- Does the obstacle come from the world, not from fate?
- Can you say it in one breath?

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Neon Nights; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the person your logline names and the one place it happens in. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Weak logline | Stronger logline |
| A hospital thriller | A coroner who hears the last ten seconds of the dead must stop a killing in a town of liars |
| A family comedy | A family of five has one hour to hide a goat from the landlord |
| A love story | Two strangers share a laundromat at 2 a.m. before one leaves the city |

## FAQ

### What is a logline?

A one-sentence summary of who wants what and what stands in the way.

### How long should a logline be?

One sentence you can say in one breath, usually 25 to 40 words.

### Can I use a logline as a prompt?

Yes. Type it into AI Create to get a draft script, cast and storyboard.

### Can AI write my logline?

AI Create drafts a script from one line; you write or refine the logline yourself.

### Do loglines need a name?

A named lead helps, but a specific description is enough.

### Is AI Create free?

Drafting and editing are free to try; new SceneMixer accounts start with 50 credits.

### Can I reuse these?

They are examples of structure; write your own with your own characters.

Updated November 3, 2026


---

# Long Scripts Are Read in Two Passes: Episodes Still Being Prepared

URL: https://scenemixer.com/guides/long-scripts-read-in-two-passes/
Scripts over the threshold are read as cast first, episodes second, with each full outline prepared when it is needed. The stages, what is ready when, and the caps.

# Long Scripts Are Read in Two Passes: What “Preparing Episodes” Means and Why Some Episodes Open Later

Quick answer: A script longer than about 10,000 characters (counted as Chinese-equivalent characters) is read in two passes: first the cast, sets and props, then the episode list with a short outline per episode. The first three episodes get their full outline in the background straight away; any other episode is prepared when you press Generate storyboard on it — the button waits a moment and then goes ahead by itself. Nothing is charged twice, and the size cap for a two-pass read is 100,000 equivalent characters.

Paste a long novel chapter set on the home page and watch the stage names in the progress bar: cast, then episodes, then preparing.

## What you see in Step 1

The reading bar runs through the same stages as any script — reading, cast, episodes — and for a long script adds a final preparing episodes stage. You can leave the page; the read carries on and picks up where it was when you come back. What each stage means and what happens if it fails: While the Script Is Being Read.

## Why two passes instead of one

A single read has a fixed budget for what it can write back, and a fifty-episode script does not fit: read in one go, the model silently delivers a fraction of the episodes. Splitting the read into cast and then episodes keeps every episode, and preparing each episode’s full outline separately keeps the detail. The parse fee on SceneMixer is charged once for the whole script, by length, whichever way it is read — Upload a Script and Parse It.

## Which episodes are ready when

After the read, the first three episodes have their full outline; the others show a shorter summary with a preparing mark in Step 3. The others are prepared when you press Generate storyboard on one: that episode and the next one are prepared first; opening an episode alone does not start it. the button shows a short wait and continues on its own once the outline is in. You never have to press twice.

## If an episode will not finish preparing

SceneMixer retries a failed preparation on its own; after repeated failures the episode is marked as prepared from its short summary instead, and the storyboard is generated from that. The result is thinner for that episode; outlines cannot be edited on the site, so add detail by editing the storyboard once it is generated. Checking the Summary, and What Re-parsing Means.

## Size caps and how characters are counted

Scripts under 10,000 equivalent characters are read in one pass; from there up to 100,000 they are read in two. Alphabetic scripts count lighter: every 3,333 characters of English, German, Spanish and other alphabetic text count as 1,000, so an English two-pass read can be up to 300,000 raw characters. The home page shows the cap for the language you paste. How the caps and the character factor work.

| Script length (equivalent characters) | How it is read | Cap |
| Under the two-pass threshold | One pass: cast, episodes and full outlines together | Up to the threshold, 10,000 |
| Over the threshold | Two passes, outlines prepared when needed | 100,000 |
| Alphabetic text, over the threshold | Same, counted at 3,333 characters per 1,000 | 300,000 raw characters |

## FAQ

### Why does episode 12 say 'preparing' when episode 1 is ready?

SceneMixer prepares only the first three in advance. Press Generate storyboard on episode 12 and it is prepared then; the button waits and continues.

### Do I pay extra for the second pass?

No. SceneMixer's parse fee is by script length, charged once.

### I pressed Generate storyboard and it just waited. Is it stuck?

SceneMixer is waiting for that episode's outline to finish preparing, then continues by itself. Give it a minute.

### Can I edit an outline?

No. Outlines cannot be edited on SceneMixer; edit the storyboard once it is generated.

### My script is over the cap. What now?

Split it into parts and parse each as its own project, or trim it. SceneMixer's home page tells you the cap for your language before you start.

Updated November 3, 2026


---

# Look Around the Set While a Shot Renders: The 360° Preview

URL: https://scenemixer.com/guides/look-around-the-set-while-a-shot-renders/
While a segment renders, the preview shows the scene’s set in 360°, with every character and camera placed. What each part means, the controls, and when it is not shown.

# Look Around the Set While a Shot Renders

Quick answer: On projects with the world model, a segment that is not generated yet, still generating or failed no longer shows a blank card in the preview. It shows the scene’s set in 360°: drag to look around, or tilt your phone. Once the positions are worked out, each character stands in the room as a figure with their name above, the numbered wedges show where each camera stands, and the line at the bottom steps through the shots. When the clip finishes while you watch, the view turns to the first camera and fades into the video. It is read-only and free: nothing you do there changes the clip or costs credits.

Open an episode in a project that has the world model, select a segment that has not been generated, and look at the preview on the right. Or drag the demo in Try it here below.

## Where the 360° preview appears

- On a computer: in the episode editor, in the preview on the right, for the segment you have selected.
- On a phone: in the editor’s preview screen.
- For segments that are not generated yet, generating, or failed. A finished segment shows its video as before.
- Only for scenes that have a 360° view. That means scenes in projects with the world model (new projects on paid plans). What a world model is.
While the segment generates, the progress shrinks to a small pill at the bottom, just above the timeline, so the set has the room. If the segment failed, its card sits at the bottom with the retry button and the error, as before.

## What you see, stage by stage

- The room on its own. Before the characters have been placed, you can look around the set but nobody is in it. People are never drawn before their positions are worked out, so an empty room means “not placed yet”, not “nobody is in this scene”.
- The people. Once placed, every character in the segment stands in the room.
- The cameras and the shots. When the shots are planned, the cameras appear and the timeline at the bottom plays through them on a loop: Camera move while people get into place, then Shot 1, Shot 2 and so on.
- The reveal. If the clip finishes while you are watching, the view swings to where the first shot’s camera looks and fades into the finished video. Open the page after it finished and you simply get the video.

## Reading the figures, cameras and timeline

- Figures. One per character, each in its own colour, with the name floating above. The pose follows the storyboard: standing, sitting, kneeling, crouching, sitting on the floor, lying down, or seated in a vehicle. The figure turns to face whoever the text says they face.
- Camera wedges. Each numbered wedge is a camera: the point is where it stands, the open side is where it looks. The frame drawn in front of it is what that shot will hold.
- The timeline. A thin line with a mark at each cut. Camera move covers people moving into place; Shot N is each shot. Hold the cursor and drag to stop at any moment; let go and it carries on looping.
The figures are deliberately simple. They show where people are and which way they face; how each person looks comes from their character image, not from the figure.

## Controls

| To do this | On a computer | On a phone or tablet |
| Look around | Drag with the mouse; let go and it glides to a stop | Drag with one finger |
| Look by moving the device | – | Tap Tilt to look, then turn the phone. iPhone asks for motion access first |
| Take a small step (sets that allow it) | W A S D or the arrow keys, or Shift + drag | Drag with two fingers |
| Pause on one moment | Hold and drag the timeline cursor | Hold and drag the timeline cursor |
Leave it alone and the view turns slowly by itself. It stops turning while the shots are playing through or while tilt is on. Tilt to look only appears on touch devices with motion sensors, and it works from the way you were holding the phone when you tapped it. Steps are limited to about half a metre from where the set was captured, so the picture stays clean.

## Try it here

This is the demo from the SceneMixer homepage: one set, four characters, four cameras. Your own segments show your own set this way while they render; in the editor the people are drawn as simple coloured figures and the cameras as numbered wedges.

Try it: drag to look around the demo set. Four characters stand in the room, the numbered cameras show where each shot is taken from, and the view moves from camera to camera along the timeline at the bottom. The editor’s preview works the same way, with simpler figures.

## Why a segment shows no 360° view

- It is finished. Finished segments show their video.
- The scene has no 360° view. Projects created before the world model launched, and projects on the free plan, use the earlier staging and show the usual card.
- It is a sample project. Samples are read-only showcases and never show the set.
- The set could not be loaded. Then the usual card comes back. Nothing is wrong with the segment; generate it as normal.

## If someone is in the wrong place

The preview is read-only: you cannot drag a figure to move it. Positions come from the storyboard text, so that is where you fix them. Say who sits where, who faces whom, and what they stand next to, then generate again once the current run has finished. Each generation is charged as usual; the earlier take stays in Version History.

How to write positions so they come out right: Write positions the world can read. To change just one shot: Edit one shot and regenerate.

| What you see | What it means |
| The room, no people | The set exists; the characters have not been placed yet |
| Figures, no cameras | People are placed; the shots are not planned yet |
| Figures, numbered wedges and a moving timeline | Everything is planned; the video is being generated from these cameras |
| The view turns and fades into a video | The clip just finished |
| The usual card, no set | Finished, a sample, a project without the world model, or the set did not load |

## FAQ

### Does looking around cost credits?

No. SceneMixer's preview is free and read-only. You pay for generating the segment, as always.

### Can I move a character by dragging the figure?

No. Positions come from the storyboard text. Change the text, then generate the segment again.

### Why is the room empty?

The characters have not been placed yet. People only appear once their positions are worked out; they are never guessed in the meantime.

### Why don’t I see it on a finished segment?

A finished segment shows its video. The set is shown only while a segment is not generated yet, generating, or failed.

### The Tilt to look button is missing. Why?

It only appears on phones and tablets with motion sensors. On an iPhone, allow motion access when asked; if you declined, reload the page and tap it again.

### Does the preview match the final video exactly?

It shows the plan the video is generated from: the room, where everyone stands and where each camera is. The video adds the real faces, light and movement, so small differences are normal.

### Why can I only step a little?

The set is captured from one point. Steps are kept to about half a metre from it so the picture does not stretch.

### My old project never shows it. Can I turn it on?

Projects created before the world model launched keep the earlier staging, and there is no switch. Start a new SceneMixer project on a paid plan to get it.

Updated November 3, 2026


---

# MAI-Image-2.6 on SceneMixer: Microsoft's Image Model, Price and Uses

URL: https://scenemixer.com/guides/mai-image-2-6-tier/
Microsoft's MAI-Image-2.6 as an image tier on SceneMixer: price per image, output size, content filter, and when to use it for an AI series' sheets and covers.

# MAI-Image-2.6 (Microsoft) for character sheets and covers

Quick answer: MAI-Image-2.6 is Microsoft AI’s image model, and on SceneMixer it is an image tier: it draws your series’ character, location and prop sheets, state images and promo cover posters for 31 credits per image (about $0.19 at the lowest credit price). Two things set it apart here. It draws at about one megapixel — fewer pixels than the other new tiers, which draw at about 2K — and it runs Microsoft’s content filter, which is strict. Coverage of the model praises its portraits and lettering. The free sheets are not drawn on this tier; they stay on the default model.

Every tier at a glance: which image tier for your sheets.

## What MAI-Image-2.6 is

Microsoft AI announced MAI-Image-2.6 on August 10, 2026, and it became available to developers on Microsoft Foundry on September 4, 2026. Microsoft describes it as a diffusion model for both text-to-image and controlled image editing. At launch it placed second on the Arena text-to-image leaderboard, with the largest gains over its predecessor reported in portraits, text rendering and 3D-style images. Sources are at the end of this page.

On SceneMixer you do not need an Azure or Microsoft account; the tier is billed in SceneMixer credits.

## What it draws here

The same things every image tier draws: each character full length on white, each location as an empty plate, each prop alone, the state images for second looks, and the three candidates of a promo cover. The video tiers then render every shot against those images.

## Where to switch it on

Settings → Model Preferences → Image Model, then pick MAI-Image-2.6. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching on SceneMixer never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

## About one megapixel

MAI-Image-2.6 has no resolution setting; on this site it draws at about one megapixel (768 by 1360 for a vertical image). The other five new tiers draw at about 2K. For a reference sheet that matters less than it sounds — the video model reads a face, a costume and a room from it — but it does matter if you plan to crop into a sheet or print a cover large. Download the cover at the size you need and check it before you use it in print.

## A strict content filter

Every description is checked by SceneMixer first; a flagged description is not charged. MAI-Image-2.6 then runs Microsoft’s own filter, and in our use it refuses more readily than the other tiers. A refusal is refunded. If a harmless description is refused, reword it — plain, concrete sentences, no repeated phrases — or switch to another tier for that one sheet.

## When to use it

- Portraits of the leads, if you like the look it gives your faces. Draw one lead on it and compare before moving the whole cast.
- Covers where the title is short and bold.
- Not for 2K location plates (SeedDream 5.0 Lite or GPT Image 2.5 Flare at 25 credits), or for a big cast on a budget (SceneMixer's default Qwen-Image 3.0 at 7 credits, with the free sheets).

## What it costs

This tier is not covered by the free sheets. Every account’s 15 free sheets are drawn on the default image model, Qwen-Image 3.0, only. On MAI-Image-2.6 each sheet costs 31 credits from the first one (about $0.19 at the lowest credit price), and the price is printed on the button before you generate. Switch back to the default and whatever free sheets you have left are still there. A generation that is refused is not charged, and most failures are refunded automatically; a few kinds are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the description tripped the site’s check, or the model maker’s own filter refused the picture | Edit the description and generate again; nothing was charged |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The model was saturated; the site kept the job queued for a while before giving up | Generate again in a few minutes, or switch tier |
| A sheet that ignores part of the description | Too many details competing in one description | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |
| Credits held for review | A rare failure where the model may already have been paid for the attempt | Nothing to do; support checks it and releases or refunds — see credits held for review |

## All the image tiers, side by side

| Image tier | Credits per image | In short |
| Qwen-Image 3.0 (default) | 7 | The free sheets; a whole cast cheaply |
| Qwen-Image 3.0 Pro (Recommended) | 18 | Leads in close-up; Chinese descriptions |
| SeedDream 5.0 Lite | 8 | 2K locations and props on a budget |
| SeedDream 5.0 Pro | 25 | 2K, photographic skin, long descriptions read to the end |
| Nano Banana 2.1 | 19 | About 2K; Google’s faster model, the cheapest of the six new tiers |
| Nano Banana Pro | 46 | About 2K; detail kept from long descriptions; the most expensive tier |
| GPT Image 2.5 Flare | 25 | About 2K; OpenAI’s faster general-purpose model |
| GPT Image 2.5 Sunburst | 25 | About 2K; OpenAI’s slower, more precise model |
| Grok Imagine Image 2.0 | 35 | About 2K; xAI’s model, strong on lettering |
| MAI-Image-2.6 | 31 | About 1 megapixel; Microsoft’s model, strict content filter |

## Sources

- Microsoft Foundry: MAI-Image-2.6 — model card; checked October 25, 2026.
- Windows Report: Microsoft releases MAI-Image-2.6 — launch and Arena placement; checked October 25, 2026.
MAI and Microsoft are trademarks of Microsoft, named here to identify the model offered as a tier. Prices are SceneMixer’s own, read from the live price list.

| | MAI-Image-2.6 on SceneMixer |
| Shown in Settings as | MAI-Image-2.6 |
| Used for | Character, location, prop and state sheets; cover posters |
| Credits per image | 31 |
| Output size | About 1 megapixel (no resolution setting) |
| Content filter | Microsoft’s own, strict; refusals are refunded |
| Free sheets | No — default model only |

## FAQ

### Can I use MAI-Image-2.6 on SceneMixer?

Yes. Pick MAI-Image-2.6 in Settings, Model Preferences, Image Model, and new sheets and covers are drawn on it.

### How much does MAI-Image-2.6 cost here?

SceneMixer charges 31 credits per image, about $0.19 at the lowest credit price.

### Is it free on SceneMixer?

No. The 15 free sheets per account are drawn on the default image model only.

### What resolution does MAI-Image-2.6 draw at?

About one megapixel on SceneMixer, for example 768 by 1360 for a vertical image. The model has no resolution setting.

### Why was my description refused on MAI-Image-2.6?

Microsoft's own content filter runs on this tier and is strict. The refusal is not charged on SceneMixer; reword the description or use another tier for that sheet.

### Do I need an Azure account?

No. The tier is part of SceneMixer and billed in SceneMixer credits.

### Is MAI-Image-2.6 Flash available?

No. The tier on SceneMixer is MAI-Image-2.6.

### Is it good at faces?

Coverage of the model reports its biggest gains in portraits. Try it on one lead and compare with the tier you use now before switching the whole cast.

Updated November 3, 2026


---

# Make a Black Friday Ad from One Photo: A Plan for the 27th

URL: https://scenemixer.com/guides/make-a-black-friday-ad-from-one-photo/
A Black Friday or Cyber Monday video ad from one product photo: a plan counted back from November 27, the end card, which tone fits a sale, and a second cut for Monday.

# How to make a Black Friday ad from one photo

Quick answer: Black Friday 2026 is Friday, November 27, and Cyber Monday is Monday, November 30. A sale ad needs one thing that cannot be improvised on the day: footage that is ready. In SceneMixer you upload one to four photos of what you are selling, pick a length of 15, 30 or 60 seconds, a tone and a shape, and the ad is written, shot and finished with an end card carrying your brand, product name, tagline and call to action. Put the offer on the end card and in the spoken line, because nothing legible is written into the picture. Make the main ad now, then change the end card for Cyber Monday and rebuild the final cut without reshooting.

One-line description for the box next to the upload: This is our own ceramic coffee set, sold directly; Black Friday weekend sale, free gift wrap. When the photo alone is ambiguous, this line outranks what the picture seems to show.

## Count back from the 27th

Thanksgiving is Thursday, November 26, so the sale weekend starts the next morning. Working backwards from there:

| When | What to do |
| Now to Nov 8 | Choose the hero product and take or clean up one to four photos of it. Decide the offer, the end date and the link. |
| Nov 9–15 | Make the 30-second main ad in portrait. Check the end card text. Make a 15-second teaser from a second photo if you have a second product. |
| Nov 16–22 | Fix any shot that missed in the editor, rebuild the final cut, export. Schedule the teaser. |
| Nov 23–26 | Teaser out. On Thanksgiving, a soft “tomorrow” post. |
| Nov 27 | Main ad out. |
| Nov 28–30 | Edit the end card for Cyber Monday (“Last day”, the new end time) and rebuild the final cut with the same shots. |
The earlier you start, the more room there is to reshoot one shot rather than all of them. The whole ad flow, step by step.

## Pick the photo

- The first photo is the main subject. The end card is built around it. A product, a dish or a garment is redrawn as a clean studio shot on white so it is lit the same in every shot.
- A shop, a car or an app screen is not redrawn. A shop becomes the place the ad happens inside. People in a photo of a place or a vehicle are removed first.
- Up to four photos of the same subject: angles, a detail, the product in use.
- If the picture is misleading, say what it is in the one-line box; that line wins over the photo.
Which photos work for an ad · An ad for a physical product · for clothing · for a restaurant · for an app.

## Pick the tone for a sale

| Tone | Good for |
| Epic | A hero product with a big reveal: one item that anchors the sale. |
| Mouth-watering close-ups | Food, drink, a gift hamper. |
| Show how it works | Gadgets, apps, tools; the demo is the pitch. |
| Funny | Mass-market items; the product is the punchline. |
| Warm | Gifts and family purchases. |
| Twist | A “you think it’s one thing” reveal that ends on the offer. |
| Minimal | Premium goods that do not shout. |
| A day in the life | Clothing and everyday items worn or used. |
Length is 15, 30 or 60 seconds (one, two or four segments of fifteen seconds); the resolution is 480p, 720p or 1080p. Length and resolution change the price in SceneMixer; tone and shape do not. Ad type, length, shape and resolution.

Choose portrait for feeds and short-video apps, landscape for a website or a pre-roll slot. Changing the shape later is a new ad. Music is written for the ad and mixed into the final cut: where the music comes from.

## What goes on the end card

The end card is the one place the picture carries readable words. Four lines, any of which can be blank:

- Brand and product name, read from your photo and description.
- Tagline: the line people should remember. Black Friday weekend only.
- Call to action: Shop the sale at your-shop.example.
Two rules for a sale ad. First, superlatives such as “the lowest price anywhere” or “number one in sales” are stripped from the script before it is produced, so say what is true: the real discount and the real end date. Second, settle the wording before you assemble: the end-card hint warns that changing it after assembling means assembling again, and paying again. The end card, line by line.

## The second cut for Cyber Monday

Once the ad is finished, Edit end card opens the four lines on its progress page. Change the tagline to the Monday version and press Rebuild final cut: the same shots are joined again with the new card, nothing is reshot, and if the rebuild fails, the previous cut stays. The ad progress page.

Before you start, the price is shown and it is a cap, not an estimate. What you actually pay.

## Make it in SceneMixer, start to finish

- Open Make an ad and add one to four photos of the product.
- Type the one-line description if the photo is ambiguous: what it is, who it is for, the sale.
- Pick the tone, the length, the shape and the resolution. The price for the whole ad is shown before you start.
- Hand it to the AI to run the whole chain, or use step-by-step mode to edit the script and reshoot single shots before paying for the rest.
- Check the end card, assemble, download and post. Make the Monday version from the progress page.

| Setting | For a Black Friday ad | Why |
| Photos | 1–4 of the hero product | The first is the main subject |
| Length | 30 s main, 15 s teaser | Two segments, then one |
| Shape | Portrait for feeds | Landscape for the website |
| End card | Offer, end date, link | The only readable text |
| Monday | Edit the end card, rebuild | Same shots, no reshoot |

## FAQ

### When is Black Friday 2026?

Friday, November 27, 2026, the day after US Thanksgiving on Thursday, November 26. Cyber Monday is Monday, November 30.

### How do I make a Black Friday ad from one photo?

Open Make an ad, add one to four photos of the product, pick a tone, length, shape and resolution, check the end card and let it run. The price is shown before you start.

### Can I put the discount in the video?

Put it on the end card and in the spoken line. Nothing legible is written into the picture itself.

### Can I say lowest price in the ad?

No. SceneMixer strips superlatives and absolute claims from the script. Say what is true: the discount and the end date.

### Can I make a second version for Cyber Monday?

Yes. Edit the end card on the progress page and rebuild the final cut. The same shots are used, and nothing is reshot.

### Is the end card free to change?

Changing it after assembling means assembling again, and the interface warns that it is charged again, so finish the wording first.

### How long should a sale ad be?

15 seconds for a teaser, 30 for the main ad, 60 if it tells a small story.

### Can I make ads for several products?

Yes, one ad per product, each from its own photos. Lead with the hero product.

Updated November 3, 2026


---

# How to Make a Movie With AI for Free: What Is Free and What Is Not

URL: https://scenemixer.com/guides/make-a-movie-with-ai-for-free/
Make a movie with AI for free: what a free account gets, the order to use it, where payment starts, and how to keep a first film to a few seconds of paid video.

# How to make a movie with AI for free

Quick answer: You can start a movie in SceneMixer for free and stay free until you render video. A new account gets 50 credits, the first 15 setup images are free, the first episode's storyboard on your first drama is free, editing text is free and you can watch a preview of about 5 seconds. The finished video is paid by the second at 10 credits a second on the default tier. So the free path is a full plan, a cast and a preview; the movie itself costs credits.

Free first session: paste a one-page script, approve the cast with the free images, read and edit the free storyboard, run the free preview, then decide.

## The free path, in order

- Write or paste the script. Reading is 4 credits per 1,000 characters; the welcome credits cover a short script.
- Approve the cast and places using the first 15 free images.
- Read the storyboard and rewrite what is wrong. It is free.
- Run the free preview.
- Stop, or render the rest.
What you get for free · Free sheets are per account.

## Where payment starts

Video is the cost. Each finished second on SceneMixer is 10 credits a second on the default tier, about $0.06 at the lowest credit price, plus 1 credit a second to join. A 60-second movie is 60 of those seconds. Videos on a free account carry a small AI label; a paid membership turns it off by default. Compositing and sharing.

## Make the paid part small

- Render only the opening first.
- Fix a shot alone rather than redoing a scene.
- Keep to one place and two characters.
- Use a 30-second film as the test.
Spending fewer credits · Free AI movie maker · Free AI drama generator.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Use the free images on the two main characters and the one place; they are the part you keep for every later version. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Step | Free? | Notes |
| Sign up | Yes | 50 credits to start |
| Character and place drawings | The first 15 images | Per account |
| Storyboard | Free for the first episode of your first drama | Editing text is free |
| Preview | About 5 seconds, once | Shows look and voices |
| The finished video | Paid | 10 credits a second on the default tier |

## FAQ

### Can I make a movie with AI for free?

You can plan it, cast it, storyboard it and preview it free on SceneMixer. The finished video is paid by the second.

### Is there a watermark?

Videos on a free SceneMixer account carry a small AI label in the corner; a paid membership turns it off by default.

### How much does one minute cost?

Sixty seconds at 10 credits a second on SceneMixer's default tier plus 1 credit a second to join.

### Do I need a card to start?

No card is needed on SceneMixer to read the script, draw the cast and read the storyboard.

### Is the preview long?

About 5 seconds, once.

### Can I edit the storyboard free?

Yes. Editing text on SceneMixer is free.

### What is the cheapest way to finish a film?

Keep it short: one place, two characters, 30 to 60 seconds.

Updated November 3, 2026


---

# AI Drama Poster and Cover Art: Make One for Your Series

URL: https://scenemixer.com/guides/make-a-promo-cover-poster-for-your-ai-drama/
SceneMixer paints three poster options from your leads' character sheets with the title drawn in, then adds the tagline and episode numbers. How it works, what it costs.

# Make a promo cover poster for your AI drama series

Quick answer: Open step 2 of your project and pick the Cover art tab. Type the title, choose the title lettering, the genre mood and up to two cover characters, add an optional tagline, and press Generate 3 options. The AI paints three posters from your leads’ reference art with the title drawn into the picture; you pick one, and the site adds the tagline and each episode’s number on top. A portrait and a landscape version come with it; other sizes are made on request. Each option is priced as one image on the image tier you have selected, and your account’s first project gets its first set of three free.

The cover is drawn on your image tier; which one to use is in which image tier for your sheets.

## What you get

- Three key-visual options — finished posters with your title painted in, built from your leads’ character sheets, so the faces and costumes are the ones in your series.
- The tagline and episode numbers on top — laid out by the site, not painted by the model, so you can change them any time without redrawing.
- Portrait 9:16 and landscape 16:9 — the project’s own shape first; once you pick one, the other is redrawn from it at no extra cost.
- Episode covers — the same poster with “EP 1”, “EP 2” and so on, in every size, if you want them.
- The share card — the cover becomes the picture people see when you share a link to the series.

## Before you start

The cover is drawn from your leads’ reference art, so at least one lead needs a character image in step 2. If none has one yet you will see “Generate a character image for at least one lead before making cover art.” Ad projects don’t have cover art.

The image tier you have selected in Settings → Model Preferences → Image Model paints the cover. Switch it before you generate if you want a particular model on the poster.

## Step by step

- Open the Cover art tab. In step 2 (Characters & Scenes) the tabs are Characters, Scenes, Items and Cover art. From the episode editor, the first card on the timeline is the cover; when it says Add cover, clicking it takes you there.
- Title. It is drawn right into the poster — 12 characters or fewer works best. The title of the series is filled in for you; edit it if you like.
- Title lettering. Four styles: Brush, Gold, Rounded and Serif. One is marked as recommended for your genre.
- Genre mood. Twelve moods — Romance, Suspense, Epic, Family, Feel-good, Workplace, Horror, Comedy, Teen, Survival, Sci-fi and Action — set the poster’s colours and atmosphere. The site reads your story and pre-selects one (“recommended for your story”); pick another if it is wrong.
- Cover characters. Pick up to two, people or animals, from everyone in the project who has a character image. Your leads are pre-selected. Characters who only appear as a voice are listed last.
- Tagline (optional). One line to hook viewers. The site can write a suggestion; it is added on top later, so you can leave it empty and add it after.
- Generate 3 options. The price is on the button. The three come back in about a minute or two.
- Pick one with Use this one. SceneMixer lays out the tagline and draws the other base shape for you.

## Writing a title that comes out right

The title is painted by the image model as part of the picture, which is why it looks like a real poster rather than text dropped on top — and also why it can come out misspelled. The site does not correct it. Keep the title short (twelve characters or fewer is the safe range), check the spelling on each option before you pick, and if all three are wrong, change the lettering style or the tier and generate again.

Tiers whose makers stress lettering are a reasonable first try for covers: Grok Imagine Image 2.0 (35 credits per option) and GPT Image 2.5 Sunburst (25). See Grok Imagine Image 2.0 and GPT Image 2.5.

## Sizes and downloads

Portrait 9:16 and landscape 16:9 are always included. Three more shapes can be made from the one you picked — recomposed with the same people, outfits and title lettering:

- 3:4 — a short-drama platform cover.
- 1:1 — a social post.
- 4:3 — a landscape card.

Each extra shape is priced as one image on your tier. Download offers This size or All sizes. On a free account the AI badge is on the cover, as on videos; members can remove it, and the same setting controls covers and videos.

## Episode covers

Once the series has a cover, Make episode covers makes one per episode, with its episode number, in every size. It is priced per episode and makes every missing episode at once; after that, changing the tagline or picking a different cover is free. New episodes added later are charged when you make their covers. In the episode editor, an episode without its cover shows a lock and Make cover.

## Changing it later

- Tagline: change it any time with Apply; the layout is redone, nothing is redrawn, no charge.
- A different option: Pick a different one from this set or an earlier set; free.
- Title, lettering, mood or characters: these are painted into the picture, so changing them makes a new set of three, at the same price per option.

## What it costs

- A set of three options: three images on your selected image tier — for example 3 × 7 credits on the default Qwen-Image 3.0.
- The first set on your account’s first project: free, once per account, on a project whose script reading you paid for (the same condition as the free sheets). On accounts created since October 23, 2026 it is made automatically once your leads have character images, marked “We made 3 from your story, free”.
- The second base shape (landscape for a vertical project, portrait for a horizontal one): free.
- 3:4, 1:1, 4:3: one image each.
- Episode covers: one image per episode, all missing episodes at once.

An option that does not come out is refunded; so is an episode-cover run that fails. The price is always on the button before you press it.

## Messages you may see

| Message | What it means |
| “Generate a character image for at least one lead before making cover art.” | No lead has a character image yet. Make one in step 2. |
| “Cover art is already being generated. Try again once this batch is done.” | A set is still being painted. |
| “The characters you picked have changed… Please pick again.” | One of them was deleted or lost its image. |
| “Pick one of the candidates first.” | Extra sizes and episode covers are made from the option you picked. |
| “The price just changed. Check the new price and try again.” | Your tier or the episode count changed while the page was open; nothing was charged. |
| “Ad projects don’t have cover art.” | Covers are for series, not ads. |

## FAQ

### Can SceneMixer make a movie poster for my AI drama?

Yes. Step 2, Cover art: it paints three poster options from your leads' character sheets with the title drawn in, and adds the tagline and episode numbers.

### Is the cover free?

In SceneMixer, the first set of three on your account's first project is free. After that each option is priced as one image on your selected image tier.

### Can I make a poster from a photo?

The cover is built from your characters' reference art. If you uploaded your own photo as a character's image, that character can be picked for the cover.

### Why is the title misspelled?

The image model paints the title as part of the picture, and SceneMixer does not correct it. Keep it short, check each option, and generate again with another lettering style or tier if needed.

### Can I change the tagline after?

Yes, any time, free. The tagline is laid out on top of the picture, not painted into it.

### Which sizes do I get?

Portrait 9:16 and landscape 16:9 are included. 3:4, 1:1 and 4:3 can be made from the option you picked, each at one image's price.

### Can I use the cover as a YouTube or TikTok thumbnail?

Yes: download the landscape version for YouTube or the portrait version for vertical platforms. You can use it the same way as your videos; see what you get to keep.

### Can I remove the AI badge from the cover?

Members can, with the same setting that controls the badge on videos.

### Can animals be on the cover?

Yes. Any character with a character image can be picked, people or animals, up to two.

Updated November 3, 2026


---

# Make a Second Language Version of Your AI Drama

URL: https://scenemixer.com/guides/make-a-second-language-version-of-your-drama/
Dialogue language is fixed once a project is read, so a second-language version is a new project from the same script. Keeping cast, names and look identical.

# Making the Same AI Drama in a Second Language

Quick answer: The dialogue language is chosen in step 1 and cannot change after the script is read, so a version in another language is a second project. Start a new project with the same script, choose the new Dialogue language, set Target Region to Follow script to keep the names and setting, and pick the same visual, scene and costume styles. In step 2, bring each character across with Pick existing so the faces match — the whole entry moves over, voice included; if the voice does not suit the new language, swap it for another from the preset library afterwards. If viewers only need to understand the story, subtitles over the existing episodes are the cheaper route.

Open the first project’s step 1 summary and note its styles and region before you start the second project.

## Why it is a second project

The dialogue language shapes how the script is read: the lines in the outline and the storyboard are written in it, and the clips speak them. That is why it is set before the reading and stays fixed afterwards. A second language means reading the script again, in a new project. What can still change after the read.

## Settings that keep the two versions alike

- Dialogue language: the new language.
- Target Region: Follow script, so names, places and faces stay as in the original. A different region relocates the story.
- Visual, scene and costume styles: the same as the first project.
- Aspect ratio and drama type: the same.

Interface, script and dialogue language.

## The same cast

In SceneMixer's step 2, open each character and press Pick existing, then choose the character from the first project; the whole entry comes across — the images, the description and the voice — so the face and costume match and the voice description is translated into the new language with the rest. If the voice itself does not sound right in the new language, pick another one afterwards from the voice row’s preset library. Scenes can be copied the same way. Reuse a character across projects · Pick character voices.

## Choosing the tier

Some languages may occasionally be mispronounced on some tiers; the tier guide says which. Generate a test segment in the new language before the whole episode. Dialogue language and video tier.

## When subtitles are enough

If the goal is reach rather than a dubbed cast, add subtitles in the new language over the episodes you already have, in any video editor or on the platform you post to. That costs nothing on SceneMixer; a second project is charged like any project.

| Setting | First version | Second version |
| Script | Your script | The same script |
| Dialogue language | Language A | Language B |
| Target Region | As chosen | Follow script, to keep names and setting |
| Styles, aspect, type | As chosen | The same |
| Cast | Generated | Pick existing from the first project |

## FAQ

### Can I switch the dialogue language of the project I already have?

No. SceneMixer fixes it once the script is read. The second language is a new project.

### Will the characters look the same?

Yes, if you bring them across with SceneMixer’s Pick existing and keep the same styles.

### Will the episode split be identical?

Usually close, but SceneMixer reads the script again, so episode boundaries can differ slightly. Check step 3 before generating.

### Is the second version charged in full?

Yes, like any SceneMixer project. Subtitles over the first version are the cheaper option.

Updated November 3, 2026


---

# Make a Short Drama with Kling 4.0 in SceneMixer, Step by Step

URL: https://scenemixer.com/guides/make-a-short-drama-with-kling/
Kling 4.0 and Kling 4.0 Flash are selectable video tiers in the episode editor: where to pick them, generating a segment, the 20-second cap on Flash, and fixes.

# How to Make a Short Drama with Kling 4.0 or Kling 4.0 Flash

Quick answer: Kling 4.0 and Kling 4.0 Flash are two video tiers you can pick per segment in the episode editor. Kling 4.0 renders at 1080P and takes segments up to 30 seconds; Kling 4.0 Flash is the 720P, 20-second one, and cheaper per second. Neither is the default — you choose the tier yourself, in the model picker in the editor's top bar or under Settings → Model Preferences → Video Model. Open a project, read the script, generate the cast sheets and an episode storyboard, pick the tier, then generate segments one by one; the price for the segment is on the button before anything is charged.

## What the two tiers are

The still above is from Neon Debt, an episode this site took from script to composite; its segments were rendered by choosing a tier per segment.

Both are video tiers in the same editor as every other tier, chosen per segment:

| Tier | Resolution | Segment length | Price |
| Kling 4.0 | 1080P | up to 30 seconds | 38 credits per second |
| Kling 4.0 Flash | 720P | up to 20 seconds | 29 credits per second |

The numbers move with the knobs: the per-second price and the segment cap come from the site's own pricing, so the exact rate for your account is the one printed on the generate button. You do not need to memorise anything — the button always shows the full price of that segment before you press it.

## Where you pick a tier

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier's name. Press it, the Select Model list opens, and you choose Kling 4.0 or Kling 4.0 Flash. In Settings: Model Preferences → Video Model. Both change the same thing: the tier your next segments render on.

Three rules that save confusion later:

- It is a per-segment choice, not a per-project lock. You can render one segment on one tier and the next on another; the composite joins whatever each segment produced.
- Already-made segments keep their tier. Switching the picker does not re-render anything; each version records the tier it was made with.
- Neither tier is the default. The default is the cheap drafting tier; a new project starts there until you change it. Choosing a Kling tier is a deliberate step, which is the point — draft on the default, finish the keepers on the tier you want.

## Steps 1 and 2: the project and the cast

If the project already exists, skip to the storyboard. If not:

- Start a project. In step 1 set the format — Short Drama, the aspect ratio (9:16 for a vertical drama) and the dialogue language. These follow the project.
- Add the script by pasting it or uploading a .txt, .docx or .md file, then let the read run: it returns the cast, the locations, the props and an episode-by-episode outline.
- Generate the sheets in step 2: one reference image per character, plus the locations and props. These are what keep a face recognisable in every segment, whatever tier renders it. Keeping characters consistent with Kling tiers goes through that layer in detail.

Your first 15 reference images are free on your SceneMixer account (characters, scenes and props share that pool); after that each one is charged at the image tier's rate.

## Step 3: the storyboard

Generate an episode's storyboard in step 3, then edit it in the editor. A storyboard is a stack of segments; each segment holds numbered shots, and each shot carries its seconds, the characters actually in frame, and the lines to be spoken. The segment's length is the sum of its shots' seconds, and that is what you pay for.

Plan the segment length against the tier's cap now rather than later. On Kling 4.0 Flash a segment of 20 seconds or less renders; on Kling 4.0, up to 30. How to write one shot covers the writing side.

## Step 4: generate a segment on a Kling tier

- Pick the tier in the model list (see above).
- Select the segment on the timeline. The generate button shows the tier and the price for that segment: seconds rounded up to whole seconds, times the per-second rate, with the price on the button itself.
- Press Generate Video. The credits are held, not yet spent.
- Wait or keep working. The segment card shows progress; you can work on other segments or close the page.
- On delivery the credits are charged and the take becomes the segment's active version. Regenerating later does not overwrite it — every take is kept in the version history.

To render a whole batch instead, Generate all plans every segment that has no clip yet and shows the count, the total seconds, the tier and one total price in a confirmation window. Each segment is charged on its own, and failures are refunded automatically. Segments beyond your concurrency limit wait in line and start on their own.

## The 20-second cap on Kling 4.0 Flash

This is the one hard rule that bites. A segment longer than 20 seconds cannot render on Flash, and the editor tells you in three places:

- In the model list, Flash is greyed out for that segment (disabled, dimmed) and carries a short note that the segment is over the cap — segment 24s over cap.
- In Generate all, the segments that are too long are left out of the batch and listed in the confirmation window: 2 segments are longer than this model allows and are left out. Switch models to generate them.
- If you force the generate anyway, the submission is rejected with This segment exceeds the duration cap of the selected model — pick a model that supports longer clips. Nothing is charged.

Two ways out. Shorten the segment — trim its shots' seconds so the total lands at 20 or under — or switch that segment to Kling 4.0, which takes up to 30. The cap is per segment, so a long scene split into two 15-second segments renders on Flash exactly as it should; a single unbroken 25-second take needs the bigger tier.

## Sound and lines on these tiers

Same as any tier here: the lines written in the storyboard are performed in the video, generated with the picture rather than dubbed on afterwards, and ambient sound comes out of the same render. The project's dialogue language is the one the cast speaks; you set it once in step 1.

If a line must land on a beat, write it into the shot with enough seconds for the character to say it. The rules for writing lines and speaking characters are in How to write one shot and Native dialogue vs dubbing.

## Regenerating and fixing a segment

In SceneMixer, editing the storyboard text is free. Change a word, move a second, rename who is in frame — nothing is charged until you render again. Regenerate renders a new take on the selected tier and costs the segment again; the old take stays in the version history and switching between versions is free.

For a clip that is already finished, the player has Edit this clip with its own tools — trim it, redub a line, edit visuals, enhance clarity. Each of those shows its own price before you confirm it. What each one does is set out in Edit this clip: which fix to use and Fix one shot without redoing the episode.

One habit worth keeping: fix the storyboard first, render last. A segment that reads wrong is cheapest to correct before it is rendered.

## What this costs

Both tiers are billed by the second, like every video tier here. A 10-second segment on Kling 4.0 costs 10 × 38; the same 10 seconds on Kling 4.0 Flash costs 10 × 29. The button shows the total, and the credit history records one line per segment.

- Held when you press, charged on delivery. A failed generation is refunded automatically; a generation whose outcome is unclear is held for review instead of being charged twice.
- No free allowance on either tier. Video is billed from the first second here; the free parts of an account are the reference images and the first project's episode-1 storyboard.
- The fix tools are priced the same way — on the button, before you confirm.

Draft elsewhere, finish here: render the whole episode on the cheap default tier while the story is still moving, then re-render only the segments you keep on the Kling tier you want. Draft at 480P, finish at 1080P is that working pattern, and Kling 4.0 vs Kling 4.0 Flash is which of the two to finish on.

## Common errors, and what to do

| What you see | What it means | What to do |
| This segment exceeds the duration cap of the selected model — pick a model that supports longer clips. | The segment is longer than this tier renders (20 s on Flash). | Shorten the segment or switch it to Kling 4.0. |
| Flash greyed out in the model list, with a note like segment 24s over cap | Same cap, shown before you press. | Pick the other Kling tier for this segment, or trim it. |
| A generation that fails and says the credits were returned | Nothing was delivered. | Press generate again; nothing was charged. |
| A generation held for review | The outcome was not clear. | Do not resubmit; the hold resolves to a refund or a delivered take. |
| A rejected generation (content check) | Text or a reference image was flagged before rendering. | Edit the flagged material and render again; rejections are refunded. |

The full list of editor messages, with reasons, is in every edit-this-clip message and content rules and checks.

## What is Kling's own specification, and what is this tier

Keep the two apart, because they are not the same thing. The tier you pick in the editor gives you what the table at the top of this page says: 1080P or 720P, a segment cap, and a per-second price, all of it shown live from this site's pricing.

Kling's own model announcements are the model's specification: as officially announced by Kling, 4.0 supports up to 4K, up to 10 keyframes, Omni Reference with up to 15 references, stereo audio in 9 languages, and 10-bit HDR marked as coming soon. Those are the officially announced capabilities of the model, not a list of what any particular tier here exposes — this site's tier is defined by the picker, the segment cap and the price.

The announcement itself is summarised at Kling's 4.0 model release page, and our own write-up of what it says is Kling 4.0: what was announced.

## FAQ

### Are Kling 4.0 and Kling 4.0 Flash the default tiers?

No. A new project starts on the cheap default tier; you pick a Kling tier yourself, per segment, from the model list in the editor or in Settings under Model Preferences.

### How long can a segment be on each tier?

Kling 4.0 takes segments up to 30 seconds; Kling 4.0 Flash up to 20. The cap is per segment, and the editor blocks a segment that is over it before anything is charged.

### What happens if a segment is longer than 20 seconds and Flash is selected?

Flash is greyed out for that segment with an over-cap note, Generate all leaves the segment out and says so, and forcing the generate is rejected with a message telling you to pick a model that supports longer clips. Nothing is charged.

### How much does a segment cost?

Seconds times the per-second rate of the tier. The rate and the total for your segment are printed on the generate button before you press it; the table above shows the two tiers side by side.

### Is the price the same on both tiers?

No. Flash is the cheaper per-second tier; that is what you trade the 1080P resolution and the extra 10 seconds of segment length for.

### Does the free first-project allowance cover these tiers?

No. Video is billed from the first second on SceneMixer, on every tier. The free parts of a new account are the 15 reference images and the first project's episode-1 storyboard.

### Do the lines get spoken, or do I add sound afterwards?

The cast performs the lines written in the storyboard in the same render as the picture, with ambient sound from the same pass. There is no separate dubbing step.

### Can I change a segment after rendering it?

Yes. Editing the storyboard text is free; Regenerate makes a new take and costs the segment again, keeping the old take in the version history. A finished clip can also be trimmed, redubbed, re-edited or enhanced from Edit this clip, with its own price shown before you confirm.

### Can I mix these tiers with other tiers in one episode?

Yes. The tier is chosen per segment and the composite joins each segment's active version, whatever tier produced it.

### Where do the officially announced 4K, keyframe and Omni Reference figures come from?

From Kling's own model announcements: 4K, up to 10 keyframes and Omni Reference are the officially announced 4.0 specification. What this site's Kling tiers expose is what the model picker here shows, with its own caps and prices.

Updated November 3, 2026


---

# Social Story Video for Kids: Make a Simple Narrated Story

URL: https://scenemixer.com/guides/make-a-social-story-video-for-kids/
Make a simple narrated social story video for kids with AI: a calm cartoon cast, one idea per shot, a steady voice, a worked script and clear limits.

# How to make a social story video for kids

Quick answer: A social story is a short, plain story that explains a situation, such as going to the dentist, and what a child can do in it. SceneMixer can turn a story you write into a simple narrated video: a calm cartoon child or animal, a steady narrator, one idea per shot and the same opening and closing every time. It is a video maker, not a therapy or teaching programme, and nothing on this page promises any result for any child. Write the story with the child's parent, teacher or specialist, and watch every second before you show it.

Brief for AI Create: A one-minute story for a child who is nervous about the dentist: a bear cub called Bo goes to the dentist; what he sees, what the dentist does, what Bo can do if he feels worried, and a calm ending at home; a slow, warm narrator, simple sentences, nothing frightening.

## What this page is, and what it is not

- It is a how-to for putting a short, simple story into pictures and a voice.
- It is not medical, therapeutic or educational advice, and it is not a substitute for the child's own teacher, therapist or doctor.
- It makes no claim that a video helps any particular child. Children differ, and the people who know this child decide what suits them.
- If a specialist already works with the child, ask them for the text. Some professionals use a specific writing method for social stories, with its own guidelines; this page does not teach or certify it. Use SceneMixer only to put their text into pictures and a voice.

## Write the story first

Plain, literal and calm. These are common ways people write this kind of story; adapt them with whoever knows the child:

| Kind of sentence | What it does | Example (the dentist) |
| What happens | Describes the place and the people | The dentist's room has a big chair that goes up and down. |
| Who is there | Names the people and what they do | The dentist wears a mask. The mask keeps germs away. |
| How others may feel | Says that feelings differ | Some children feel a little worried. That is okay. |
| What I can do | Gives a choice or a signal | If Bo feels worried, he can raise his paw. Then the dentist stops. |
| What happens next | Closes the situation | When the check is finished, Bo goes home. |
- Short sentences, one idea each; say what usually happens, and use “usually” where it varies.
- Avoid sayings and jokes that are not meant literally.
- Say what to do rather than what not to do.
- Change the details to match what will really happen: the real room, the real order of events.

## Choose how it looks and sounds

| Choice | Suggestion | Why |
| Look | Watercolor Storybook or 3D Toy Cartoon | Round, bright and slow; nothing frightening |
| Main character | An invented cartoon child or animal, drawn once and used in every story | The same friendly face each time; no real child's photo needed |
| Narrator | An Off-screen character with one calm voice, the same every time | A steady voice carries the story |
| Shots | 4 to 6 seconds, one idea each, still or a slow push-in, one place per segment | Time to take each picture in |
| Sound | Quiet ambient lines in the shots; turn the Music switch off for the version without the score | Nothing sudden |
| Language | The language spoken at home, from the 15 dialogue languages; subtitles on accounts that have made a purchase | Familiar words |

Visual Style: 3D Toy Cartoon
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
Character voices · Background music and the Music switch · Add subtitles · AI videos for your classroom · A kids bedtime story series.

## A worked script: Bo goes to the dentist

Each narrator line is given enough seconds to be said calmly: SceneMixer plans English speech at about three words a second plus a pause at each sentence end. About 50 seconds in all.

| Beat | Shot |
| The waiting room | [Shot 1] 5s: Wide shot, still. A bright waiting room in soft watercolor, @Bo the bear cub on a small chair beside his mother's bag. @Narrator (voice-over, calm and slow): Today Bo goes to the dentist. |
| The fish tank | [Shot 2] 5s: Medium shot. @Bo looks at a fish tank. @Narrator (voice-over): The waiting room has a fish tank and some books. |
| The room | [Shot 3] 6s: Wide shot. The dentist's room, a big chair, a round lamp. @Narrator (voice-over): The dentist's room has a big chair that goes up and down. |
| The dentist | [Shot 4] 6s: Medium shot. @Dentist, a kind badger in a mask, waves gently. @Narrator (voice-over): The dentist wears a mask. The mask keeps germs away. |
| The check | [Shot 5] 5s: Close-up on a small mirror and a tiny light. @Narrator (voice-over): The dentist looks at Bo's teeth with a little mirror. |
| The feeling | [Shot 6] 5s: Close-up on @Bo's face, eyes wide. @Narrator (voice-over): Some children feel a little worried. That is okay. |
| What Bo can do | [Shot 7] 7s: Medium shot. @Bo raises his paw; @Dentist stops and nods. @Narrator (voice-over): If Bo feels worried, he can raise his paw. Then the dentist stops. |
| The end | [Shot 8] 5s: Medium shot. @Bo smiles and climbs down from the chair. @Narrator (voice-over): When the check is finished, Bo goes home. |
| Outside | [Shot 9] 4s: Wide shot. @Bo and his mother walk out into the sunlight. No dialogue. Ambient: birdsong. |
| Home | [Shot 10] 4s: Medium shot. @Bo plays on a rug at home. @Narrator (voice-over): Then Bo goes home and plays. |
Voice-over lines · Shot seconds and timing.

## Situations to start with

| Situation | What the story shows | What the child can do |
| Going to the dentist | The room, the chair, the mask | Raise a paw to pause |
| Getting a haircut | The cape, the scissors, the noise of the clippers | Say “stop” or ask for a break |
| A fire drill | A loud bell, lining up, walking outside | Follow the teacher and walk |
| Waiting in line | Other people ahead, a turn that comes | Wait, count, ask how long |
| A loud place | The supermarket, the noises, the lights | Wear headphones or ask to step out |
| When the plan changes | A sign changes; a trip is cancelled | Ask what will happen instead |
| Saying hello | Someone waves; people say hello in different ways | Wave, or say hello |
| Sharing a toy | Two children and one toy | Take turns, ask for a turn |
| Riding the bus | Steps, seats, stops, a bell | Sit, hold on, press the bell |
| A visit to the doctor | The waiting room, the scale, the nurse | Ask a question or hold a grown-up's hand |
| The first day in a new classroom | A new room, new faces, a place to sit | Ask a teacher for help |
| Asking for help | Someone nearby who is able to help | Use words or a signal to ask |

## Before you show it

- Watch every second with the sound on. Generated pictures can contain mistakes in hands, faces and objects, and a child may notice what you do not.
- Fix single shots, not the whole video. Edit one shot and regenerate.
- Use the free 5-second preview to check the look and the voice before you render. The free preview.
- Show it to the parent, teacher or specialist first.
- Use an invented character, not a photo of a real child. Uploads need your confirmation that every real person shown has agreed, and the video provider's face check can flag photos of real people. Why a photo can be flagged.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Hospital Ward, Pastoral Healing; Costume Style Soft Everyday; Visual Style Watercolor Storybook, 3D Toy Cartoon; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw Bo, the dentist and the dentist's room as locations; the narrator is an Off-screen character with a voice and no drawing. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Words in a narrator line | Shot length it needs (one sentence) |
| 6 words | 3.5 seconds |
| 8 words | 4 seconds |
| 10 words | 5 seconds |
| 12 words | 5.5 seconds |
| 14 words | 6 seconds |

## FAQ

### What is a social story?

A short, plain story that explains a situation and what a child can do in it, usually written by or with a parent, teacher or specialist.

### Can SceneMixer make a social story video?

It can turn a story you write into a simple narrated video with a cartoon cast and a calm voice. It is a video maker, not a therapy or teaching programme.

### Does a video like this work for my child?

Nothing here promises any result for any child. Ask the child's teacher, therapist or doctor what suits them.

### Should I use my child's photo as the main character?

We suggest an invented cartoon child or animal. Uploads need confirmation that every real person shown has agreed, and photos of real people can be flagged by the video provider.

### Which look is calmest?

Watercolor Storybook or 3D Toy Cartoon: round, bright and slow.

### How long should the video be?

About a minute for one situation, with one idea per shot.

### Can it be in my child's home language?

Yes. In SceneMixer, dialogue and narration can be in any of 15 languages.

### Can I turn the music off?

Yes. SceneMixer's Music switch gives a version without the score.

### Does a free account's video carry a label?

Yes. Videos on a free account carry a small AI label, and a video shared through a SceneMixer link always shows it.

Updated November 3, 2026


---

# AI Anime Fight Scenes: Write Battles That Read on Screen

URL: https://scenemixer.com/guides/make-ai-anime-fight-scenes/
Make anime fight scenes with AI: one move per shot, power-ups as states, energy and impact written as light and effect, and damage shown before and after.

# How to make AI anime fight scenes

Quick answer: An anime fight that works on screen is a chain of short shots, each with one move: the stance, the charge, the clash, the reaction. In SceneMixer, write each beat as its own shot of two to four seconds, describe energy attacks as light and effect (“a ring of blue fire forms around her fist”), make power-up forms a state of the character, and keep injuries out of close-up. Pick an anime visual style in Step 1 and the whole fight stays in it.

Brief for AI Create: An anime battle episode: two former best friends meet in the final of a sword tournament; one of them is fighting to free her brother, the other to keep a promise she cannot explain; one episode of about two minutes, ending on a broken blade.

## Break the fight into beats

| Beat | Shot to adapt |
| The stance | [Shot 1] 3s: Wide shot, a stone arena at sunset, crowd in the stands behind. @Aya and @Ren face each other, swords drawn. |
| The charge | [Shot 2] 2s: Low angle. @Aya launches forward, dust kicking up behind her. |
| The clash | [Shot 3] 3s: Close-up, two blades meet; sparks and a ring of light burst outward. |
| The face | [Shot 4] 3s: Close-up on @Ren gritting her teeth. @Ren (strained): You're holding back. (voice: Ren) |
| The turn | [Shot 5] 3s: Medium shot. @Aya's eyes glow gold as she pushes Ren back across the stones. |
How to write one shot · Fights, crowds and group scenes.

## Power-ups, weapons and damage

- Power-up forms are states. Glowing eyes and white hair after the awakening: a state of the same character. States.
- Signature weapons are props, the same sword every episode. A broken sword is a second state of the prop. Props.
- Energy is described plainly: colour, shape, where it comes from. Describe the cause.
- Damage before and after. Torn clothes and dirt are fine (a state); wounds and blood are blocked. Content rules.

## Look and pace

Anime or Painterly Anime for classic action, Painted Manhua for martial arts and cultivation, Cel-shaded 3D for a game-like fight. Use the 30-second segment length for a long battle so more beats fit into one take. 15 or 30-second segments · Visual styles.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style to match the arena; Costume Style to match the fighters; Visual Style Anime, Painterly Anime; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: fighters with distinctive faces and power-up states; weapons as props; the arena as a location. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Beat | What it shows | Shot length |
| Stance | Who is fighting and where. | 3–4 s |
| Charge | One movement. | 2 s |
| Clash | Impact as light and sparks. | 2–3 s |
| Reaction | A face and a line. | 3–4 s |
| Turn | A power-up or a new move. | 3 s |

## FAQ

### How do I make anime fight scenes with AI?

Write the fight as short shots of one move each, describe energy as light and effect, make power-up forms states, and choose an anime visual style.

### Why not one long fight shot?

AI video does best with one clear action per shot; a long brawl in one take blurs.

### How do I show a power-up?

Add the powered-up look as a state of the character in SceneMixer and pick it from that shot on.

### Can I show blood?

No. SceneMixer blocks wounds and gore; show impact, torn clothes and reactions.

### How do I keep a weapon the same?

Make it a prop with a reference picture; add a broken state if it breaks.

### Which visual style suits fights?

Anime or Painterly Anime; Painted Manhua for martial arts; Cel-shaded 3D for a game look.

### Can I recreate a famous anime fight?

No. Known characters and moves trigger the IP check.

### 15 or 30-second segments?

30 seconds lets more beats fit in one segment for a long battle.

Updated November 3, 2026


---

# AI Brainrot Videos: Absurd Characters of Your Own

URL: https://scenemixer.com/guides/make-ai-brainrot-videos/
Make absurd AI brainrot videos: invent your own hybrid creatures, give them sing-song names and a narrator, keep each clip to 15 seconds, and build a recurring cast.

# How to make AI brainrot videos

Quick answer: Brainrot videos are absurd hybrid creatures (a shark in trainers, a cappuccino that dances ballet) with sing-song names, a dramatic narrator and a few seconds of nonsense story. In SceneMixer, invent your own creature as a non-human character with an exact description, pick Realistic 3D or 3D Toy Cartoon for the glossy look, write the narrator's line, and keep a clip to one segment of up to 15 seconds. Because every creature has a sheet, the same cast can come back in every video, which most brainrot does not manage.

Brief for AI Create: Absurd shorts: Bananino Turbino, a banana with jet-engine legs, races a teapot on wheels called Teiera Velocissima around a kitchen table; an over-serious narrator announces each lap; ten 15-second episodes.

## Invent the creature

| Part | Example |
| What it is | A ripe banana, glossy, life-size |
| What is wrong with it | Two small jet engines for legs, white sneakers |
| Face | Big round eyes, a tiny moustache |
| Name | Bananino Turbino, said with drama every time |

Write all of it into the character description; the sheet draws it once and every shot uses it. Non-human characters · Fruit drama.

## Your own creatures, not the famous ones

The known meme characters are other people's creations; their names and designs can set off the IP check and they are not yours to build a channel on. Invent a creature and you own the joke. IP check.

## The glossy look

Visual Style: 3D Toy Cartoon

Visual Style: Family 3D Animation

Visual Style: Game Engine Realism
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

Realistic 3D gives the uncanny shine; 3D Toy Cartoon is softer and safer for kids. Pick once in Step 1. Visual styles.

## Fifteen seconds of nonsense

| Shot | What happens |
| [Shot 1] 3s | Wide shot, a kitchen table as a race track. Narrator (voice-over, very serious): Bananino Turbino. |
| [Shot 2] 3s | Close-up, the jet engines firing. |
| [Shot 3] 4s | Tracking shot along the table edge; the teapot on wheels overtakes. |
| [Shot 4] 3s | Close-up, Bananino's moustache trembling. |

One clip is one segment of up to 15 seconds, rendered at 10 credits a second on the default tier. The narrator is a voice-only character. Segment length · Voice-only characters.

## What we do not make

The split-screen “study” brainrot (a lecture over game footage) is an editing format, not something SceneMixer generates. Music that rhymes the names is not generated either; the episode gets a background score you can switch off. Background music.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Nordic Minimal, Pastoral Healing; Costume Style None (creatures are described, not dressed); Visual Style Realistic 3D, 3D Toy Cartoon; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: each creature as a character with what it is, what is wrong with it, its face and its name; the narrator as a voice-only character; one tabletop or kitchen as the location. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Creature | The joke |
| 1 | Bananino Turbino | Loses a race to a teapot |
| 2 | Teiera Velocissima | Cannot stop |
| 3 | Both | A rematch on the fridge |
| 4 | A new rival | A toaster that only moves when nobody looks |

## FAQ

### How do I make AI brainrot videos?

Invent a hybrid creature, describe it exactly as a character, add a narrator line and render one 15-second segment per clip.

### Can I use the famous brainrot characters?

Better not: they are other people's creations and may trigger the IP check. Invent your own.

### Which style gives the brainrot look?

Realistic 3D for the glossy uncanny look, 3D Toy Cartoon for a softer one.

### How long is each clip?

In SceneMixer, one segment is up to 15 seconds; longer clips are several segments joined.

### Can the same creatures come back?

Yes. Each has a sheet, so it looks the same in every video.

### Can it speak Italian-sounding lines?

Set Italian as the dialogue language, or keep English with dramatic names.

### Can I make the study split-screen kind?

No; that is an editing format, not generated video.

Updated November 3, 2026


---

# Make AI Fairy Tale Videos: Public-Domain Tales, Your Own Designs

URL: https://scenemixer.com/guides/make-ai-fairy-tale-videos/
Make fairy tale videos with AI: public-domain tales retold with your own designs, Watercolor Storybook and Family 3D styles, and a castle, a forest and a cottage.

# How to make AI fairy tale videos

Quick answer: The classic fairy tales (the Brothers Grimm, Hans Christian Andersen, Perrault, Aesop) are public domain, so you can retell them freely; the famous film versions are not, so the dresses, characters and designs must be your own. In SceneMixer, choose Watercolor Storybook for a picture-book look or Family 3D Animation for a feature-film one, Medieval Castle and Elven Forest for the world, and give the wolf, the witch and the princess their own sheets.

Brief for AI Create: A retelling of Little Red Riding Hood for children aged 4 to 7: the girl is a clever map-maker and the wolf is a lost, hungry, polite wolf who needs directions; gentle, funny, one moral; a five-minute episode.

## Tales you can retell

- Public domain: the Grimm tales, Andersen, Perrault, Aesop's fables, and folk tales from around the world.
- Not free to copy: the designs, songs and characters of famous film and TV versions; they trigger the IP check. Change colours, faces and costumes so they are clearly your own. IP check.
- A twist helps: a modern setting, a role reversal or a new hero makes the retelling worth watching.

## Look and world

| Tale | Scene Style | Visual Style |
| Castles and princesses | Medieval Castle or European Palace | Watercolor Storybook, Family 3D Animation |
| Forest tales | Elven Forest or Pastoral Healing | Watercolor Storybook |
| Sea tales | Undersea Kingdom | Family 3D Animation |
| Fables with animals | Pastoral Healing | 3D Toy Cartoon, Claymation |
Bedtime story series · Cinderella stories.

## The wolf, shot by shot

| Beat | Shot to adapt |
| The path | [Shot 1] 4s: Wide shot, a watercolour forest path, tall pines, a red hood among the trees. |
| The map | [Shot 2] 3s: Close-up. @Ruby unfolds a hand-drawn map. |
| The wolf | [Shot 3] 3s: Medium shot. @Wolf, thin and grey, steps out holding his hat in his paws. |
| The line | [Shot 4] 4s: Close-up on @Wolf. @Wolf (embarrassed): Excuse me. Do you know the way to anywhere with soup? (voice: Wolf) |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Medieval Castle, Elven Forest; Costume Style Medieval Nobility, Fantasy Adventurer; Visual Style Watercolor Storybook, Family 3D Animation; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: each character with your own design and a voice; the cottage, the forest and the castle as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Tale | The twist | Moral |
| Little Red Riding Hood | The wolf is lost, not wicked. | Help, but check the map. |
| The Three Little Pigs | The pigs are architects. | Plan before you build. |
| The Tortoise and the Hare | A city race. | Steady wins. |
| The Frog Prince | The princess is the frog. | Keep your promises. |
| The Elves and the Shoemaker | A modern sneaker shop. | Kindness returns. |

## FAQ

### How do I make fairy tale videos with AI?

Retell a public-domain tale with your own designs, choose Watercolor Storybook or Family 3D Animation, and give each character a sheet.

### Can I use any fairy tale?

The classic tales are public domain; the famous film versions' designs and characters are not.

### Can I make it look like a famous animated film?

No. Studio looks, songs and character designs trigger the IP check.

### Which style suits fairy tales?

Watercolor Storybook for a picture book, Family 3D Animation for a feature look.

### Can animals talk?

Yes. They are characters with voices.

### How long should a fairy tale video be?

Three to eight minutes for a full tale; under a minute for a Short.

### Can I make a series of tales?

Yes. In SceneMixer, choose TV Series and one tale per episode.

### Vertical or landscape?

16:9 Landscape for YouTube; 9:16 Portrait for Shorts.

Updated November 3, 2026


---

# Make AI Talking Animal Videos: Pets and Wildlife with Voices

URL: https://scenemixer.com/guides/make-ai-talking-animal-videos/
Make talking animal videos with AI: real animals drawn as animals with a voice, cartoon animals with faces and hands, a recurring cast, and comedy in short beats.

# How to make AI talking animal videos

Quick answer: A talking animal video is a real animal with a voice that fits too well, or a cartoon animal with a whole personality. In SceneMixer an animal is a character with a sheet and a voice. Say exactly what it is (“a ginger tabby with a white chest”) and whether it has hands and clothes: a real animal is drawn as an animal on all fours; a cartoon animal can wear a scarf and hold a cup. Choose live action for pets, an animation style for a cartoon cast.

Brief for AI Create: A talking animal comedy: a house cat narrates the family's day like a nature documentary presenter while the golden retriever ruins every plan; one-minute episodes, one disaster each.

## Real or cartoon

| Kind | How to describe it | Visual Style |
| Real pets | Breed, colour, markings; on four legs, no hands | Digital HD, Soft Glow |
| Wildlife | Species, size, one mark | Digital HD, Shot on Film |
| Cartoon animals | Face, clothes, whether it has hands | 3D Toy Cartoon, Family 3D Animation, Claymation |
Non-human characters · Cat and pet dramas.

## Voices and comedy

- A voice that fits too well is the joke: a posh cat, a nervous dog. Pick from the preset library. Voices.
- Narration works as voice-over: the cat describing the household. Voice-over.
- Short beats: a look, a line, a disaster. One shot.
- A recurring cast turns clips into a series people follow. Faceless channels.

## The documentary, shot by shot

| Beat | Shot to adapt |
| The kitchen | [Shot 1] 4s: Wide shot, a family kitchen. @Ginger, a ginger tabby, sits on the counter beside a cake. |
| The narration | [Shot 2] 4s: Close-up on @Ginger. @Ginger (voice-over, hushed): Here we observe the cake, unguarded. (voice: Ginger) |
| The dog | [Shot 3] 3s: Medium shot. @Rufus, a golden retriever, skids into the kitchen. |
| The disaster | [Shot 4] 3s: Close-up on the empty plate. @Ginger (voice-over): Nature is cruel. (voice: Ginger) |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Suburban Mansion, Soft Romance; Costume Style Soft Everyday; Visual Style Digital HD, 3D Toy Cartoon; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: each animal as a character with an exact description and a preset voice; the kitchen and the garden as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Disaster | Narrator's last line |
| 1 | The cake. | Nature is cruel. |
| 2 | The vet. | We do not speak of the vet. |
| 3 | The new sofa. | It was beige. Briefly. |
| 4 | The bath. | Betrayal has a smell. |
| 5 | The holiday. | They came back. Eventually. |

## FAQ

### How do I make a talking animal video with AI?

Make the animal a character with a sheet and a preset voice, describe it exactly, and write short comedy beats.

### Will a real dog stand on two legs?

No. SceneMixer draws a real animal as an animal on all fours; cartoon animals can have hands and clothes.

### Can I use my own pet's photo?

Uploading a photo as a character is a membership feature on SceneMixer.

### Can animals narrate?

Yes, as voice-over in the animal's voice.

### Which style suits pets?

Digital HD or Soft Glow for real pets; 3D Toy Cartoon for cartoon animals.

### How long should a video be?

Under a minute for short-video feeds.

### Can I make a series?

Yes. Keep the same animals as recurring characters.

### Can I copy a famous cartoon animal?

No. Known characters trigger the IP check.

Updated November 3, 2026


---

# Make AI Talking Baby Videos: Funny Babies with Grown-Up Voices

URL: https://scenemixer.com/guides/make-ai-talking-baby-videos/
Make funny talking baby videos with AI: the baby as a character with a sheet and a voice, short comedy beats, a 3D or live-action look, and a recurring cast.

# How to make AI talking baby videos

Quick answer: Talking baby videos are comedy: a baby with the opinions of a tired accountant. In SceneMixer the baby is a character like any other, with a sheet that keeps the same face in every video and a voice chosen from the preset library (play them before choosing). Short beats work best: a reaction, a line, a reaction. A recurring baby with a recurring problem (daycare politics, a rival baby) turns clips into a series.

Brief for AI Create: A talking baby comedy series: a nine-month-old who thinks she runs the household holds a daily meeting with the family dog about the parents' behaviour; each episode is one minute, one complaint and one plan that fails.

## The baby as a character

- A sheet with one mark: a single curl, a red bib, a gap in the gums. Describe a face.
- A voice from the preset library; a grown-up voice is the joke. Voices.
- Write what the baby does, sitting, crawling, held by a parent; babies are not written standing and walking.
- The dog is a character too. Talking animal videos.

## Look and format

- Live action: Digital HD or Soft Glow with Soft Romance or Suburban Mansion rooms.
- Cartoon: 3D Toy Cartoon or Family 3D Animation.
- Format: 9:16 Portrait and under a minute for Shorts, Reels and TikTok. Posting.
- Keep it wholesome. Content involving minors is held to the strictest standard. Content rules.

## The meeting, shot by shot

| Beat | Shot to adapt |
| The setup | [Shot 1] 3s: Medium shot, a nursery, morning sun. @Baby Mia sits in her high chair in a red bib; @Biscuit the beagle sits beside it. |
| The complaint | [Shot 2] 4s: Close-up on @Baby Mia. @Baby Mia (tired, deadpan): They put me down at seven. Seven, Biscuit. (voice: Mia) |
| The dog | [Shot 3] 3s: Close-up on @Biscuit tilting his head. |
| The plan | [Shot 4] 3s: Close-up on @Baby Mia narrowing her eyes. @Baby Mia (whisper): Tonight, we cry in shifts. (voice: Mia) |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Soft Romance, Suburban Mansion; Costume Style Soft Everyday; Visual Style Digital HD, 3D Toy Cartoon; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the baby and the dog as characters with one mark and a voice each; the nursery as a location. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | The complaint | The plan that fails |
| 1 | Bedtime is too early. | Crying in shifts. |
| 2 | The new baby next door. | A charm offensive. |
| 3 | Broccoli. | Feeding it to the dog. |
| 4 | Daycare. | A staged sneeze. |
| 5 | Grandma's visit. | Being too cute to leave. |

## FAQ

### How do I make a talking baby video with AI?

Make the baby a character with a sheet and a preset voice, write short comedy beats, and render in 9:16 for short-video feeds.

### Will the baby look the same in every video?

Yes. SceneMixer uses the baby's sheet in every shot.

### Can the baby have an adult voice?

Yes. Pick any voice from the preset library.

### Can I use a photo of my own baby?

Uploading a photo as a character is a membership feature on SceneMixer; only use photos you have the right to use.

### Can the baby walk?

Write babies sitting, crawling or being held; that reads naturally.

### Which style is funnier?

Live action with Digital HD for realism; 3D Toy Cartoon for a cartoon series.

### How long should a video be?

Under a minute for Shorts, Reels and TikTok.

### Can I make a series?

Yes. Keep the same baby and the same dog as recurring characters.

Updated November 3, 2026


---

# Make AI Video Look Realistic: Faces, Light, Cause and Tier

URL: https://scenemixer.com/guides/make-ai-video-look-realistic/
Make AI video look real: a live-action style, faces with specific flaws, light that comes from somewhere, actions written as causes, fewer people, the right tier.

# How to make AI video look more realistic

Quick answer: Realism comes from five decisions, not from magic words. Pick a live-action look (Digital HD or Shot on Film), describe faces with specific, imperfect features, give the light in each shot a source, write actions as physical causes, and keep two or three people per shot. Then render the shots you keep on a higher tier and enhance the best ones up to 2K. “8K, hyperrealistic, masterpiece” adds nothing the model can put in the frame.

Try it: write one character as a man in his fifties, a crooked front tooth, grey stubble, reading glasses pushed up into thinning hair and one shot as late afternoon sun through the blinds, striping his desk, then compare with a generic description.

## Start with a live-action look

| Visual Style | What it looks like |
| Digital HD | Crisp and true-to-life, like today's streaming dramas |
| Shot on Film | Film grain, warm colour, soft blacks |
| Soft Glow | Hazy highlights, pale and dreamy |
| Black & White Film | Rich blacks and greys, like a classic film |

Visual Style: Shot on Film

Visual Style: Soft Glow

Visual Style: Black & White Film
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

Realistic 3D and Game Engine Realism are computer graphics, not live action. In SceneMixer, the look is fixed once you continue past Step 1. Choosing the visual style.

## Faces that belong to someone

A generic description gives the generic AI face: smooth skin, symmetrical, too pretty. Real faces have one or two things wrong with them: a crooked tooth, a broken nose, freckles, uneven eyebrows, a scar. Write those, plus age and build, into the character sheet. Do not aim for a famous actor: a face too close to a real person is refused. Describe a face · Real faces.

## Light with a source

Unreal light comes from nowhere. In each shot, say where it comes from: a desk lamp, the fridge door open, headlights across the ceiling, overcast noon. Time of day and weather come from the script for the scene. Sheets are evenly lit on purpose, so the mood belongs in the shot. Why sheets are plain.

## Write the cause, not the effect

“A dramatic spill” is a picture the model has to invent; the train brakes and her coffee tips into her lap is physics it can follow. Small, physical, ordinary actions read as real. Describe the cause · Say what is there.

## Fewer people, one action

Hands, crowds and five people talking at once are where AI video gives itself away. Two or three named people per shot, one action each; the rest of the room goes in the background line. Background people · Group scenes.

## Tier and resolution

Higher tiers render finer detail and steadier faces. A cheap way to work: draft every segment at 480P, fix what is wrong, regenerate the keepers at a higher tier, then raise the best clips one step at a time with Enhance, up to 2K. Draft at 480P, finish at 1080P · Enhance a clip · Which model.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Street Life, Nordic Minimal; Costume Style Smart Casual, Soft Everyday; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: faces with specific features written into every character; locations described with their layout and light source; the objects that matter as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| It looks fake because | Fix |
| Faces are smooth and generic | Write age, build and one or two imperfections |
| The light is flat or everywhere | Name a light source in the shot |
| Movement looks floaty | Write the physical cause of the action |
| Hands and crowds go wrong | Fewer people per shot, one action each |
| Detail is soft | Higher tier for the keepers, then Enhance |
| Text in the frame is garbled | Keep signs and screens out; add text in your editor |

## FAQ

### How do I make AI video look more realistic?

Choose a live-action look, write specific imperfect faces, give the light a source, write causes, and keep shots to two or three people.

### Do words like 8K or hyperrealistic help?

No. They add nothing the model can place in the frame and push the useful words down.

### Which visual style is the most realistic?

Digital HD for modern; Shot on Film for a warmer cinema look.

### Why do faces look like AI?

Generic descriptions produce average faces; write the details that make someone a person.

### Can I make it look like a real actor?

No. Faces close to real people are refused by the face check.

### Does a higher tier make it more realistic?

Higher tiers give finer detail; draft cheap, regenerate the keepers higher, then Enhance.

### Why does text in the frame look wrong?

Generated video does not spell reliably; add text afterwards.

Updated November 3, 2026


---

# AI Videos for TikTok Shop: Product Clips in Portrait, Several at Once

URL: https://scenemixer.com/guides/make-ai-videos-for-tiktok-shop/
Make TikTok Shop product videos with AI: 15- and 30-second portrait clips from one to four photos, the product in the first second, several tones from one set.

# How to make AI videos for TikTok Shop

Quick answer: A TikTok Shop video is short, vertical and starts with the product. In SceneMixer, upload one to four photos of the item, pick Portrait, 15 or 30 seconds and a tone, and the ad is written, shot and finished with an end card. The product is redrawn as a clean studio shot so it looks the same in every clip; the people holding it are generated. Make three versions from the same photos with different tones (Show how it works, Funny, A day in the life) and let the feed tell you which one sells. Check TikTok's current upload rules in Seller Center before you post; the format and length limits change, and the portrait 9:16 frame is the constant.

One line for the box next to the upload: A collapsible silicone lunch box that fits in a coat pocket; for commuters who hate carrying a bag; ships from our US warehouse.

## What a Shop video has to do

- Show the product in the first second. The feed decides in three. Write it into the line if the photo alone is ambiguous: the blue box is the product.
- Say the one thing it does. One claim, spoken out loud: folds flat, fits a pocket. Price and offer go on the end card and in your caption.
- Portrait, 15 or 30 seconds. Fifteen is one beat (the product and one idea); thirty is a tiny story (problem, product, relief).
- End on the product and a button. The end card carries the brand, the product name and Shop now.
Length, format and quality · The end card.

## Several versions from the same photos

The feed rewards testing. The same four photos make a 15-second Show how it works (the fold, the pocket), a 15-second Funny (a lunch bag that explodes, then yours), and a 30-second A day in the life (a commuter's morning). Each is a separate ad with its own quote; tone does not change the price, length and quality do. Post them a day apart and keep the one that holds attention. UGC-style ads for the hand-held, talking-to-camera feel.

| Tone | TikTok Shop use |
| Show how it works | Gadgets, kitchen things, tools: the demo is the pitch. |
| Funny | Mass-market items; the product is the punchline. |
| A day in the life | Clothing, bags, everyday things worn or used. |
| Twist | You think it is one thing, then it is the product. |
| Mouth-watering close-ups | Snacks, drinks, candles: steam, pour, crunch. |

## The photo and the line

One to four photos of the same item: front, in use, a detail, the packaging. The product is cleaned into a studio shot and reused in every scene; people in your photos are removed and the ad casts its own. The one line wins over the photo when they disagree, so use it to name the product, the buyer and where it ships from. SceneMixer removes claims like best seller or viral from the script. Which photos work · An ad for a physical product.

## Dialogue in the buyer's language

The spoken line is performed in the language you choose, so the same product can have a Spanish clip for one audience and an English one for another: two ads, same photos. Captions are added after export, in TikTok's own tool or yours; nothing legible is drawn into the picture. Subtitles · Text on screen.

## Before you upload

TikTok Shop publishes its own video requirements in Seller Center and changes them; the stable parts are a portrait 9:16 frame, MP4, and short clips. Download the 1080P version for the Shop and the 720P for testing. Members can download without the SceneMixer badge; shared links keep it. Video without a watermark · Download original vs play version.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Picker | Options | Changes the price? |
| Vibe | Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life | No |
| Length | 15, 30 or 60 seconds (one, two or four segments) | Yes |
| Format | Portrait 9:16 or Landscape 16:9 | No |
| Quality | 480P, 720P or 1080P | Yes |

## FAQ

### Can I make TikTok Shop videos with AI?

Yes. Upload one to four product photos in SceneMixer, choose Portrait and 15 or 30 seconds, pick a tone, and the clip is written, shot and finished with an end card.

### How many versions should I make?

Three from the same photos with different tones is a good first test; tone does not change the price.

### Does it make UGC-style videos?

The cast is generated and the tone can be hand-held and plain; see the UGC-style ads answer for the recipe.

### Can the video show the price?

On the end card and in the spoken line, not in the picture. Put the offer in your caption too.

### What format does TikTok Shop need?

Portrait 9:16 MP4 is the constant; check the current length and size limits in Seller Center before uploading.

### Can I use a real model?

Real faces in your photos are removed; the people in the clip are generated to fit the buyer you describe.

### Can I make it in Spanish or another language?

Yes. The spoken line follows the language you pick; make a second ad for a second market.

### What does one clip cost?

In SceneMixer, one figure is quoted before it starts, and it is a ceiling; 15 seconds at 720P is the cheapest useful clip.

Updated November 3, 2026


---

# AI Videos for Your Classroom: Lessons as Short Dramas and Explainers

URL: https://scenemixer.com/guides/make-ai-videos-for-your-classroom/
Make classroom videos with AI: a three- to five-minute lesson as a small drama or an explainer with a recurring guide, in a child-safe look and your language.

# How to make AI videos for your classroom

Quick answer: A classroom video in SceneMixer is a short drama or an explainer built from your lesson: paste the text, pick TV Series, 16:9, a look suited to the age (3D Toy Cartoon for the youngest, Family 3D Animation or Watercolor Storybook for primary, Digital HD for teenagers), and create a recurring guide and a sidekick who carry every episode. Each episode teaches one idea, shows it happening and ends with the sentence to remember. You check the facts; the characters say exactly what the script says. Three to five minutes is the length that fits a lesson slot, and the dialogue can be in the language you teach.

Brief for AI Create: A classroom series for ages 9 to 11: Professor Pim, a nervous owl, and Rook, a confident crow who is usually wrong, test one science question per episode; episode 1: why ice floats; four minutes, one kitchen, one experiment with a glass of water.

## Three kinds of classroom video

| Kind | Shape | Guide |
| An explainer with a guide | One question, one experiment, one sentence to remember | Educational cartoon for kids |
| A scene from history or a book | A short drama of the event or the chapter, with the real stakes | History drama · A shot list for history |
| A language lesson | A situation (a shop, a station) with the target phrases spoken by the characters | Language-learning drama |
| A social-skills or behaviour scene | Two characters, one situation, the better choice shown | Training drama (the same method) |

## A guide who comes back

Children follow a character before they follow a topic. Create a guide and a sidekick in Step 2 and reuse them for the whole term: the same owl, the same crow, the same kitchen. Non-human characters work well and are never refused by the face check. Give the sidekick a flaw (always wrong, always late) and the lesson has somewhere to go. Non-human characters · Reuse a character across projects · A personality the video shows.

## Looks by age

| Age | Visual Style | Why |
| 4–7 | 3D Toy Cartoon, Watercolor Storybook | Round, bright, slow; nothing frightening |
| 7–11 | Family 3D Animation, Anime | Expressive faces, readable experiments |
| 11–16 | Digital HD, Shot on Film, American Comic | Taken seriously; history and literature land better live-action |

Visual Style: Watercolor Storybook

Visual Style: Family 3D Animation

Visual Style: Anime

Visual Style: Digital HD

Scene Style: Campus Youth
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

Content rules are strict and that suits a classroom: no gore, no sexual content, no real people's faces. A historical figure is played by an invented character who looks the part. Content rules.

## The lesson, shot by shot

| Beat | Shot to adapt |
| The question | [Shot 1] 4s: Medium shot, a sunny kitchen. @Pim the owl holds up a glass of water with an ice cube in it. @Pim (puzzled): Why does it float, Rook? |
| The wrong answer | [Shot 2] 3s: Close-up on @Rook the crow. @Rook (confident): Because it is lighter. Obviously. |
| The experiment | [Shot 3] 4s: Close-up on the glass. @Pim drops a second ice cube; both float. Ambient: water, a clink. |
| The idea | [Shot 4] 4s: Two-shot. @Pim: Water gets bigger when it freezes. Same weight, more room, so it floats. @Rook (quietly): I knew that. |

One idea per episode, said plainly at the end. You check the facts; the script is what the characters say. Dialogue in the storyboard.

## Subtitles, language and sharing

Choose the dialogue language in Step 1: the language of instruction, or the language being taught. Subtitles can be added on accounts that have made a purchase, which helps early readers and language classes. Share the episode with a link or download it for the classroom player; shared clips carry a small badge. Subtitles · Share links · Using it as a team.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Pastoral Healing, Campus Youth; Costume Style Soft Everyday, School Uniform; Visual Style Family 3D Animation, Watercolor Storybook; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the guide and the sidekick with one mark and a voice each; the classroom or kitchen as a location; the experiment's objects as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post. Subtitles are available on accounts with a purchase.

| Episode | Question | The one sentence |
| 1 | Why does ice float? | Water takes more room when it freezes. |
| 2 | Where does the rain go? | Down, into the ground, and back up as vapour. |
| 3 | Why is the sea salty? | Rivers carry salt from rocks and leave it there. |
| 4 | How does a seed know which way is up? | It feels gravity and grows against it. |
| 5 | Why do we have leap years? | A year is a little longer than 365 days. |

## FAQ

### Can teachers make videos with AI here?

Yes. In SceneMixer, paste the lesson text, choose TV Series and a look for the age group, create a recurring guide, and generate three- to five-minute episodes.

### Who checks the facts?

You do. The characters say exactly what the script says; nothing is added.

### Which look is right for young children?

3D Toy Cartoon or Watercolor Storybook; Family 3D Animation for primary; Digital HD for teenagers.

### Can it be in the language I teach?

Yes. Choose the dialogue language in Step 1; subtitles can be added on accounts with a purchase.

### Can I show a historical figure?

Played by an invented character who looks the part; real people's faces are not generated.

### Can students appear in it?

No. Real faces are refused; use characters.

### How long should a classroom video be?

Three to five minutes: one idea, one experiment, one sentence to remember.

### What does an episode cost?

Video is 10 credits a second on SceneMixer's default tier; a four-minute episode is about 2,400 credits of video before retakes. The first setup images and first storyboard are free.

### Can a school account be shared?

One account can be used by a small team; see the team guide for how projects and credits are shared.

Updated November 3, 2026


---

# AI 1980s Small-Town Mystery: Bikes, Static and a Town That Knows

URL: https://scenemixer.com/guides/make-an-ai-1980s-small-town-mystery/
An eighties small-town mystery with AI: American Small Town scenes, '80s costume, a film-grain look, props like walkie-talkies, and danger told without harm on screen.

# How to make an AI 1980s small-town mystery

Quick answer: A 1980s small-town mystery is a look, a set of objects and a question the whole town is not answering. In Step 1 pick the scene style American Small Town — wide streets and low buildings, clapboard houses with porch swings, a main street of two-storey brick — and the costume style American ’80s: neon-contrast nylon windbreakers, high-waist washed jeans, wide-shoulder blazers, big hair. The town itself carries no decade, so the eighties come from the clothes, the props and your script. Choose Shot on Film or Soft Glow for the grain and warm colour, Western for the region and English for the dialogue, and put the walkie-talkie, the bicycle and the rotary phone in as props.

Brief for AI Create: A 1980s small-town mystery: on the night of the first snow, the town radio station plays thirty seconds of a voice reading names; one of them is the new girl at school’s mother, who has not lived there for ten years; 8 short episodes, English dialogue, every episode ends on a radio going to static.

## The choices that make it 1985

| Step 1 setting | Choose | Why |
| Scene Style | American Small Town | Midwest or New England: wide streets, low buildings, clapboard houses with porch swings, a brick main street. No decade built in. |
| Costume Style | American ’80s | Neon-contrast nylon windbreakers, high-waist washed jeans, wide-shoulder blazers, big hair. |
| Visual Style | Shot on Film or Soft Glow | Film grain, warm colour and soft blacks, or a hazy low-contrast glow; both read as a decade ago. |
| Target Region | Western | Western names and faces for an American town. |
| Dialogue language | English | Spoken as each clip renders. |
In SceneMixer, scene and costume style are written into the descriptions when the script is read, so choose them first. Why they are fixed afterwards · Every scene and costume style.

## Put the decade in the props

The small town is the stage; what makes it 1985 is what people hold. Make each important object a prop in Step 2 with a reference image, and name it in the shots with @ so it is the same object every time:

- The walkie-talkie with a stubby antenna — the voice that should not be on that channel.
- The bicycles leaning against a fence, and the paper bags on the handlebars.
- The rotary phone on a kitchen wall, with a cord that reaches the whole kitchen.
- The tape: a labelled video cassette or audio cassette that holds the clue.
- The arcade, the diner, the drive-in: scene cards for the places the town meets.
The same prop in every hand · Props and locations in a shot.

## A town is many places and one night

A mystery works best in a few places the audience learns by heart. Give each its own scene card, and keep each segment to one place and one stretch of time.

| Place | What it is for |
| The main street at dusk | Everyone passes it; no one is looking at the same thing. |
| The radio station | The voice, the booth, the glass between. |
| A kitchen | Parents saying less than they know. |
| The school hallway | Rumours, locker conversations, the new face. |
| The drive-in or the edge of town | Where the clue turns out to lead. |
Time and weather live on each segment header: set Night and Fog or Snow for the scene, and the run of segments in it follows. The segment header, field by field.

## Danger without harm on screen

Kids-on-bikes mysteries put young characters near something frightening. The video checks do not allow harm to children, so tell the fear, not the harm:

- The thing glimpsed: a light in the trees, a shape in a window, then the character turning away.
- The disappearance: the bicycle on its side, the porch light left on, the empty swing.
- The chase: pedalling hard, the headlight behind, then the garage door pulled down.
- The adults: let the grown-ups do the scary parts, and keep them off screen when it turns physical.
Invent your own town and your own mystery; characters from a known series can trigger the IP check. If a clip is blocked · Murder mystery · Horror short.

## A radio scene, shot by shot

| Beat | Shot to adapt |
| The street | [Shot 1] 4s: Wide shot, fixed camera. A small-town main street at dusk in light snow, brick storefronts, one neon sign flickering. @Dana wheels her bicycle along the pavement in a neon-green windbreaker. |
| The radio | [Shot 2] 3s: Close-up on the @Walkie-talkie clipped to her handlebars. Static, then a low voice reading names. |
| The name | [Shot 3] 3s: Close-up on @Dana, stopping the bike. @Dana (barely audible): That’s my mom’s name. (voice: Dana) |
| The window | [Shot 4] 3s: Medium shot across the street. The lights go out in the upstairs window of the diner. |
| The static | [Shot 5] 3s: Close-up on the @Walkie-talkie. The voice stops mid-name. Static. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- The signal: a voice on a channel it should not be on, getting closer.
- The new kid: someone who arrives and sees what the town has learned not to look at.
- The ten-year-old case: an unsolved disappearance and a tape that was never played.
- The sheriff who stayed: a quiet adult who knows more than anyone and says less.
- The last night of the fair: the whole town in one place, and one person missing from it.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief in AI Create.
- Step 1: Type Short Drama or TV Series; scene style American Small Town; costume style American ’80s; visual style Shot on Film or Soft Glow; Target Region Western; English dialogue; 9:16 Portrait or 16:9. Read the script.
- Step 2: a specific face and one mark per character; the walkie-talkie, the bicycle and the tape as props. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). Night and weather on the headers.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | The main street in the first snow. | The radio reads her mother’s name, then stops. |
| 2 | The new girl at school. | A locker note: Don’t listen after nine. |
| 3 | The radio station after hours. | The booth is empty and the dial is warm. |
| 4 | An old tape in a kitchen drawer. | The voice on it is the same. |
| 5 | The drive-in at night. | A car with its engine off, its radio on. |

## FAQ

### How do I make an 1980s small-town mystery with AI?

In SceneMixer's Step 1, choose the scene style American Small Town and the costume style American '80s, a film-look visual style, Western and English; then write or generate the script and render episode by episode.

### Does the American Small Town style look like the eighties?

It carries no decade. The eighties come from the American '80s costume style, your props and your script.

### What does the American '80s costume style include?

Neon-contrast nylon windbreakers, high-waist washed jeans, wide-shoulder blazers and big hair.

### Which visual style gives the film look?

Shot on Film gives film grain, warm colour and soft blacks; Soft Glow gives hazy highlights and low contrast.

### How do I keep the same walkie-talkie in every scene?

Make it a prop in Step 2 with a reference image and name it with @ in each shot.

### Can children be in danger on screen?

Not harmed on screen. Tell the fear: the light in the trees, the empty swing, the closed garage door. The video checks stop harm to children.

### Can I borrow a famous series' characters or monsters?

No. Known characters are refused by the IP check; invent your own town and your own mystery.

### Will the kids look the same in every episode?

Yes. Each character is drawn once and that sheet is used in every shot.

Updated November 3, 2026


---

# Make an AI 3D Cartoon Movie: Family 3D, Toy Cartoon and More

URL: https://scenemixer.com/guides/make-an-ai-3d-cartoon-movie/
Make a 3D animated cartoon film with AI: Family 3D Animation, 3D Toy Cartoon, Cel-shaded 3D or Realistic 3D, Film type, a cast that holds, and an original world.

# How to make an AI 3D cartoon movie

Quick answer: A 3D cartoon movie in SceneMixer is a Film project in one of the 3D visual styles: Family 3D Animation for a feature-film family look, 3D Toy Cartoon for toy-like characters, Cel-shaded 3D for a game-like outline, Realistic 3D or Chinese 3D Animation for something else. The scene and costume cards switch to 3D pictures to match. The cast is drawn once and stays the same through every sequence.

Brief for AI Create: A 3D animated film: a pizza-delivery robot in a city of talking pets is the only one who notices the city's pigeons have stopped flying; with a grumpy cat and a nervous puppy he follows the clues to the rooftop; a family comedy told in sequences.

## Choose the 3D look

| Look | Visual Style |
| Feature-film family animation | Family 3D Animation |
| Toy-like, rounded, bright | 3D Toy Cartoon |
| Outlined, game-like | Cel-shaded 3D |
| Lifelike CG | Realistic 3D or Game Engine Realism |
| Chinese 3D fantasy | Chinese 3D Animation |
Studio names are not used in the presets because they trigger the IP check; the presets describe the look instead. Visual styles.

## Film type and the cast

- Type Film for one long story; TV Series for episodes. Series or film.
- Animals, robots and toys are characters with sheets and voices. Non-human characters.
- Locations (the pizzeria, the rooftop) are drawn once in 3D. The same room.
- Render a sequence, watch it, then continue. Checklist.
Animated short films · AI movie generator.

## A comedy beat, shot by shot

| Beat | Shot to adapt |
| The rooftop | [Shot 1] 4s: Wide shot, a sunny rooftop full of pigeons standing perfectly still. |
| The robot | [Shot 2] 3s: Medium shot. @Dot, a small round delivery robot with a pizza box on its head, rolls to a stop. |
| The cat | [Shot 3] 3s: Close-up on @Mr Whiskers, a grey cat, narrowing his eyes. |
| The line | [Shot 4] 4s: Close-up on @Mr Whiskers. @Mr Whiskers (deadpan): I've waited my whole life for this, and it's weird. (voice: Whiskers) |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style American Small Town, New York Metropolis; Costume Style Soft Everyday; Visual Style Family 3D Animation, 3D Toy Cartoon; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: animals, robots and people as characters with one mark and a voice each; the city's key places as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Sequence | What happens | Where it ends |
| 1 | A normal delivery day. | The pigeons stop flying. |
| 2 | The team forms. | The cat agrees, for a fee. |
| 3 | Following the clues. | A trail of breadcrumbs up a fire escape. |
| 4 | The rooftop. | The pigeons are waiting for something. |
| 5 | The big sky. | Everyone flies, even the robot. |

## FAQ

### How do I make a 3D cartoon movie with AI?

Choose a 3D visual style such as Family 3D Animation or 3D Toy Cartoon in SceneMixer's Step 1, set Type to Film, and render the script sequence by sequence.

### Which 3D styles are there?

Chinese 3D Animation, Realistic 3D, Cel-shaded 3D, 3D Toy Cartoon, Family 3D Animation and Game Engine Realism.

### Is there a style named after a famous studio?

No. Studio names trigger the IP check; SceneMixer's presets describe the look instead.

### Can animals and robots talk?

Yes. They are characters with voices.

### Will the cast stay consistent?

Yes. Each sheet is used in every shot.

### Can it be a series instead of a film?

Yes. Choose TV Series as the type.

### Vertical or landscape?

16:9 Landscape for a film.

### Can I use famous cartoon characters?

No. Create your own.

Updated November 3, 2026


---

# AI Action Movie: Chases, Fights and Stunts Shot by Shot

URL: https://scenemixer.com/guides/make-an-ai-action-movie/
Make an action movie with AI: one move per shot, chases built from segments, stunts shown before and after, and settings in Step 1 that give every fight a place.

# How to make an AI action movie

Quick answer: Action on AI video works when every shot does one thing: a run, a jump, a punch, a reaction. Long choreography in one take is where models fail. In SceneMixer, choose Film or Short Drama, a look such as Digital HD or Game Engine Realism, write each beat of a fight or chase as its own shot of two to four seconds, and keep injuries off screen. Positions are tracked across cuts, so the hero stays on the left of the corridor.

Brief for AI Create: An action film: a courier who never opens her packages is chased across a rain-soaked port city after she finally does; one night, three chases, one rooftop, a quiet last shot.

## Give every fight a place

| Kind of action | Scene Style | Costume Style |
| City chase at night | Neon Nights | Streetwear, Near-Future Techwear |
| Office tower siege | Corporate Cold | Power Tailoring, Professional Uniforms |
| Desert or border run | Frontier Vastness or Gulf Metropolis | Outdoor Gear |
| Military operation | Military Base | Modern Military |
| Underground fight club | Retro Hong Kong | Sportswear |
Every scene and costume style.

## One move per shot

Write a fight as a list: the swing, the block, the counter, the fall, the reaction. Each is a shot with its own seconds. A shot that asks for five moves gets a blur of limbs. Chases work the same way: the run, the obstacle, the jump, the landing, the look back. Fights, crowds and group scenes.

Keep the action readable with positions: who is left, who is right, who is higher. Everyone stays put.

## A chase, shot by shot

| Beat | Shot to adapt |
| The start | [Shot 1] 3s: Wide shot, night, rain. @Kai bursts out of a warehouse door and runs left along the wet dock. |
| The pursuers | [Shot 2] 2s: Medium shot. Two men in dark coats come through the same door, running. |
| The obstacle | [Shot 3] 3s: Low angle. @Kai vaults a stack of crates. |
| The look back | [Shot 4] 2s: Close-up on @Kai, breathing hard, glancing over her shoulder. Ambient: footsteps, rain. |
| The escape | [Shot 5] 4s: Wide shot. @Kai drops from the dock edge onto the deck of a moving ferry. |

## Stunts without injuries

Graphic injury is blocked. Show the moment before a hit lands and the moment after: the car spinning, then the driver climbing out. Explosions read best as light and debris behind a character. Content rules · Describe the cause, not the result.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Neon Nights, Corporate Cold; Costume Style Streetwear, Power Tailoring; Visual Style Digital HD, Game Engine Realism; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: one sheet per fighter; a worn or bruised state for the climax; the dock, the tower and the roof as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Act | Set piece | Shots |
| Opening | A chase that sets up the hero's skill | 8–12 short shots |
| Midpoint | A fight in a tight space | One move per shot |
| Low point | The hero loses something: a partner, a car, a plan | A quiet scene |
| Climax | The biggest location: a roof, a bridge, a ship | Wide, close, wide |
| Ending | Aftermath and a last look | Two or three slow shots |

## FAQ

### How do I make an AI action movie?

Write each fight or chase beat as its own short shot, choose a scene style that gives the action a place, and render with Film or Short Drama type.

### Can AI do fight scenes?

Yes, one move per shot. Long choreography in a single take is where current video models break down.

### Can I show blood or wounds?

No. Show the moment before and after the hit; SceneMixer blocks graphic injury.

### Which style suits action?

Digital HD for a modern film look, Game Engine Realism for a video-game feel, Shot on Film for a seventies grade.

### How long should action shots be?

Two to four seconds each; let the cut make it fast.

### Can I do car chases?

Yes, as sequences: the turn, the near miss, the reaction, each its own shot.

### Will the hero look the same through the whole film?

Yes. SceneMixer uses one reference sheet in every shot; damage and dirt can be a separate state.

### Can I copy a famous action franchise?

No. Famous characters and franchises trigger the IP check; invent your own.

Updated November 3, 2026


---

# How to Make an AI Ad from One Product Photo | SceneMixer

URL: https://scenemixer.com/guides/make-an-ai-ad-from-one-photo/
Upload a photo of what you sell and SceneMixer writes the script, casts it, shoots it and adds an end card. What each setting changes, and the rules that keep copy legal.

# How to Make an Ad from One Photo

Quick answer: Upload one to four photos of whatever you are selling — a product, a dish, a car, a shop front, an app screen — and SceneMixer reads them, works out what the thing is, writes a short script around it, generates the cast and locations, shoots each segment and finishes with an end card carrying your brand, product name, tagline and call to action. You pick the length, the shape and the resolution; everything else is one button.

The ad pipeline runs on the same engine as the drama pipeline, so the storyboard it produces is the same working document you can see in this sample project.

## Step 1 — Add the photos

Up to four photos of the same subject: different angles, a detail shot, the thing in use. The first one is treated as the main subject — it is the one the end card is built around, and, for a product, a dish or a garment, the one that gets redrawn as a clean shot on white. A shop, a car or an app screen is not redrawn (see below). The rest become locations.

SceneMixer reads them before anything is charged. If the read comes back unusable — too blurry, nothing recognisable, a screenshot at a strange aspect ratio — you get told on the spot and nothing is spent.

## Step 2 — Say what it is, if the photo is ambiguous

There is a one-line box next to the upload. Use it when the picture alone is misleading: “this is our shop, not a stock photo” changes the whole treatment, because a shop is a place the ad happens inside, while a product is a prop the ad happens around.

Where the sentence and the photo disagree, the sentence wins — this is the one place where what you type outranks what the model sees.

## Step 3 — Pick length, tone, shape and resolution

Length is 15, 30 or 60 seconds, which is one, two or four segments of fifteen seconds. Tone picks the narrative recipe — spectacle, funny, warm, minimal, twist, appetite, demo or lifestyle — and each one carries its own art direction as well as its own structure.

Resolution is 480p, 720p or 1080p. In SceneMixer, the price on each option already includes every step, so the number you see is the number you pay.

## Step 4 — Let it run, or take the wheel

Handing it to the AI runs the whole chain unattended and quotes one price up front: clean the photo, write the script, break it down, generate the cast and sets, shoot each segment, composite with the end card.

The step-by-step mode charges less up front and hands you the normal editor after the script is written, so you can rewrite a line, change a character or reshoot one segment before paying for the rest.

## What it does to your photo, and why

SceneMixer redraws products, food and clothing as a clean studio shot on white so they can be lit consistently in every segment. Shops, restaurants, cars and houses are not redrawn — cutting them out of their surroundings destroys the signage, the depth and the sense of scale that make them recognisable. App and web screenshots are not redrawn either; redrawing turns readable interface into mush.

If there are people in a photo of a place or a vehicle, they are removed before anything else happens. A model given a reference frame copies the faces in it, and putting a real bystander into your ad is not something we will do.

## The copy rules

SceneMixer strips superlatives and absolute claims from the script before it is produced — “the lowest price anywhere”, “number one in sales”, “cures everything”. Brand and product names go through the same pass, but only against that shorter list of outright claims — an ordinary trademark is left alone, while a name that itself makes a claim (“#1”, “100%”, “guaranteed”) is softened.

Every segment carries at least one spoken line, and the tagline is said out loud before the final freeze. Nothing legible is written into the picture — video models cannot render readable text, so the words that have to land are delivered by voice and by the end card.

| Setting | What it does |
| Photos | 1–4 of the same subject; the first is the main one |
| Length | 15s / 30s / 60s → 1 / 2 / 4 segments |
| Tone | Eight narrative recipes, each with its own art direction |
| Resolution | 480p / 720p / 1080p, price shown per option |
| End card | Brand, product, tagline, call to action |
| Music | Written for the ad and mixed in at the final cut |

## FAQ

## Selling for Black Friday?

There is a plan counted back from November 27, with the end card and a second cut for Cyber Monday: Make a Black Friday ad from one photo.

### Do I need a script?

No. SceneMixer writes the script from the photos and the one-line description. In step-by-step mode you can rewrite it before anything is shot.

### Can I use a photo of my shop rather than a product?

Yes, and it is handled differently on purpose: a place becomes the location the ad happens inside, and SceneMixer rebuilds the first scene from your own photo rather than imagining it from text.

### Will the ad show my logo or price on screen?

Only on the end card. Video models cannot render readable text reliably, so SceneMixer puts anything that must be read exactly on the card at the end, and everything else is spoken.

### What happens if my photo has customers in it?

SceneMixer removes them before the shoot for shops, restaurants and vehicles. Screenshots containing people are declined outright rather than edited.

### Can I change the ad after it is finished?

You can recomposite with different end-card text, and in step-by-step mode you can reshoot individual segments. The finished file on SceneMixer can also be downloaded and shared with a link.

Updated November 3, 2026


---

# AI Alien Drama: First Contact, Alien Worlds and Who to Trust

URL: https://scenemixer.com/guides/make-an-ai-alien-or-first-contact-drama/
Make an alien or first contact series with AI: aliens as characters with their own sheets, Alien World and Alien Tribal styles, disguises as states.

# How to make an AI alien or first contact drama

Quick answer: An alien drama is about the first conversation: what the visitor wants and who believes it. Make each alien a character in Step 2 and describe its body plainly (“grey-blue skin, four fingers, eyes that reflect light like a cat's”) so it is the same creature in every shot. Step 1 gives you the worlds: Alien World (purple sky, two moons, glowing fungal forests, crystal pillars) with Alien Tribal costumes that look grown from the body; on Earth, any modern setting. An alien in human disguise is one character with two states.

Brief for AI Create: A first contact drama: a radio astronomer in a small desert town finds an injured alien hiding in her barn; it can only speak through the radio, and the government team arriving tomorrow is led by her brother; 10 episodes, each ends with a new signal.

## Where the aliens meet us

| Kind of story | Scene Style | Costume Style |
| A visitor in a small town | American Small Town or Wild West | Soft Everyday, Outdoor Gear |
| Humans on an alien planet | Alien World | Space Explorer for humans, Alien Tribal for aliens |
| The invasion of a city | New York Metropolis | Modern Military, Smart Casual |
| A colony on Mars | Mars Base | Space Explorer |
| A ship that picked something up | Starship | Starship Uniform |
Space operas · Sci-fi short dramas.

## Aliens as characters

- Describe the body. Skin, limbs, eyes, how it moves; no “otherworldly” or “strange”. Describe the cause.
- A disguise is a state. The alien in a human body and in its own form share one card. States.
- A voice that sounds the same every time. Pick a preset voice, or keep it off-screen through a radio. Voices · Voice-only characters.
- Ships and craft that come back are props or locations with their own picture. Props and locations.
Non-human characters.

## The barn, shot by shot

| Beat | Shot to adapt |
| The light | [Shot 1] 4s: Wide shot, a desert farm at night, a pale blue glow leaking through the gaps of the barn door. |
| The door | [Shot 2] 3s: Medium shot. @Rosa pulls the barn door open with a torch in her other hand. |
| The visitor | [Shot 3] 4s: Medium shot. @Visitor crouches in the straw, grey-blue skin, four long fingers pressed to a wound of light on its side. |
| The voice | [Shot 4] 4s: Close-up on the radio on @Rosa's belt crackling. @Visitor (through static, off-screen): Do not call them. (voice: Visitor) |

## Story shapes that work

- The visitor in hiding: E.T.-shaped, but with grown-up stakes.
- The interpreter: the one person who can talk to them.
- The invasion seen from one apartment.
- The alien romance: a disguise that cannot last.
- The stranded crew on a planet that is not empty.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Alien World, American Small Town; Costume Style Alien Tribal, Soft Everyday; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: each alien as a character with a body description and a disguise state if it has one; the barn, the observatory and the ship as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The observatory. | A signal that repeats her name. |
| 2 | The barn. | The visitor speaks through the radio. |
| 3 | The diner. | Her brother's team rolls into town. |
| 4 | The desert at dawn. | The ship is not where it crashed. |
| 5 | The observatory. | A second signal, from much closer. |

## FAQ

### How do I make an alien drama with AI?

Make each alien a character with its own sheet, pick a scene style such as Alien World or an Earth setting in SceneMixer's Step 1, and keep its voice and look the same with a preset voice and states.

### Will the alien look the same in every shot?

Yes, if it has a sheet in Step 2. Describe its body plainly and give it one mark.

### How do I show an alien disguised as a human?

Two states of one character: the human disguise and the true form.

### What does the Alien World scene style look like?

An alien planet: a purple sky with two moons, glowing fungal forests taller than buildings, crystal pillars, floating rock and fluorescent lakes.

### Can the alien speak?

Yes, with a voice from SceneMixer's preset library, on screen or off-screen.

### Can I show an invasion?

Yes, through short shots of craft, crowds and faces. Keep injuries off-screen.

### Can I use a famous alien design?

No. Known characters trigger the IP check. Design your own species.

### Which visual style suits aliens?

Digital HD or Game Engine Realism for live action; Realistic 3D or Painterly Anime for a drawn look.

Updated November 3, 2026


---

# How to Make an AI Animated Short Film: Full Workflow

URL: https://scenemixer.com/guides/make-an-ai-animated-short-film/
Anime, 3D family film or claymation: pick an animation style that holds, design characters that do not drift, and take an animated short from script to finished file.

# How to Make an AI Animated Short Film

Quick answer: An AI animated short starts the same way as a live-action one — a script, a cast, a storyboard — with one decision that matters more: the style. In step 1 SceneMixer offers three groups of animation presets (2D Animation, 3D Animation and Stop-motion) and the style you pick is drawn into every character sheet, location and shot, on every video tier. Choose Film as the Type for a short with room to breathe, give each character one unmistakable visual trait so animated faces do not drift between shots, draft cheaply — flat, cel-shaded looks tend to hold up at 480P better than painterly ones — and composite the finished film into one file.

Brief for AI Create: An animated short: a lighthouse robot who has never seen another robot finds a small broken drone washed up on the rocks; he spends the winter fixing it; it flies away on the first day of spring and comes back with others; gentle, wordless, hopeful.

## Pick the style first, and carefully

The visual style is fixed once you continue past step 1; to try another, you start a new project with the same script. Look at the sample card for each preset before choosing:

| Group | Presets | Good for |
| 2D Animation | Anime, Watercolor Anime, Painterly Anime, Chinese Painted Style, Ink Wash Animation, Korean Webcomic, Painted Manhua, American Comic, Watercolor Storybook, Classical Oil Painting | Anime shorts, storybook tales, graphic-novel thrillers |
| 3D Animation | Realistic 3D, Family 3D Animation, 3D Toy Cartoon, Cel-shaded 3D, Chinese 3D Animation, Game Engine Realism | Family films, toys and creatures, game-style stories |
| Stop-motion | Claymation | Handmade, tactile, often funny |
Every video tier renders the style; there is no separate animation tier. The presets, one by one · Stylised looks across the tiers.

Visual Style: Watercolor Anime

Visual Style: Painterly Anime

Visual Style: Family 3D Animation

Visual Style: 3D Toy Cartoon

Visual Style: Claymation

Visual Style: Watercolor Storybook

Visual Style: American Comic
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Design characters that hold across shots

Animated faces are simpler than real ones, which makes them easier to confuse: two anime girls with brown hair become one. Give every character one trait nobody else has, and put it first in the description:

- A white streak in her black fringe, round glasses, a yellow raincoat she never takes off.
- A short, round robot with one large blue eye-lens, a dent on his left side, rust on his feet.
- A tall fox in a green scarf, one ear torn at the tip.
Sheets are drawn on a plain background with even light on purpose: they are references, not posters. The mood comes in the shots. Why sheets are plain · Robots, animals and creatures.

Character sheets from the sample Rust City Knockout: one reference image per character, used in every shot they appear in.

## Write for animation

- Bigger expressions. Animation reads emotion from the whole body. Delivery notes and actions can be broader than in live action: he throws both arms up, her ears flatten.
- Wordless is fine. Many animated shorts have no dialogue at all. Mark those shots No dialogue and let the Ambient line carry the sound: wind, a creak, a beep.
- Light as storytelling. Animation lives on colour. Write the light into each shot: the lighthouse beam sweeps across the dark sea, spring sunlight turns the rocks gold.
- One action per shot. Even in a cartoon, a jump, a landing and a fall are three shots.

| Beat | Shot to adapt |
| The discovery | [Shot 2] 4s: Medium shot, low angle. @Beacon kneels on the wet rocks and lifts the @Drone in both hands; its single light flickers. No dialogue. Ambient: waves, a faint electronic chirp. |
| The winter | [Shot 5] 5s: Wide shot, fixed camera, the workshop at night. @Beacon at the bench, the @Drone in pieces under a lamp; snow against the window. No dialogue. Ambient: a soldering hiss, the wind. |
How to write one shot.

## Settings to pick in step 1

- Type: Film — slower pace, no episode cap, written as one piece. Short Drama, TV Series or Film.
- Aspect Ratio: 16:9 Landscape for a film; 9:16 Portrait for an animated series on phones.
- Scene Style and Costume Style: they work with any visual style. A storybook village (Pastoral Healing), a sci-fi station (Starship), a magic school (Magic Academy). Set all three together; a period costume in a modern set reads as a mistake.
- Region and language: pick where the world is and what the cast speaks; animated characters speak any of the dialogue languages natively.

## Draft cheaply, finish the keepers

Generate the whole short on the default tier first — 10 credits a second — and watch it cut together. Flat, cel-shaded styles tend to hold up at 480P; painterly and finely textured ones usually gain more from a higher tier, so regenerate the shots that will be seen most at a higher resolution. Regenerating one segment costs only that segment. Draft at 480P, finish at 1080P.

## Make it in SceneMixer, start to finish

- Script: your own on Upload Script, or a brief in AI Create with the character, what they want, what is in the way and the ending.
- Step 1: Type Film, the aspect ratio, and the animation style you chose from the sample cards; read the script.
- Step 2: generate the cast with one distinguishing trait each, the locations and the props that recur. The first 15 setup images on an account are free.
- Storyboard: read it like an animator: one action per shot, light in every shot, No dialogue where nobody speaks.
- Video and finish: generate, regenerate what misses, composite, download the original. Composite and share.
See two animation styles in finished samples: Rocket Pup in Family 3D Animation, and Rust City Knockout in Painterly Anime.

| Story | Style | Why |
| A family adventure with animals | Family 3D Animation | Rounded, expressive, made for all ages |
| A teen drama or romance | Painterly Anime or Watercolor Anime | Luminous light, emotional close-ups |
| A bedtime or picture-book tale | Watercolor Storybook | Soft, gentle, readable for children |
| A dark thriller | American Comic | Heavy ink, hard shadows |
| A comedy with props | Claymation or 3D Toy Cartoon | Tactile, playful, forgiving |
| A Chinese fantasy | Ink Wash Animation or Chinese 3D Animation | Brushwork or immortal-fantasy CG |

## FAQ

### How do I make an animated short film with AI?

Write or generate a short script, choose Film and an animation style in step 1, design each character with one distinguishing trait, then generate the storyboard and the video segment by segment and composite them into one file.

### Which animation styles are there?

2D Animation (ten presets from cel shading to oil painting), 3D Animation (six presets from Family 3D Animation to Game Engine Realism) and Stop-motion (Claymation).

### Can I change the style halfway through?

No. The visual style is fixed once you continue past step 1. Start a new project with the same script to try another style.

### Is there a special animation video tier?

No. Every video tier on SceneMixer renders the project's style; tiers differ in resolution, segment length and price.

### Why do my animated characters look alike?

Animated faces are simpler, so give each character one trait no one else has and put it first in the description, then regenerate the sheet.

### Can an animated short have no dialogue?

Yes. Mark shots No dialogue and use the Ambient line for sound. Many animated shorts work best that way.

### Is animation cheaper than live action?

No. On SceneMixer, prices are per sheet and per second whatever the style.

### Can animated characters speak other languages?

Yes. Choose the dialogue language in step 1; the characters speak it natively.

Updated November 3, 2026


---

# Make an AI Anime Movie: A Feature-Length Script in a Drawn Style

URL: https://scenemixer.com/guides/make-an-ai-anime-movie/
Make an anime film with AI: Film type for movie pacing, a long script read into sequences, an anime visual style held across every shot, and one cast throughout.

# How to make an AI anime movie

Quick answer: An anime movie in SceneMixer is a Film project in a drawn style. Set Type to Film in Step 1 for movie pacing without episode cliffhangers, choose Painterly Anime or another anime look, and upload the screenplay: the read splits it into sequences you storyboard and render one at a time. The cast is drawn once and holds across the film; length is stacked from segments rather than requested in one go, so the bill follows what you actually render.

Brief for AI Create: An anime film: on the last night before a seaside town is flooded for a dam, a girl and her late grandfather's robot try to deliver forty letters he never sent; one night, forty doors, a feature told in sequences.

## Film type and the drawn look

- Type Film: pacing for one long story, no forced hooks at the end of each part. Short drama, series or film.
- Visual style: Painterly Anime for a cinematic look, Watercolor Anime for a gentle one, Anime for classic. It cannot change after Step 1. Visual styles.
- Aspect: 16:9 Landscape for a film. Aspect ratio.
The AI movie generator explains how long a script it reads and how film mode differs.

## Holding a feature together

- One cast, many sequences. Sheets are used in every shot of every sequence. Why a whole series stays consistent.
- Places that come back (the harbour, the post office) are locations with one picture. The same room.
- Sequence by sequence. Storyboard and render the first sequence, watch it, then continue. Checklist before you generate.
- Cost: video is charged per second rendered; check the price before you render a sequence. Where credits go.

## An opening, shot by shot

| Beat | Shot to adapt |
| The town | [Shot 1] 5s: Wide shot, a seaside town at dusk under a painted orange sky, a half-built dam above the valley. |
| The letters | [Shot 2] 3s: Close-up, a tin box of forty yellowed letters on a kitchen table. |
| The robot | [Shot 3] 3s: Medium shot. @Tobi, a small round copper robot, lifts one letter with two careful fingers. |
| The line | [Shot 4] 4s: Close-up on @Mio. @Mio (quiet): Then we deliver all of them. Tonight. (voice: Mio) |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style to match the world; Costume Style to match the world; Visual Style Painterly Anime, Watercolor Anime; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the cast drawn once with distinctive faces; recurring places as locations; render the first sequence before the rest. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Sequence | What happens | Where it ends |
| 1 | The town's last evening. | The tin of letters. |
| 2 | The first ten doors. | A letter to a woman who moved away. |
| 3 | The harbour. | The robot's battery is failing. |
| 4 | The dam. | The last letter is to Mio. |
| 5 | Dawn. | The water starts to rise. |

## FAQ

### How do I make an anime movie with AI?

Set Type to Film and an anime visual style in SceneMixer's Step 1, upload the screenplay, and storyboard and render it sequence by sequence.

### Can it be feature length?

Length is stacked from segments; long scripts are read into many sequences you render one by one.

### Will characters look the same for the whole film?

Yes. Each sheet is used in every shot of every sequence.

### Which anime style suits a film?

Painterly Anime for a cinematic look, Watercolor Anime for a gentle one, Anime for classic.

### What is the difference between Film and Short Drama?

Film paces one long story without a hook at the end of each part; Short Drama builds episodes with cliffhangers.

### How much does it cost?

SceneMixer charges video per second rendered; the price shows before you render each part.

### Can I adapt an existing anime film?

No. Known characters trigger the IP check. Write an original story.

### Vertical or landscape?

16:9 Landscape for a film.

Updated November 3, 2026


---

# Make an AI Anime Series for YouTube: Same Characters, Every Episode

URL: https://scenemixer.com/guides/make-an-ai-anime-series-for-youtube/
Make an original AI anime series for YouTube: Anime, Painterly Anime or Korean Webcomic styles, a cast that stays on model, episodes joined into one video, and Shorts.

# How to make an AI anime series for a YouTube channel

Quick answer: An anime channel grows on a series, not on clips: the same characters, a story that continues and a new episode on a fixed day. In SceneMixer one project is one show. Pick the drawn look in Step 1 (Anime, Painterly Anime, Watercolor Anime, Korean Webcomic or Painted Manhua) and it is fixed for every sheet and shot; the cast is drawn once in Step 2 and stays on model; each episode is storyboarded, rendered and joined into one video, and a 9:16 cut of the best moment becomes the Short.

Brief for AI Create: An original anime series: a delivery girl in a floating city discovers her bicycle can ride on wind currents only she can see; a sky-pirate crew wants the bicycle and a city guard wants her; 12 episodes of about two minutes, each ending on a chase.

## Pick the drawn look once

| Look | Visual Style | Suits |
| Classic anime | Anime | Action, comedy, school stories |
| Rich, painted backgrounds | Painterly Anime | Cinematic fantasy, sci-fi, romance |
| Soft and gentle | Watercolor Anime | Slice of life, healing stories |
| Webtoon look | Korean Webcomic | Romance, office, revenge |
| Chinese comic look | Painted Manhua | Cultivation, palace, martial arts |
The visual style is fixed once you continue past Step 1, and every tier keeps it. Scene and costume styles change to drawn pictures to match. Visual style presets · Anime looks across tiers.

## Characters that stay on model

- A face that is not generic. Hair shape and colour, eye colour, one mark (a bandage on the cheek, a hairpin). Describe a face.
- Outfit changes are states. School uniform and battle outfit. States.
- Your own designs can be uploaded as sheets on a membership plan. Upload your own image.
- Voices from the preset library stay the same across episodes. Voices.

## A channel rhythm that works

- Season plan: 8–12 episodes, each ending on a hook. Hooks.
- Landscape for the channel: 16:9 Landscape for full episodes; the Short is a 9:16 project or a crop of the best moment. Vertical or landscape.
- Join each episode into one video and download it. Composite an episode.
- Post on a fixed day. Posting to YouTube Shorts, TikTok and Reels · Can you monetise AI videos on YouTube and TikTok?.

## Keep it original

Characters, names and iconic outfits from existing anime are refused by the IP check, and studio names are not used in the presets for the same reason. Your channel is safest, and easiest to grow, with a world of your own. IP check · Faceless YouTube channels.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style to match the world; Costume Style to match the world; Visual Style Anime, Painterly Anime; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the cast drawn once, each with a distinctive face, a mark and outfit states; the recurring places as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | What happens | The ending beat |
| 1 | A delivery goes wrong over the canal. | Her bicycle lifts off the ground. |
| 2 | She learns to ride the currents. | A sky-pirate flag on the horizon. |
| 3 | A chase through the market towers. | The guard captain sees her face. |
| 4 | The pirates' ship. | Their captain has the same bicycle. |
| 5 | The city council. | A reward on her head. |

## FAQ

### How do I make an AI anime series for YouTube?

Pick an anime visual style in SceneMixer's Step 1, draw the cast once in Step 2, storyboard and render each episode, join it into one video and post on a fixed schedule.

### Will my characters look the same in every episode?

Yes. Each character's sheet is used in every shot; give each one a distinctive face and mark.

### Which anime styles are there?

Anime, Painterly Anime, Watercolor Anime, Korean Webcomic and Painted Manhua, plus other drawn styles such as American Comic and Ink Wash Animation.

### Can I use my own character designs?

Yes, on a SceneMixer membership plan: upload them as the character's sheet.

### Can I make fan anime of an existing series?

No. Known characters and names are refused by the IP check.

### Can I monetise it?

Platform rules decide; an original continuing series is the strongest case. See our answer on monetising AI videos.

### Vertical or landscape?

16:9 Landscape for full episodes, 9:16 Portrait for Shorts.

### How long should episodes be?

One to three minutes works well for a new channel; each ends on a hook.

Updated November 3, 2026


---

# AI Bible Story Videos: How to Make a Bible Story Series, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-bible-story-series/
Turn Bible stories into an AI video series: an epic live-action look or a storybook for children, consistent characters, narration and lines, and what to watch for.

# How to Make an AI Bible Story Series

Quick answer: An AI Bible story series works best as one well-known story per episode or a short arc split into episodes — Joseph and his brothers, Ruth and Naomi, David and Goliath — told with a narrator and a few spoken lines. In SceneMixer you paste the passage or your own retelling as the script; it is read into a cast, the places and an episode outline, each character is drawn once and stays the same in every shot, and every episode is rendered and joined into one video. Pick the look in step 1: Digital HD or Shot on Film for an epic, Watercolor Storybook or Family 3D Animation for children.

Brief for AI Create: The story of Joseph for children, in four episodes: the coat, the pit, the dream of the king, the brothers in Egypt; a warm narrator, short lines, every episode ends on a question.

## Choose who it is for, then the look

| | Family and children | Adult audience, epic |
| Visual Style | Watercolor Storybook, Family 3D Animation, 3D Toy Cartoon | Digital HD, Shot on Film, Classical Oil Painting |
| Episode length | One to two minutes, one story beat | Two to four minutes, a full scene |
| Telling | A warm narrator carries it; short lines | Mostly dialogue; narration for time jumps |
| Violence | Implied, off screen | Shown briefly, never in detail |
SceneMixer fixes the visual style once you continue past step 1, so decide the audience first. Visual style presets.

## Set the world in step 1

- Scene Style: Ancient Egypt for the Exodus and Joseph in Egypt; Middle Eastern Souq and Silk Road City for towns and markets; Frontier Vastness for the desert and the wilderness; Greco-Roman for the New Testament’s Roman world; Epic Scale for armies and temples.
- Costume Style: Middle Eastern Robes for most stories; Ancient Egyptian Attire for Pharaoh’s court; Greco-Roman Attire for Roman soldiers and officials.
- Target Region: Middle East so faces, names and places stay in the story’s world, or Follow script if your retelling decides.
These are starting directions, not costumes for one person: each character’s own clothing is still written on their card. Scene style and costume style.

Visual Style: Family 3D Animation

Visual Style: Shot on Film

Scene Style: Ancient Egypt

Scene Style: Middle Eastern Souq

Costume Style: Middle Eastern Robes

Costume Style: Ancient Egyptian Attire
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Characters who stay the same for the whole series

A series that follows Joseph from boy to governor needs him recognisable in every episode. Give each character a face that can be drawn and one mark: a lean young man, olive skin, dark curly hair to the shoulders, a narrow face with heavy brows, a small scar on his chin; a long robe of many coloured bands. Older or changed versions of the same person — Joseph in Pharaoh’s court, years later — are states of the same character, drawn from the main sheet, so the face carries over. Costume changes and scene states · Make a face specific.

For a children’s series, keep the cast small — four or five named characters per story — and let crowds stay crowds. Background people in a scene.

## Narration and lines

Bible stories are usually told, not only acted. Write the narrator as a voice-over line and keep the characters’ own words short. A voice-over is heard over the shot and nobody on screen mouths it.

| Beat | Shot to adapt |
| The narrator sets it up | [Shot 1] 5s: Wide shot, slow push in. A dusty well in an empty plain at noon. In frame only: @Brothers, gathered around the well. @Narrator (voice-over, warm): His brothers saw him coming from far away. Ambient: wind, sheep bells. |
| One spoken line | [Shot 2] 3s: Medium shot. @Reuben steps forward, hand raised. @Reuben (urgent, low): Do not hurt the boy. Ambient: wind. |
| The image that ends the episode | [Shot 3] 4s: Close-up. The coat of many colours lies in the dust. No dialogue. Ambient: wind dropping to silence. |
The narrator is a character like the others, marked Off-screen: it gets a voice and no image, and the same voice tells every episode. If the reading does not create one, add it on SceneMixer's Characters tab. Leads, supporting and voice-only characters · Lines, voice-over and captions · Character voices.

## Miracles and set pieces

Big moments — the sea parting, the walls falling, the burning bush — work as a sequence of short shots rather than one long one: the people reacting, the thing itself in a wide shot, a detail, the people again. Describe what is physically there (two walls of water, the dry sea bed, the people walking between them), not what it means. Writing crowds and group scenes · Say what is there, not what is not.

## Content checks to know about

Every shot and description is checked. The checks look at violence shown in detail, sexual content, real public figures and political slogans; the brutal parts of some stories — battles, the flood, the crucifixion — should be implied or shown at a distance, not in close detail. We have not tested every biblical name against the checks: if a description or a shot is refused, it is not charged, and rewording it as a description (the shepherd boy, the old prophet) is the usual fix. Content rules and checks · Blocked by moderation.

## Make it in SceneMixer, start to finish

- Script: paste your retelling or the passage on Upload Script, or give AI Create a brief that names the story, the audience and how many episodes.
- Step 1: Type TV Series for a multi-story run or Short Drama for one story in episodes; the visual, scene and costume styles above; 9:16 Portrait for Shorts and Reels, 16:9 for YouTube; the dialogue language. Then read the script.
- Step 2: check every character has a drawable face and one mark; regenerate any sheet that looks generic. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project); make the narrator a voice-over line.
- Video: generate on the default tier at 10 credits a second; composite; post.

To keep going, add the next story to the same project so the cast carries over. Make the next episode · Post to YouTube, TikTok and Reels.

| Episode | The story beat | How it ends |
| 1 | Joseph’s dreams and the coat | The brothers watch him walk away in it. |
| 2 | The pit and the caravan | The coat in the dust; Jacob’s cry heard far off. |
| 3 | Prison and the king’s dream | “Bring me the one who reads dreams.” |
| 4 | The brothers in Egypt | Joseph turns away so they will not see him weep. |
| 5 | “I am Joseph.” | The whole family on the road to Egypt. |

## FAQ

### How do I make AI Bible story videos?

Paste a passage or your own retelling as a script, or give SceneMixer's AI Create a short brief. Choose a storybook or 3D style for children, or a live-action style for an epic, then generate episode by episode.

### Can I make Bible stories for kids?

Yes. On SceneMixer, pick Watercolor Storybook or Family 3D Animation in step 1, keep episodes short, let a warm narrator carry the story, and keep any violence off screen.

### Will Moses look the same in every episode?

Yes, if his card describes a specific face and one mark. SceneMixer uses the same image in every shot where he is named with @; older versions of him are states drawn from that sheet.

### Can a narrator tell the story?

Yes. Write the narrator's lines as voice-over: they are heard over the shot and nobody on screen mouths them.

### Which aspect ratio for YouTube?

16:9 for regular YouTube videos, 9:16 Portrait for Shorts, TikTok and Reels. It is set in SceneMixer’s step 1.

### Will biblical names be blocked?

We have not tested every name. The checks target real public figures, political slogans, sexual content and detailed violence. If SceneMixer refuses something, it is not charged; describing the person instead of naming them usually passes.

### Can I make it in another language?

Yes. Choose the dialogue language in step 1 and the lines and narration are spoken in it; 15 languages are available on SceneMixer.

### How much does an episode cost?

Video is priced per second of the tier you pick; SceneMixer's default tier is the cheapest. The generate button shows the price before you spend anything.

Updated November 3, 2026


---

# AI Billionaire Romance Drama: Contract Marriage to Face-Slap

URL: https://scenemixer.com/guides/make-an-ai-billionaire-romance-drama/
The CEO, the contract, the gala and the in-laws who regret it: how to write a billionaire romance short drama and generate it in SceneMixer, scene by scene.

# How to Make an AI Billionaire Romance Drama

Quick answer: A billionaire romance is a fantasy of power changing hands: an ordinary heroine is humiliated by people with money, a man with more money takes her side, and every episode ends with someone being proved wrong about her. To make one in SceneMixer, write those reversals into the script, pick a luxury set and wardrobe in step 1 (Urban Opulence or Corporate Cold with Power Tailoring or Haute Couture), turn the contract and the ring into props with their own images, and give the heroine’s transformation — the gala gown — its own state. Describe wealth by what it looks like, never by brand name: a named luxury brand trips the IP check, a black long-wheelbase sedan does not.

Brief for AI Create: A billionaire romance: a temp who spills coffee on the CEO is offered a one-year contract marriage to stop his family's board coup; his mother wants her gone by the gala; she wins the board vote for him and walks out; sharp and glossy.

## The engine of the genre: humiliation, then the reversal

Viewers do not come for the money. They come for the moment the money stops protecting the people who were cruel. Build every episode around one injustice and one answer to it:

- The contract. A marriage, a pretend engagement or a deal with a deadline. It explains why two people from different worlds keep sharing scenes.
- The hidden status. Either the CEO hides who he is, or she does — the “poor” girl is the heiress of a bigger family. The reveal is a guaranteed payoff.
- The gatekeepers. A mother-in-law, an ex-fiancée, a snobbish assistant, a stepsister. Each one is a villain the audience waits to see answered.
- The face-slap. The short-drama term for a public reversal: the card declines for the bully, the “nobody” is announced as the new owner, the evidence plays on the big screen.
- The ex who regrets. Someone who discarded her early on comes back when she is untouchable. Their regret is a hook for several episodes.
Keep the lead couple’s scenes short and charged; keep the villains’ scenes loud and public. The contrast is the rhythm.

## Settings to pick in step 1

- Type: Short Drama — this genre lives on fast reversals and a hook every episode.
- Aspect Ratio: 9:16 Portrait if you will post to phones; the face-to-face close-ups the genre depends on fill a vertical frame well.
- Visual Style: Digital HD for the polished app look, Shot on Film if you want a richer grade.
- Scene Style: Urban Opulence for penthouses and galas, Corporate Cold for glass boardrooms, Suburban Mansion for the family estate, New York Metropolis for a city-skyline story.
- Costume Style: Power Tailoring for the CEO and his board, Haute Couture for galas and the rival heiress, Minimal Luxury for quiet-money elegance. The heroine’s early, plainer clothes go in her own character description.
- Target Region: Western for English names and Manhattan-style streets; Middle East with Gulf Metropolis for a Gulf-billionaire story.
In SceneMixer, the scene and costume styles are read with the script and cannot be changed afterwards in that project. When scene and costume style take effect.

Scene Style: Urban Opulence

Scene Style: Corporate Cold

Scene Style: Suburban Mansion

Costume Style: Power Tailoring

Costume Style: Haute Couture
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Show money without naming it

Wealth is the genre’s scenery, and the easiest way to get a segment blocked is to write it as a shopping list of brands. Well-known brand names and logos trigger the provider’s IP suspected check, the same way celebrity faces do. Describe the object instead:

| Instead of | Write |
| A brand-name supercar | A low matte-black two-door sports car, doors open upward |
| A named luxury handbag | A structured orange leather handbag with a gold clasp |
| A named watch | A heavy steel watch with a deep blue dial |
| A famous hotel by name | A rooftop bar forty floors up, glass railings, the city lit below |
| A brand-name black card | A matte black metal card with no visible lettering |
The same rule applies to your props and location descriptions in step 2. “IP Suspected”: celebrity faces, brands and franchises.

## The contract, the ring and the gown

Three things recur through a billionaire series, and each needs to look the same every time it appears.

- The contract and the ring are props. If the story keeps returning to them, make sure each has a card in step 2 with its own image, and name it with @ in the shots where it is handled. A ring that changes design between episodes breaks the romance faster than a bad line. Same prop in every hand · When a prop gets a reference image.
- The gala gown is a state. A full change of outfit changes what the reference image shows, so it needs its own image: on the heroine’s card choose Add a state, name it Gala gown, and write only the difference — a floor-length midnight-blue silk gown, bare shoulders, hair up, small diamond earrings. Then, in the editor for the gala episode, open her from the episode’s assets and choose Use this state in this episode. One outfit per character sheet.
- The text on paper is not for reading. A model draws a contract, not its clauses. Let the signature, the seal and the reaction carry the scene, and put what the document says into the dialogue. The same goes for a phone screen with a bank alert: show the face reading it, say the number aloud. Text in the picture.
Character sheets from the sample The Glass Floor: one reference image per character, used in every shot they appear in.

## Three scenes every billionaire drama needs, written as shots

These follow the storyboard shape — marker and seconds, framing, who is in frame, one action, the line — so they can be pasted into a segment and adapted.

| Scene | Shot to adapt |
| The boardroom humiliation | [Shot 2] 5s: Medium shot, fixed camera. @Vivian at the head of the glass table, @Nora standing by the door holding the coffee tray. @Vivian (smiling, not looking up): The help can wait outside. @Nora lips closed, no lip movement. Ambient: air conditioning hum. |
| The entrance | [Shot 5] 4s: Wide shot, slow push in. The ballroom doors open; @Nora steps in wearing the gown, alone. The guests turn toward her. No dialogue. Ambient: a string quartet stops mid-phrase. |
| The reversal | [Shot 8] 4s: Close-up. In frame only: @Vivian. Her smile freezes as the announcement plays over the speakers. No dialogue. Ambient: a microphone squeal, then silence. |
The guests in the entrance shot belong in the segment’s background line, not as named characters — a crowd of named people is a crowd of faces to keep consistent. Background people · How to write one shot.

## Voices: cold money, warm heroine

Write each Voice description as a contrast to the others. The CEO: low, even, clipped, never hurries. The heroine: warm and quick, a laugh close to the surface, steadies when she is cornered. The mother-in-law: precise, cultured, every word a little too polite. Delivery that changes line to line — a whisper, a line through the teeth — goes in the brackets at the start of that line. How character voices work · Dialogue in the storyboard.

## Make it in SceneMixer, start to finish

- Script: paste yours on Upload Script, or let AI Create write one from a brief that names the contract, the villain and the ending. Writing the one-line brief.
- Step 1: pick the settings above and read the script. In the summary, check that every episode has a reversal and ends on a hook.
- Step 2: generate the cast, the penthouse, the boardroom and the ballroom; give the ring and the contract their images; add the gown state. Your first 15 setup images on the account are free.
- Storyboard: generate episode 1 (free on your first project) and tighten it: open on the conflict, end on the turn.
- Video: generate the segments, starting with the default tier at 10 credits a second; regenerate only the shots that miss.
- Finish: composite, download and post. Plan the next episode while the first one is out. Make the next episode.
For the look of a glass-walled corporate world built this way, watch The Glass Floor, a corporate thriller where a firing turns into a public accusation, and The Veil at Dawn, where a woman walks out on a wedding in a marble-and-gilt mansion.

| Episode | The injustice | The answer | The hook |
| 1 | The temp is blamed for the CEO’s ruined deal and fired in front of the floor. | He stops her at the lift with a contract. | “Marry me for a year.” |
| 2 | His mother calls her a gold-digger at the family dinner. | She answers with the one fact about the company nobody else knew. | The ex-fiancée walks in. |
| 3 | The ex-fiancée plants a story in the press. | The heroine turns it around on a live call. | Her own father appears on the news — as the owner of a rival firm. |
| 4 | The board moves to vote him out. | She brings the swing vote to the meeting. | The vote is tied. |
| 5 | The gala: the family plans to humiliate her in public. | She arrives in the gown, and is announced by her real surname. | He realises she never needed the contract. |

## FAQ

## When she is the one hiding something

If the heroine’s secret is her family or her name, plan it as a ladder of who knows what: Make an AI heiress drama.

### Can I name real luxury brands in my billionaire drama?

Not in anything that gets generated. Brand names and logos trigger the IP check on images and video. Describe the object by colour, shape and material and the scene still reads as expensive.

### How do I keep the heroine's gala dress the same in every shot?

Make the dress a state of her character with its own image, then assign it to the gala episode or to the segments that need it. Everything outside those segments keeps her main look.

### Can the contract be readable on screen?

Do not rely on it. Generated documents and phone screens rarely hold legible text. Show the signature and the reaction, and let the dialogue say what the paper says.

### Should a billionaire drama be vertical or horizontal?

Vertical, 9:16, if it is going to TikTok, Reels or short-drama apps. Horizontal suits YouTube and a big screen. Pick in step 1; it is fixed for the project.

### How do I make the ballroom look full?

Put the guests in the segment's background line instead of naming them as characters. Name only the people whose faces matter to the scene.

### Can the dialogue be in another language for another market?

Yes. Choose the dialogue language in step 1. For a second market, start a second project from the same script with another language and region.

### How many episodes should a billionaire romance have?

Enough for the contract to run its course; many AI-made series run twenty to sixty short episodes. Make the first three, post them, and continue the series in the same project.

### Can AI Create write the script for me?

Yes. Give it the genre, the heroine and what she wants, the villain, the ending and the tone in one or two sentences, pick a length and press Start Writing.

Updated November 3, 2026


---

# AI Birthday Video: A Short Story Starring the Birthday Person

URL: https://scenemixer.com/guides/make-an-ai-birthday-story-video/
Make a birthday video with AI that tells a story: the birthday person as the hero of a short animated adventure, a spoken wish at the end, and what photos can do.

# How to make an AI birthday story video

Quick answer: A birthday video people watch twice is a short story with the birthday person as the hero: one adventure, thirty seconds to two minutes, a line at the end. In SceneMixer, describe them in words (hair, glasses, the favourite jumper, the dog), pick a cartoon look such as Family 3D Animation or Watercolor Storybook, write a one-line brief, and render two to four segments. Uploading their photo is a paid-plan feature that needs their consent, and a realistic photo of a real person is refused by the video's face check, so a drawn look is the route that works.

Brief for AI Create: A one-minute birthday story: seven-year-old Lily, curly red hair and a yellow raincoat, flies her cardboard rocket to the moon to bring back the last slice of birthday cake, with her dachshund Pip as co-pilot; ends with the whole family singing to her.

## A story, not a slideshow

| Who | The adventure |
| A seven-year-old | A rocket to the moon to rescue the cake |
| A grandmother at eighty | Sailing back to the harbour town where she grew up |
| A friend at thirty | A heist to steal the last slice |
| A partner | The night you met, told as a film |

## Describe them in words

The details people recognise are the ones you write: curly red hair, gap in the front teeth, yellow raincoat, a dachshund called Pip. They go into the character description and the sheet draws them in the project's style. SceneMixer's free setup images (15 per account) cover a small cast. Describe a face · The dog as a character.

Visual Style: 3D Toy Cartoon

Visual Style: Watercolor Storybook

Visual Style: Claymation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Photos: what works

On the current plans, uploading a photo is a membership feature. Upload only with the person's consent (for a child, a parent's). Choose Optimize into a reference sheet so the photo is redrawn in the project's style; a realistic photo of a real person is refused by the video's face check, while drawn and 3D faces are rarely flagged. Using your own photo · Real faces.

## The wish at the end

Make the last shot a line someone says: @Mum (smiling): Happy birthday, Lily. Spoken lines come out right; written text in the frame does not, so put a title card on in your editor if you want one. The video carries a background score; a birthday song is something to add in your editor. Dialogue in the storyboard.

## Length, cost and sending it

A one-minute story is about four 15-second segments, rendered at 10 credits a second on the default tier; there is a free 5-second preview once per account. Composite the episode, then download it or send a share link, which plays without an account. Composite and share · Share links · What is free.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday; Visual Style Family 3D Animation, Watercolor Storybook; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the birthday person described in words (or a consented photo, optimised); family members and pets as characters; home and the adventure's place as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Second | Shot |
| 0–15 | Lily finds the cake plate empty; a crumb trail to the window |
| 15–30 | The cardboard rocket, Pip in a helmet, lift-off |
| 30–45 | The moon, the slice on a crater, a grumpy moon mouse |
| 45–60 | Home through the window; the family, candles, the wish |

## FAQ

### How do I make a birthday video with AI?

Write a short adventure with the birthday person as the hero, describe them in words, pick a cartoon look and render a minute or so.

### Can I use their photo?

On a paid plan, with consent, optimised into a drawn sheet; realistic photos of real people are refused by the face check.

### Can it say happy birthday with their name?

Yes, as a spoken line. Written text in the frame is unreliable; add it in your editor.

### Can it include the Happy Birthday song?

The video has a background score; add a song in your editor.

### How long should it be?

Thirty seconds to two minutes: two to eight segments.

### How do I send it?

Download the composite or send a share link that plays without an account.

### Which look suits a birthday video?

Family 3D Animation, 3D Toy Cartoon, Watercolor Storybook or Claymation.

### Can the pet be in it?

Yes. A dog or cat is a character with its own sheet.

Updated November 3, 2026


---

# AI BL Drama: How to Make a Boys' Love Romance Series, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-bl-romance-drama/
A BL romance series made with AI: two leads who never get mixed up, campus or city settings, tropes that fit short episodes, and where the content checks draw the line.

# How to Make an AI BL (Boys’ Love) Romance Drama

Quick answer: A BL drama lives on two leads and the tension between them, so the first job is making sure the audience — and the model — never confuses the two. In SceneMixer each lead is a character card with its own face, one unmistakable mark and its own voice; every shot is generated against those sheets, and in projects with the world model, two-person scenes keep each man on his own side of the frame. Pick a campus, office or Bangkok setting in step 1, cut the story into one-to-three-minute episodes that each end on a look or a line, and keep intimacy to glances, closeness and an almost-kiss: explicit content is refused.

Brief for AI Create: A BL campus romance: a cold, top-ranked student is forced to tutor the basketball captain who failed his exam; they fight in every session until a rainy night at the bus stop; 8 short episodes, each ends on a look.

## Two leads who never get mixed up

Two young men of similar age, height and build are the hardest pair for any video model to keep apart. Make them different on purpose:

- Hair: one short and black, one longer and dyed ash brown.
- Face: one long face with sharp, narrow eyes; one round face with wide eyes and a dimple.
- One mark each: a silver earring in the left ear; a small mole under the right eye.
- Clothes: one always in a pressed white shirt, one in a team hoodie. A second outfit is a state, not a second clause.
Make a face specific · Same face in every shot · One outfit per sheet.

## Settings in step 1

| Setting | Scene Style | Costume Style |
| Campus | Campus Youth, American High School | School Uniform, Streetwear |
| City office | Corporate Cold, Urban Opulence | Power Tailoring, Smart Casual |
| Night city | Neon Nights | Streetwear |
| Thai-style romance | Southeast Asian Town, Island Resort | Smart Casual, School Uniform |
Visual Style: Digital HD or Soft Glow for live action; Korean Webcomic or Anime for an illustrated BL. Target Region: East Asia or Southeast Asia to cast the faces and names of the market you are writing for. Target region.

Visual Style: Soft Glow

Visual Style: Korean Webcomic

Visual Style: Anime

Scene Style: Campus Youth

Scene Style: Neon Nights

Costume Style: School Uniform

Costume Style: Streetwear
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Two people in one frame

The scenes that matter are two-shots: across a desk, under one umbrella, in a narrow corridor. Name both with @ and place each one in the room, not in the frame: @Ren at the window, @Kai by the door. On projects with SceneMixer’s world model (new projects on paid plans), everyone has a place in one measured room and the cameras are planned in that room, so a shot-reverse-shot conversation keeps each man on his side. Why characters swap sides between cuts · Write positions the world can read.

| Beat | Shot to adapt |
| The standoff | [Shot 1] 4s: Medium two-shot, fixed camera. An empty classroom after school, late sun. @Ren sitting at the front desk, @Kai leaning on the door frame. @Kai (teasing): You actually waited. Ambient: cicadas, a distant whistle. |
| The reply | [Shot 2] 3s: Close-up on @Ren, not looking up from the book. @Ren (cold): Sit. You have forty minutes. Ambient: a page turning. |
| The look that ends it | [Shot 3] 3s: Close-up on @Kai, the grin fading as he watches Ren. No dialogue. Ambient: cicadas. |

## Tropes that fit short episodes

- Enemies to lovers: every episode one argument, every episode a little less cold.
- Forced proximity: roommates, a shared project, a stuck lift.
- Fake relationship: they pretend for a reason, and the pretending stops being pretend.
- Senior and junior: the club captain and the new member.
- Second chance: they meet again years later; the earlier version of each is a state of the same character.
End every episode on a look, a line or a touch interrupted. Hooks and cliffhangers.

## Where the content checks draw the line

Romance is fine; explicit sexual content is refused in the script, the shots and the pictures. Write intimacy as closeness and restraint — a hand on a wrist, a forehead almost touching, the kiss cut before it lands — which is also how the genre’s best scenes work. A refused generation is refunded. Content rules and checks.

## Make it in SceneMixer, start to finish

- Script: your own on Upload Script, or a brief in AI Create with the two leads, the trope and the number of episodes.
- Step 1: Type Short Drama; 9:16 Portrait for feeds; the scene, costume and visual styles above; the dialogue language and target region. Read the script.
- Step 2: make the two leads different at a glance; give each a voice description (low and clipped against bright and fast). The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project); place both leads in the room in every two-shot.
- Video: generate on the default tier at 10 credits a second; composite; post.
Make ReelShort-style dramas · Post to TikTok, Shorts and Reels.

| Episode | The scene | The ending beat |
| 1 | Ren is told to tutor Kai. | Kai: “Forty minutes? I’ll make it four.” |
| 2 | The first session goes badly. | Kai finds Ren’s notes left for him. |
| 3 | The basketball match. | Ren is in the stands. He says he is not. |
| 4 | Kai passes the exam. | No reason to meet any more. Ren waits anyway. |
| 5 | Rain at the bus stop, one umbrella. | Kai steps closer. The bus arrives. |

## FAQ

### How do I make an AI BL drama?

Write a short script or give SceneMixer's AI Create a brief with two leads and a trope, pick a campus or city setting in step 1, make the two leads different at a glance, and generate episode by episode.

### How do I stop the two male leads looking the same?

Give them different hair, face shapes, one mark each and their own outfits. SceneMixer uses each character's own sheet in every shot, so distinct sheets give distinct men.

### Can the characters kiss?

Romance and closeness are fine; SceneMixer refuses explicit sexual content. Most BL scenes work better as almost-kisses and glances anyway.

### Can I make a Thai-style or Korean-style BL?

Yes. Pick East Asia or Southeast Asia as the target region and a matching scene style, and choose the dialogue language.

### Can it be illustrated instead of live action?

Yes. Korean Webcomic or Anime in step 1 gives an illustrated BL; the style is fixed after step 1.

### Do two-person scenes stay consistent?

Place each character in the room, not in the frame. In projects with the world model, new projects on paid plans, the cameras are planned in one room, so each man keeps his side across cuts.

### How long should an episode be?

One to three minutes, one scene, ending on a look or a line.

### Can I post it on TikTok?

Yes. Pick 9:16 Portrait in step 1 and post each composited episode.

Updated November 3, 2026


---

# AI Bodyguard Romance: The Protector, the Heiress and the Threat

URL: https://scenemixer.com/guides/make-an-ai-bodyguard-romance/
Make a bodyguard romance with AI: Urban Opulence and Gulf Metropolis worlds, Power Tailoring suits, an earpiece as a prop, and danger shown before it lands.

# How to make an AI bodyguard romance

Quick answer: A bodyguard romance is two people who cannot be apart and must not be together. In SceneMixer, set the client's world in Step 1 (Urban Opulence, Gulf Metropolis or Island Resort), dress the bodyguard in Power Tailoring and give him an earpiece and a scar as marks. Danger is shown before it lands: a red dot, a car that stays too long, a hand pulling her down, rather than wounds.

Brief for AI Create: A bodyguard romance: a pop star with death threats is assigned the one bodyguard who quit her father's security team after a scandal; the threats are coming from inside her own entourage; 10 short episodes, each ends with him stepping between her and something.

## The client's world

| Client | Scene Style | Costume Style |
| An heiress in the city | Urban Opulence | Power Tailoring for him, Haute Couture for her |
| A royal or a sheikh's family | Gulf Metropolis | Power Tailoring, Middle Eastern Robes |
| A star on tour | Neon Nights or New York Metropolis | Streetwear, Power Tailoring |
| A holiday gone wrong | Island Resort | Beach Resort |
Billionaire romance · Desert romance.

## Marks, props and the threat

- The bodyguard's marks: an earpiece, a scar, the same dark suit. Same face.
- The threat is a prop: the letter, the photo, the phone. Props.
- Action is short shots of one move each; no injuries on screen. Fights.
- Crowds and paparazzi are background. Background people.

## The red carpet, shot by shot

| Beat | Shot to adapt |
| The carpet | [Shot 1] 4s: Wide shot, a red carpet at night, camera flashes from a crowd behind barriers. @Skye walks in a silver gown. |
| The watcher | [Shot 2] 3s: Close-up on @Cole two steps behind her, earpiece in, eyes on the crowd. |
| The flash | [Shot 3] 3s: Close-up on one camera in the crowd that is not flashing. |
| The move | [Shot 4] 4s: Medium shot. @Cole steps in front of @Skye. @Cole (low): Walk. Don't look. (voice: Cole) |

## Story shapes that work

- The star and the guard.
- The heiress who hates being protected.
- The fake relationship for cover.
- The female bodyguard and the CEO.
- The guard with a secret link to the threat.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Neon Nights; Costume Style Power Tailoring, Haute Couture; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the client and the bodyguard with one mark each; the threat letter as a prop; the carpet, the suite and the bus as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The record label office. | Her new guard is the man who quit on her father. |
| 2 | The red carpet. | A camera that is not flashing. |
| 3 | The hotel suite. | A letter under the door, inside the security perimeter. |
| 4 | The tour bus. | Someone in the band has the same handwriting. |
| 5 | The rooftop. | He kisses her, then steps back. |

## FAQ

### How do I make a bodyguard romance with AI?

Choose the client's world in SceneMixer's Step 1, Power Tailoring for the bodyguard, give him one clear mark and make the threat a prop.

### Can I show action scenes?

Yes, as short shots of one move each; injuries are kept off-screen.

### Can the bodyguard be a woman?

Yes. Describe her on her sheet and keep Power Tailoring for the role.

### How do I show paparazzi?

Put the crowd and flashes in the background line.

### Which visual style suits it?

Digital HD or Soft Glow.

### Can I use a real celebrity?

No. Real people cannot be depicted.

### Vertical or landscape?

9:16 Portrait for short-drama feeds.

### How long are the episodes?

One to two minutes, each ending on one protective moment.

Updated November 3, 2026


---

# AI Bollywood-Style Family Drama: Weddings, Palaces and Big Feelings

URL: https://scenemixer.com/guides/make-an-ai-bollywood-style-family-drama/
Make a Bollywood-style family drama with AI: Indian Bazaar Streets and Indian Royal Palace scenes, South Asian Sari and Indian Royal costumes, wedding scenes as states.

# How to make an AI Bollywood-style family drama

Quick answer: A Bollywood-style family drama is a wedding, two families and a love that does not fit the plan. In SceneMixer, Step 1 gives you India: Indian Bazaar Streets (colonial facades under bright shopfronts, cloth awnings, garlands) for a city family, Indian Royal Palace (red sandstone, white marble, mirror mosaics) for royalty, with South Asian Sari or Indian Royal costumes and Target Region South Asia. Hindi is not one of the dialogue languages, so lines are spoken in English or another available language.

Brief for AI Create: A family drama: two weeks before her arranged wedding, a Jaipur jeweller's daughter learns the groom's family bought her father's debt; she agrees to the marriage and plans to bring the family down from inside; 12 short episodes, each ends at a wedding ritual.

## India in Step 1

| Kind of family | Scene Style | Costume Style |
| A big city family | Indian Bazaar Streets | South Asian Sari |
| Royals and old palaces | Indian Royal Palace | Indian Royal |
| New money in a glass tower | Urban Opulence | Minimal Luxury |
| An Indian family abroad | New York Metropolis or English Manor | Smart Casual |
Target Region South Asia for faces and names. Target Region.

## Dialogue language, honestly

Dialogue on SceneMixer can be any of 15 languages; Hindi is not one of them. Make the drama in English, add subtitles if you need them, and keep names, places and food in the script exactly as you want them said. Dialogue language · Subtitles.

## The wedding, shot by shot

Wedding outfits are states: the everyday sari and the wedding lehenga are two looks of one character. States.

| Beat | Shot to adapt |
| The courtyard | [Shot 1] 4s: Wide shot, a palace courtyard strung with marigold garlands, guests in bright silk in the background. |
| The bride | [Shot 2] 3s: Medium shot. @Meera sits for the henna ceremony in red and gold. |
| The mother-in-law | [Shot 3] 3s: Close-up on @Kamala adjusting a heavy gold necklace on Meera's neck. |
| The line | [Shot 4] 4s: Close-up on @Kamala. @Kamala (sweetly): This was meant for my son's real choice. (voice: Kamala) |
The music is scored for every episode automatically; there is no song picker. Background music.

## Story shapes that work

- The arranged marriage.
- Two families, one feud.
- The mother-in-law who rules the house.
- The royal heir and the commoner.
- Coming home from abroad for a wedding.
Family inheritance dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Indian Bazaar Streets, Indian Royal Palace; Costume Style South Asian Sari, Indian Royal; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: both families with one mark each and wedding states; Target Region South Asia; the courtyard, the shop and the family office as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The jeweller's shop. | The debt papers with the groom's name. |
| 2 | The engagement. | She says yes and smiles at the wrong person. |
| 3 | The henna ceremony. | The necklace that was meant for someone else. |
| 4 | The family office. | The groom's brother sees her copying a ledger. |
| 5 | The wedding night. | The groom confesses he never wanted this either. |

## FAQ

### How do I make a Bollywood-style drama with AI?

Choose Indian Bazaar Streets or Indian Royal Palace in SceneMixer's Step 1, South Asian Sari or Indian Royal costumes, Target Region South Asia, and give wedding outfits as states.

### Can the characters speak Hindi?

No. Hindi is not one of SceneMixer's dialogue languages. Use English or another available language, with subtitles if needed.

### How do I show wedding outfits?

Add them as a state of each character in SceneMixer’s Step 2.

### Can the characters dance?

Short movements work on SceneMixer; long choreographed numbers do not. Cut between faces and the crowd.

### Can I use a famous film's song?

No. Music on SceneMixer is scored automatically and real songs are someone else's IP.

### Which visual style suits it?

Digital HD or Soft Glow; Shot on Film for a period story.

### Can I use real film stars?

No. Real people cannot be depicted.

### Vertical or landscape?

9:16 Portrait for short-drama feeds, 16:9 Landscape for a series.

Updated November 3, 2026


---

# AI Book Trailer: Two Ways to Make One for Your Novel

URL: https://scenemixer.com/guides/make-an-ai-book-trailer/
Turn your novel into a trailer: a cinematic teaser with your characters and a narrator, or a product-style trailer with your cover and an end card with the title.

# How to Make an AI Book Trailer for Your Novel

Quick answer: A book trailer has one job: make a reader want to know what happens. SceneMixer gives an author two ways to make one. The story trailer is a short drama project: you write a sixty-to-ninety-second trailer script — a narrator, three or four key moments, a final line — and your characters are cast, kept consistent and filmed like a film. The cover trailer is an ad project: you upload the cover as the product, pick a length and tone, and it ends on an end card with your name, the title, a tagline and a call to action such as Pre-order now. Either way, the title reaches the viewer by voice or on the end card, not as text in the shots, because generated video does not write legible text.

Story trailer, first line of your trailer script: NARRATOR (voice-over): In the town of Halloway, nobody has opened the lighthouse door in forty years. Tonight, someone knocked from inside.

## Which kind of trailer

| | Story trailer | Cover trailer |
| Starts from | A short trailer script you write | One to four photos: the cover first |
| Where | Upload Script (or AI Create) | The Ad Video tab |
| Shows | Your characters and settings, in scenes | Your book as the hero object, in styled scenes |
| Title and author | Spoken by the narrator; add text later in your editor | Printed on the end card |
| Length | Whatever your script runs to | 15, 30 or 60 seconds |
| Best for | Fiction with a strong premise and a face on it | A release announcement, a pre-order push, non-fiction |

## The story trailer, step by step

- Write the trailer script, not the book. Pasting a whole novel gives you the whole novel, episode by episode. A trailer is its own short script: a narrator, three or four moments that raise questions, and a last line. One page is plenty. Use a screenplay layout.
- In step 1, choose Film for a cinematic pace and 16:9 Landscape for YouTube and your website, or 9:16 Portrait for book communities on TikTok and Instagram. Pick the visual style that matches your cover: Shot on Film for a thriller, Soft Glow for a romance, Watercolor Anime or Watercolor Storybook for fantasy or a children’s book.
- Cast your characters as you see them. In step 2, rewrite each character’s description to match the book — the scar, the coat, the colour of the eyes — and regenerate. You know them better than any reader. Describe a face so it is not generic.
- Make the narrator voice-only. A narrator who is heard and never seen is set to Voice only (off-screen); describe the voice you hear when you read the book aloud.
- Generate, fix, composite. Regenerate the shots that miss, then composite the trailer into one file and download the original.

Visual Style: Soft Glow

Visual Style: Watercolor Anime

Visual Style: Watercolor Storybook
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## A trailer script you can adapt

| Beat | What to write |
| The world | A wide shot of the one place the book is about, at the time of day it feels like. Narrator: one sentence of premise. |
| The lead | A close-up of the main character doing something that shows who they are. No dialogue, or one line. |
| The disturbance | The moment the story starts: a letter, a knock, a stranger. Narrator: the question. |
| The stakes | Two or three quick shots of what could be lost. Sound carries them. |
| The last line | A close-up and one line of dialogue that the reader will meet again in the book. |
Keep spoken lines short: about three words a second in English. A sixty-second trailer holds far fewer words than you think. How to write one shot · Lines and voice-over.

## The cover trailer, step by step

- Open the Ad Video tab and add your cover first. Up to four images; the first one is the hero. A photo of the printed book — on a table, in someone’s hands — often works better than the flat cover file. Which photos work.
- Say what it is in the one-line box: a hardback psychological thriller, launching in March. Where the sentence and the photo disagree, the sentence wins.
- Pick length, tone, shape and resolution. Tone sets the treatment — Twist or Epic for a thriller, Warm for a romance, Minimal for literary fiction. Ad options.
- Hand it to the AI, or take the wheel. One click runs the whole chain at one quoted price; expert mode stops after the script so you can rewrite lines first. One-click or expert mode.
- Set the end card. Brand = your author name, Product name = the title, a tagline, and a call to action such as Pre-order now. You can edit the card and rebuild the cut without reshooting. The end card.

## Text, rights and covers with faces

- No readable text in the shots. Generated video does not write legible words, so a title painted across a scene will come out garbled. Let the narrator say the title, use the end card in a cover trailer, or add the title and release date in your own video editor after downloading.
- Only your own book. Make trailers for work you hold the rights to. Someone else’s novel, a famous series or its characters are not yours to adapt, and well-known franchises trigger the IP check.
- Covers with a real person’s photo. Real people are not put into generated video: in an ad, people in a photo are removed before anything is made, and a reference that is a real photograph of a face is rejected by the face check. A cover built around a photograph of a person may lose the person or not be usable at all. Illustrated covers and object covers work best. Which photos work for an ad.
- The AI label. Retailers and festivals may have their own rules for disclosing AI-made trailers; check them before you publish. The AI badge switch.

## What it costs

A story trailer is priced like any short project: the script read, the setup images (the first 15 on an account are free), the storyboard, and video by the second — 10 credits a second on SceneMixer's default tier, so a sixty-second trailer is sixty seconds of video. A cover trailer shows one all-in price on each length and resolution option before you start, and that price is a cap. The ad price is a cap.

| Genre | Trailer kind | Visual style or tone |
| Thriller, crime | Story trailer | Shot on Film |
| Romance | Story trailer | Soft Glow |
| Fantasy, YA | Story trailer | Watercolor Anime or Painterly Anime |
| Children’s picture book | Story trailer | Watercolor Storybook |
| Non-fiction, business | Cover trailer | Minimal or Show how it works |
| Pre-order or launch push | Cover trailer | Epic or Twist |

## FAQ

### Can AI make a trailer for my book?

Yes. Write a short trailer script and make it as a short film project, or upload your cover on SceneMixer's Ad Video tab and make a product-style trailer that ends on an end card with your title.

### Can I paste my whole novel?

You can, but that makes the novel into a series, not a trailer. For a trailer, write a one-page trailer script with a narrator and three or four key moments.

### Can the trailer show the book title on screen?

Not inside the generated shots; video models do not write legible text. The narrator can say it, a cover trailer prints it on the end card, or you can add it in your own editor afterwards.

### How long should a book trailer be?

Thirty to ninety seconds. SceneMixer's cover trailers come in 15, 30 or 60 seconds; a story trailer runs as long as its script.

### Will my characters look like I imagined them?

Rewrite their descriptions in step 2 to match the book and regenerate their images until they do. The same faces are then kept in every shot.

### Can I make a trailer for someone else's book?

Only with the rights. Trailers for books you do not own, or for famous series and their characters, are not allowed.

### Vertical or horizontal?

Horizontal (16:9) for YouTube, your website and retailer pages; vertical (9:16) for TikTok and Instagram book communities.

### Can I use music from my favourite film?

No, only music you have the rights to. Projects can carry their own generated score, and an ad gets music written for it.

Updated November 3, 2026


---

# AI Boxing or Fight Movie: One Punch Per Shot, Then the Corner

URL: https://scenemixer.com/guides/make-an-ai-boxing-or-fight-movie/
Make a boxing or fight movie with AI: one punch or one reaction per shot, the corner between rounds, training as states, and how to keep fight shots from being blocked.

# How to make an AI boxing or fight movie

Quick answer: A boxing film is three places (the gym, the corner, the ring) and one question: does the fighter get up. In SceneMixer, write each round as short shots with one punch or one reaction each, give the fighter states for training, fight night and after, play the corner between rounds as the real drama, and keep injuries light: sweat, tape, a swollen eye. The sample Rust City Knockout is a fight story in Painterly Anime.

Brief for AI Create: A boxing film: a forty-year-old cleaner at a run-down gym talks the owner into one amateur fight, against the owner's son; training, the night, three rounds.

## One action per shot

| Shot | What happens |
| [Shot 1] 2s | Medium shot. @Rosa jabs; @Kai slips it. |
| [Shot 2] 2s | Close-up. @Kai's right hook lands on @Rosa's guard. |
| [Shot 3] 2s | Low angle. @Rosa staggers back to the ropes. |
| [Shot 4] 3s | Close-up on @Rosa's eyes. She steps forward again. |

A shot with a whole exchange in it turns to mush; a run of two-second shots reads like a fight. Change the framing on every cut. Fights, crowds and group scenes · Action movie.

## The corner is the drama

Sixty seconds on a stool: water, ice, one sentence from the trainer. That is where the film is. Write it as close-ups with dialogue and no punches. @Coach (low): He drops his left after the third jab. Wait for it.

## Training as states

| State | What the sheet shows |
| Before | Cleaner's overalls, soft around the middle, hair loose |
| Training | Grey sweatshirt, hand wraps, hair tied back |
| Fight night | Red shorts and vest, gloves, a gumshield |
| After | A swollen eye, a towel round the neck |

States keep the face the same through the change. States. The training montage is a run of short shots across these states: skipping, the bag, the stairs at dawn.

## Look and sample

Character sheets from the sample Rust City Knockout: one reference image per character, used in every shot they appear in. Rust City Knockout is drawn in Painterly Anime; the same approach works in Digital HD or Shot on Film.

Visual Style: Shot on Film

Visual Style: Painterly Anime

Scene Style: Street Life

Scene Style: Neon Nights

Costume Style: Sportswear
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Keep it in the ring

Blood and serious wounds are blocked as graphic violence, and even ordinary fight shots are sometimes blocked and pass after a rewording. A swollen eye, sweat and tape are enough; the knockdown can be the fall and the referee's count rather than the impact. A blocked attempt is normally not billed. Content rules · Blocked by moderation. For MMA, street fights and anime battles see anime fight scenes and kung fu.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Street Life, Neon Nights; Costume Style Sportswear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the fighter with before, training, fight and after states; the coach and the opponent; the gym and the ring as locations; gloves and the stool as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Act | Place | Purpose |
| The nobody | The gym, mopping | Why she wants it |
| Training | Gym, stairs, the bag | The change, in states |
| Round 1 | The ring | She is outclassed |
| The corner | The stool | One piece of advice |
| Round 3 | The ring | She gets up |

## FAQ

### How do I make a boxing movie with AI?

Write each round as two-second shots with one punch or reaction, give the fighter states, and play the corner as the drama.

### Why do my fight shots look messy?

Too much happens in one shot; split it into one action per shot and change the framing on each cut.

### Can there be blood?

Keep injuries light; SceneMixer blocks blood and serious wounds.

### How do I show training?

A run of short shots across a before, training and fight-night state.

### Which look suits a fight film?

Digital HD or Shot on Film; Painterly Anime for an anime fight.

### Can I use a real boxer or a famous fight film?

No; real people and famous films trigger the IP check. Invent the fighters.

### Can it be MMA or a street fight instead?

Yes; the same one-action-per-shot rule applies.

Updated November 3, 2026


---

# How to Make an AI Brand Film: Story First, Name Last

URL: https://scenemixer.com/guides/make-an-ai-brand-film/
Make an AI brand film: a 60 to 180 second story about why the business exists, five beats, values shown as scenes, and the name delivered as readable text at the end.

# How to make an AI brand film

Quick answer: A brand film is a short story about why a business exists, not a spot about a product. In SceneMixer, write it as a script of five beats — the maker, the problem, the turn, the work, the line — drawn with invented characters, and keep the name for the end. Video models do not draw readable text, so a logo or a name inside a shot is unreliable: put the name on a title card in your editor, or use the ad route, whose end card delivers brand, product, tagline and call to action as readable text.

Brief for AI Create: A 90-second brand film for a family bakery: the founder's daughter finds her mother's first recipe card, we see why a loaf is shaped by hand at five in the morning, a customer returns after ten years; warm, unhurried, the shop's name only at the end.

## A brand film is not an ad

| Point | Brand film | Ad |
| Starts from | A story and a script | Up to four photos of one subject |
| Length | 60 to 180 seconds | A spot, with an end card |
| The name | On a title card you add | On the end card: brand, product name, tagline, call to action |
| Where it lives | A website, a festival, a launch, a pitch | A feed or a shop page |
Make an AI ad from one photo · AI ad video generator · The ad end card.

## Five beats for 90 seconds

| Beat | Seconds | What it shows |
| The maker | 15 | A person at work before anyone is watching |
| The problem | 20 | What was wrong with how it used to be done |
| The turn | 20 | The decision that made the business |
| The work | 25 | The craft, in close-ups of hands and materials |
| The line | 10 | One spoken sentence, then silence, then the name |
That is 90 seconds, six 15-second segments or three of 30. How to make a short movie with AI.

## Turn values into scenes

Never say the value; show the scene that proves it.

| Value | A scene that shows it |
| Care | Someone re-ices a cake at 5 a.m. that no customer would have noticed |
| Patience | Dough left to rise while the founder sits and waits, the clock moving |
| Belonging | A regular's order is ready before they reach the counter |
| Honesty | A tray of imperfect loaves goes on the shelf with a handwritten tag |

## Where the name goes

- On a title card in your editor. Finish the film, download it, and add the logo and the name as your own graphics. That is the reliable way for a story film.
- Through the ad route. An ad's end card is the one place the name is delivered as readable text: brand, product name, tagline, call to action, each editable, with the cut rebuilt without reshooting. Make the ad separately from product photos and join the two in your editor.
- Spoken once. You can have a narrator say the name at the end; listen to the take, because a name that is not a common word may be said differently from how you say it.

## People, places, product

Use invented characters, not your staff: every real person whose face appears in what you upload has to have consented, and the video provider's face check refuses images that look like a real person (the credits come back, but the segment fails). Describe your place rather than uploading a logo'd photo of it, and show your product through the ad route's photos. Reference image flagged as a real person's face · Using your own photo.

## A hero film and a social cut

The aspect ratio is fixed once you continue past step 1, so the 16:9 film for your website and a 9:16 cut for social are two projects from the same script; the cast can come across from the asset library. Vertical or landscape · Reuse a character, scene or voice.

## Before it goes out

Free accounts' videos carry the small AI badge, and shared links always do. Where you publish, check the platform's AI rules. The AI badge switch · What YouTube's pages say about AI videos.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday, Smart Casual; Visual Style Soft Glow, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the founder, her daughter and the one customer, and the bakery as a single place; no logo, no sign. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Route | Input | How the name appears | Suits |
| Story film | A script | A title card you add in your editor | A website, a launch, a pitch |
| Ad | Up to four photos of one subject | The end card: brand, product, tagline, call to action | A feed, a shop page |
| Both | A script and photos | Join them in your editor | A hero film plus a closing spot |

## FAQ

### Can AI make a brand film?

Yes, as a short story film with invented characters, and with the name added on a title card or through the ad route's end card.

### Can it show our logo?

Not reliably: video models do not draw readable text. Add the logo in your editor, or use the ad route's end card for the name.

### How long should a brand film be?

60 to 180 seconds is a common range; 90 seconds is five beats of 10 to 25 seconds.

### Can I use our staff or customers as characters?

Use invented characters. Real people's faces need their consent, and the video provider's face check can refuse images that look like a real person.

### Can I make a vertical cut for social?

Yes, as a second project from the same script, since the aspect ratio is fixed after step 1.

### What is the difference from an AI ad?

A SceneMixer ad starts from photos of one subject and ends on a card with brand, product, tagline and call to action; a brand film starts from a story.

### Does it carry an AI label?

Free accounts' videos carry a small badge; on a paid membership it is off by default, and shared links always carry it.

### What does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join, so 90 seconds is 90 seconds of each, before retakes.

Updated November 3, 2026


---

# AI Cartoon Horror: Make a Scary Cartoon Series (No Gore)

URL: https://scenemixer.com/guides/make-an-ai-cartoon-horror-series/
Make a scary cartoon series with AI: pick a look that holds darkness, build fear from a rule, a sound and a cut, and plan six episodes. Spooky, not gory.

# How to make an AI cartoon horror series

Quick answer: Cartoon horror works because the picture is friendly and the event is not. In SceneMixer pick a stylised look in Step 1 — American Comic for heavy black shadows, Claymation for handmade unease, Ink Wash Animation for quiet dread, 3D Toy Cartoon for toys that should not move — then build every episode on one rule, one sound before the sight and one shock at the end. Gore is blocked by the content rules, which suits the genre: the scare comes from timing, not blood. It is a series for older kids, teens and adults, not for preschoolers.

Brief for AI Create: A six-episode cartoon horror series: a boy inherits his grandmother's button shop and finds that every jar of buttons is a night someone lost; each episode he opens one jar and that night plays out in the shop; spooky and funny, one shock per episode, the last episode answers why the shop never closes.

## Why a cartoon scares differently

- The gap. A round, bright world makes one wrong thing louder. A smile one notch too wide does more than a monster.
- Exaggeration. Long shadows, tall doorways, tiny people: the drawing can lie, so use it.
- Timing. A held shot, then a cut. In a cartoon the pause is the jump scare.
- No explanation. Leave the creature half-seen. The question it leaves is the hook for the next episode.

## Pick the look that holds the dark

| Visual Style | What the card says | The horror it suits | Sets that pair with it |
| American Comic | Bold ink lines and heavy black shadows | Monster-of-the-week, night streets, pulp scares | Gothic Dark, American Small Town |
| Claymation | Hand-pressed clay figures and miniature sets | An uncanny toy shop, small rooms with a rule | English Manor, Suburban Mansion |
| Ink Wash Animation | Shapes painted in ink washes on rice paper | Ghost stories, empty halls, quiet dread | Japanese Machiya, Chinese Folk Occult |
| 3D Toy Cartoon | Rounded, matte toy-like figures in bright colors | Toys that should not move, a bright house at night | Suburban Mansion, American Small Town |
| Anime | Clean lines and flat colors, like TV anime | School ghost stories, a ghost in the club room | Japanese High School, Campus Youth |

Visual Style: Claymation

Visual Style: Ink Wash Animation

Visual Style: 3D Toy Cartoon
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
In SceneMixer, the look is chosen in Step 1 and fixed once you continue. All visual styles · Scene and costume styles.

## The escalation ladder

Every episode climbs the same five steps. Write one shot or two for each:

| Step | What changes | In the button shop |
| 1. Normal | Establish the room and its one rule | Jars of buttons to the ceiling; Grandmother's rule: never open a jar after dark |
| 2. One thing off | A single small wrongness | The shop clock ticks one beat slow |
| 3. Repeat | It happens again, nearer | A jar slides along the shelf an inch |
| 4. The rule breaks | The character does the forbidden thing | Milo opens the jar |
| 5. The shock | One image, then a cut to black | The night inside the jar is playing out in the shop |

## A cold open, shot by shot

| Beat | Shot |
| The shop | [Shot 1] 3s: Wide shot. A tiny button shop at night, one lamp lit in the window, glass jars stacked to the ceiling. Ambient: a clock ticking. |
| The jar | [Shot 2] 3s: Close-up. @Milo's hand lifts a jar lid; one button rolls out and spins on the counter. Ambient: a click, then silence. |
| The almost-reveal | [Shot 3] 4s: Medium shot, low angle. @Milo with his back to the shelves; in the dark window glass behind him, a jar rises by itself. No dialogue. Ambient: the ticking stops. |
| The line | [Shot 4] 3s: Close-up on @Milo, frozen. @Milo (whisper): Grandma? (voice: Milo) |
| The cut | [Shot 5] 2s: Black. A single button drops on the floor. Ambient: one soft tap. |
That is 15 seconds. The sound arrives before the picture, the thing behind him is seen only in the glass, and the episode cuts away before it is explained. For the full method of lighting and sound in the dark, see the horror short guide.

## Keep it a cartoon, keep it dark enough

- Light every dark shot. Name a lamp, a window or a glow; a shot that says only “dark” comes out muddy.
- Silhouettes read well. In these looks a shape in a doorway is clear even in shadow.
- Make the creature a state. The boy and the boy with button eyes are one character with two states, so the face matches. Costume changes and states.
- Toys and animals are characters. Say what they are and whether they have hands. Non-human characters.

## Who it is for, and how to say so

A scary cartoon draws children even when it is not made for them. Say who it is for in the title and description, and follow each platform's rules for children's content and AI disclosure. Do AI videos need a label? · Content rules · AI horror movie generator · A haunted house drama.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Gothic Dark, American Small Town; Costume Style Soft Everyday; Visual Style American Comic, Claymation; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw Milo, the button shop and the jar as a prop; each episode's creature is added as a state when its episode comes. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Ep. | The jar | The sound first | Ends on |
| 1 | The jar that is always warm | The clock stops | The lid lifts on its own |
| 2 | A rainy night | Rain, indoors | The floor is wet to the front door |
| 3 | A child's laugh | Half a bar of a music box | Milo's laugh answers from the jar |
| 4 | The jar with no label | Footsteps in the flat upstairs | The ceiling light swings, once |
| 5 | Buttons from his own coat | A single button dropping | His coat has one button fewer |
| 6 | The shop's last jar | The bell over the door | Grandmother's handwriting on the label gives today's date |

## FAQ

### Can AI make a scary cartoon?

Yes. In SceneMixer, choose a stylised look such as American Comic, Claymation, Ink Wash Animation or 3D Toy Cartoon, and build each episode on a rule, a sound and a cut.

### Is gore allowed?

No. SceneMixer's content rules block detailed gore, so the fear comes from timing, sound and what is not shown.

### Is cartoon horror suitable for children?

Not by default. It is written for older kids, teens and adults; say who it is for in the title and follow each platform's rules for children's content.

### Which look is scariest?

American Comic for heavy black shadows; Claymation for handmade unease; Ink Wash Animation for quiet dread. Test the same shot in two and keep the one that holds the dark.

### Can the monster change shape between episodes?

Yes. Make each form a state of the character, so the face still matches.

### Can I use a famous cartoon's style or characters?

No. The IP check on SceneMixer refuses known characters; make your own.

### Can I make it vertical?

Yes. Choose 9:16 Portrait in Step 1.

### How long are the episodes?

A Short Drama episode on SceneMixer runs up to 240 seconds; scare episodes usually work at 60 to 120.

Updated November 3, 2026


---

# AI Cat and Pet Drama: How to Make Animal Story Videos, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-cat-or-pet-drama/
Real cats with an inner voice, or cartoon cats who talk: how to cast animal characters that stay the same in every shot, write their lines and turn them into a series.

# How to Make an AI Cat or Pet Drama

Quick answer: AI cat and pet dramas come in two kinds. In the first, the animals look real — a ginger tabby on four legs, no clothes — and the drama comes from a voice-over: what the cat is thinking while the humans argue. In the second, the animals are cartoon characters who stand, dress and talk, and the story is a soap opera with whiskers. SceneMixer does both: each animal is a character with its own card, image and voice; the description decides whether it is drawn as an animal or as a character, and the visual style you pick in step 1 (Digital HD for real animals, Family 3D Animation or 3D Toy Cartoon for talking ones) sets the world.

Brief for AI Create: A cat drama told by the cat: a ginger tabby watches his owner fall for the new neighbour, whose dog he despises; he sabotages every date, until the dog saves his life; deadpan, warm, the cat narrates.

## Choose the kind before anything else

| | Real animals, inner voice | Cartoon animals who talk |
| Visual Style | Digital HD or Shot on Film | Family 3D Animation, 3D Toy Cartoon, Anime |
| How the animal is drawn | As itself: four legs, no hands, no clothes | As a character: can stand, hold things, wear clothes |
| How it speaks | Voice-over: its thoughts are heard while its mouth stays shut | On screen, mouth moving to the words |
| Best for | Deadpan comedy, heart-warming stories, humans in the cast | Soap operas, family stories, all-animal casts |
Pick one per project. The visual style is fixed once you continue past step 1, and mixing a photoreal cat with a cartoon dog in the same world never quite works. Visual style presets.

Visual Style: Family 3D Animation

Visual Style: 3D Toy Cartoon

Scene Style: American Small Town

Scene Style: Pastoral Healing
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Write the animal’s card

The card description is what the sheet is drawn from. Say what the animal is and how much of a person it is.

- A real cat: A large ginger tabby tom, green eyes, white chest and white front paws, a notch in his left ear, thick fur, heavy and slow-moving. Leave the clothing empty: on a person an empty field is a problem; on a cat it is correct.
- A talking cartoon cat: A small grey cat who walks upright, oversized yellow eyes, a pink nose, wears a red knitted scarf and a tiny satchel; expressive eyebrows.
- A dog beside it: breed, colour, size against the cat or a person, one mark: a golden retriever, a white patch on his chest, waist-high to an adult, always carrying a tennis ball.
The one mark — the notched ear, the white paws — is what keeps a tabby recognisable across fifty shots, because tabbies otherwise look alike. Non-human characters · Make a face specific.

Character sheets from the sample Rocket Pup: one reference image per character, used in every shot they appear in. An animal lead is a card like any other, with its own sheet and voice.

## Lines: inner voice or on-screen speech

For real animals, the cat’s thoughts are a voice-over line: the shot shows the cat, its mouth stays shut, and the voice carries the joke. Start the delivery note with (inner voice) — it is read the same way as (voice-over): the words are heard over the shot and nobody mouths them — and write what the animal does in plain words.

| Beat | Shot to adapt |
| The stare | [Shot 2] 4s: Medium shot, fixed camera. In frame only: @Biscuit, sitting on the windowsill, watching the street. @Biscuit (inner voice, deadpan): The dog is back. The dog has brought flowers. Ambient: birds, a car door. |
| The sabotage | [Shot 3] 3s: Close-up. @Biscuit pushes the glass of wine off the table with one paw, slowly, looking at the camera. No dialogue. Ambient: the glass breaking. |
| The human | [Shot 4] 3s: Medium shot. @Anna turns from the stove. @Anna (exasperated): Biscuit! Ambient: a dog barking next door. |
For cartoon animals, write lines exactly as for people, with a voice tag, and give the animal a mouth in its description. Lines, voice-over and captions · How to write one shot.

## People and animals in the same shot

- Name both with @ and say who is where: @Biscuit on the sofa arm, @Anna sitting below him.
- Keep the size in the animal’s description — knee-high, fits on a lap, waist-high — so it is scaled right next to people.
- One action per shot. An animal jumping, landing and knocking something over is three shots, not one.
- Animals who never appear — a dog barking next door, a cat heard through a wall — are sounds in the Ambient line, not characters.

## Voices that sell the joke

The voice does most of the comedy. Cast against the animal: a tiny kitten with a weary, deep, unhurried voice; a huge dog with a high, breathless, delighted one. Write it in each animal’s Voice description and it applies to every line they speak. How character voices work.

## Settings and series

- Type and Aspect Ratio: Short Drama and 9:16 Portrait for TikTok, Reels and Shorts; one situation per episode, a punchline or a hook at the end.
- Scene Style: American Small Town or Soft Romance for a home-and-neighbourhood story, Pastoral Healing for a farm or countryside one.
- Same home, every episode. The flat, the windowsill and the garden are locations with their own images. Same room in every episode.
- Keep going. Post the first episodes and add new ones to the same project, so the cast and the rooms carry over. Make the next episode.

## Make it in SceneMixer, start to finish

- Script: your own on Upload Script, or a brief in AI Create naming the animal, what it wants, who is in its way and how it is told.
- Step 1: the style for the kind you chose; read the script.
- Step 2: check each animal’s description says what it is, its size and its one mark; regenerate any sheet that looks generic. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project); make the animal’s thoughts voice-over lines if it is a real animal.
- Video: generate on the default tier at 10 credits a second; composite; post.
For a cartoon animal lead, watch Rocket Pup, a 3D animated family adventure with a dog at its centre.

| Episode | The situation | The punchline or hook |
| 1 | A new neighbour moves in. With a dog. | The cat’s inner voice: “This means war.” |
| 2 | The first date dinner. | The cat knocks the wine over at exactly the wrong moment. |
| 3 | The dog brings the cat a gift: a tennis ball. | The cat hides it. Then sleeps on it. |
| 4 | The cat is stuck on the roof in the rain. | The dog will not stop barking until someone comes. |
| 5 | The two humans and the two animals on one sofa. | “Fine. He can stay.” |

## FAQ

### How do I make AI cat story videos?

Write a short script with the cat as a character, or give SceneMixer's AI Create a brief. Choose Digital HD for a real cat with a voice-over, or a 3D animation style for a talking cartoon cat, then generate episode by episode.

### Will my cat look the same in every episode?

Yes, if its card describes it precisely, with one mark such as a notched ear or white paws. The same image is used in every shot where the cat is named with @.

### Can a realistic cat talk?

It works better as a voice-over: the cat's thoughts are heard while its mouth stays shut. If you want animals that talk on screen, use an animation style and give them mouths.

### Will my cat be drawn as a person in a cat costume?

Not if the description says it is a cat and gives it no clothes. The sheet follows the description.

### Can people and animals share a shot?

Yes. Name both with @, say where each one is, and keep the animal's size in its description.

### How long should a pet drama episode be?

Short: one situation and one punchline or hook, about a minute, for feeds. Pick Short Drama and 9:16 in step 1.

### Can I use my own pet's photo?

On a SceneMixer paid plan you can upload your own photo as a character's reference image. Otherwise describe your pet on the card: breed, colouring, size and one mark.

### Do animal characters cost more?

No. On SceneMixer, images and video are priced the same for every character, animal or human.

Updated November 3, 2026


---

# AI Chef and Cooking Drama: Kitchens, Rivals and the One Dish

URL: https://scenemixer.com/guides/make-an-ai-chef-or-cooking-drama/
Make a cooking or restaurant drama with AI: the kitchen as a recurring location, chef's whites as a state, dishes as props, and service rush told in short shots.

# How to make an AI chef or cooking drama

Quick answer: A cooking drama is a kitchen, a dish that means something and a service that cannot go wrong. In SceneMixer, make the kitchen and the dining room locations you return to every episode, dress the brigade in Professional Uniforms (one cut per role, so the head chef reads at a glance), and make the signature dish a prop with its own picture. Food looks best in close-ups with the light written in: “steam rising under a warm lamp”.

Brief for AI Create: A cooking drama: a disgraced chef takes over her grandmother's failing trattoria and enters a televised competition against the restaurant that stole her grandmother's recipes; 10 short episodes, each ends with a dish tasted in silence.

## Kitchens in Step 1

| Kind of kitchen | Scene Style | Costume Style |
| A family trattoria | Mediterranean Coast | Professional Uniforms, Soft Everyday |
| A fine-dining restaurant | Urban Opulence or Nordic Minimal | Professional Uniforms |
| A street food stall | Street Life or Southeast Asian Town | Soft Everyday |
| A competition show | Corporate Cold | Professional Uniforms |
| A palace kitchen | Imperial Splendor | Song Restraint |

## Dishes, whites and the rush

- The signature dish is a prop, the same plate in episode 1 and episode 10. Props.
- Chef's whites and street clothes are two states. States.
- Write the light and steam. Food reads through texture and light. Describe the cause.
- Service rush is short shots: a ticket, a pan, a face, a plate. One shot.

## The tasting, shot by shot

| Beat | Shot to adapt |
| The plate | [Shot 1] 3s: Close-up, a plate of hand-cut pasta under a warm lamp, steam rising. |
| The judge | [Shot 2] 3s: Medium shot. @Chef Bruni lifts a fork. |
| The wait | [Shot 3] 3s: Close-up on @Giulia behind the counter, hands flat on the steel. |
| The line | [Shot 4] 4s: Close-up on @Chef Bruni. @Chef Bruni (quietly): Where did you learn this? (voice: Bruni) |

## Story shapes that work

- The grandmother's recipes.
- The competition.
- Rival restaurants, one street.
- The palace cook in an ancient court.
- The food truck romance.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Mediterranean Coast, Urban Opulence; Costume Style Professional Uniforms; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the chef with whites and street-clothes states; the signature dish as a prop; the kitchen and the dining room as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The trattoria. | Grandmother's recipe book is missing. |
| 2 | The rival restaurant. | Her grandmother's dish on their menu. |
| 3 | The competition kitchen. | Round one: the judge is the rival's owner. |
| 4 | The trattoria at night. | A note in the flour tin. |
| 5 | The tasting. | Silence, then one question. |

## FAQ

### How do I make a cooking drama with AI?

Make the kitchen and dining room locations in SceneMixer, dress the brigade in Professional Uniforms, make signature dishes props, and tell service in short shots.

### How do I make food look good?

Close-ups with the light and steam written in.

### Can a dish look the same across episodes?

Yes. Make it a prop with a reference picture.

### Will the head chef look different from the cooks?

Professional Uniforms keeps one cut per role; say who is who in the script.

### Can it be a competition show?

Yes. Make the studio a location and the judges characters.

### Which visual style suits a cooking drama?

Digital HD or Soft Glow.

### Can I use a real restaurant brand?

Invent it. Real brands can trigger the IP check.

### Vertical or landscape?

9:16 Portrait suits food and feeds.

Updated November 3, 2026


---

# AI Christmas Ad: A Holiday Spot From One Product Photo

URL: https://scenemixer.com/guides/make-an-ai-christmas-ad/
Make a Christmas ad with AI from one to four photos: a warm or twist-ending spot of 15, 30 or 60 seconds, snow and lights written in, the offer on the end card.

# How to make an AI Christmas ad

Quick answer: Christmas 2026 is Friday, December 25. A Christmas ad in SceneMixer starts the same way as any ad: upload one to four photos of what you sell, pick a tone, a length and a format, and the spot is written, shot and finished with an end card carrying your brand and offer. What makes it a Christmas ad is in the one-line description: a gift for…, our Christmas edition, order by the 19th for delivery. Snow, lights and a tree come from the words, the music is written for the ad, and nothing legible is drawn into the picture, so the offer lives on the end card and in the spoken line. Make it in November: the picture is ready weeks before the sale, and a second cut for the January sale reuses the same shots.

One line for the box next to the upload: Our hand-poured soy candles in a Christmas gift box of three; for people who hate giving gift cards; free shipping until December 19.

## Count back from the 25th

| When | What to do |
| Now | Choose the gift product. Take or clean up one to four photos: the product, the box, a detail, the product in use. |
| Early November | Make the 30-second main ad in portrait. Check the end card: brand, product name, tagline, Order now. |
| Mid November | Fix any shot that missed, rebuild the cut, export. Make a 15-second teaser from a second photo. |
| Late November | Teaser out, behind the Black Friday noise or after it. The Black Friday ad. |
| December 1–19 | Main ad out. The spoken line and end card carry the last shipping date. |
| December 20–24 | Edit the end card to Gift cards and local pickup and rebuild the cut: same shots, new ending. |
| December 26 onwards | Change the end card again for the January sale. No reshoot. |

The earlier the first cut exists, the more room there is to reshoot one shot rather than all of them. The whole ad flow, step by step.

## Which photo, and which tone

The first photo is the main subject and the end card is built around it. A product is redrawn as a clean studio shot; a shop front or a cafe is kept as it is and becomes the place the ad happens in, with any customers in the photo removed first. Up to four photos of the same thing: the gift box closed, open, a detail, the thing being used. Which photos work.

| Tone | Christmas use |
| Warm | Family, a table, a gift opened; the classic Christmas spot. |
| Twist | It looks like a snowstorm, then it is the glitter on the box; ends on the offer. |
| Minimal | The product, one light, one line; for brands that do not shout. |
| Mouth-watering close-ups | Food and drink: steam, pour, cut, for hampers and festive menus. |
| Funny | A gift that goes wrong, then yours; mass-market items. |
| A day in the life | Clothing and everyday things worn through a December morning. |

In SceneMixer, length and quality move the price; tone and format do not. Portrait for feeds and stories, landscape for a website or a screen in the shop. Length, format and quality.

## Snow, lights, and where they come from

You do not pick a Christmas set. The one line and the tone are read, and the script puts the product where Christmas happens: a kitchen with a tree in the corner, a street with lights, a snowy doorstep. If you need a specific setting, write it into the line: filmed in a small bookshop with a window display. Decorations are generated; the product stays exactly as photographed. Real people in your photos are removed; the ad's cast is generated to fit the buyer you describe. An ad for a shop · An ad for a product.

## The offer goes on the end card

No video model writes legible text reliably, so SceneMixer never draws prices or dates into the picture. Three places carry the offer: the spoken line (Free shipping until the nineteenth), the end card (brand, product name, a tagline with the offer, and Order now), and your post caption. Edit the end card after the cut is made and rebuild without reshooting: the same spot serves the sale, the last-shipping-day push and the January clearance. The end card · The price is a cap.

## A Christmas video that is not an ad

If what you want is a Christmas story rather than a spot, that is a drama project: a Christmas romance, a bedtime story for the week before, a greeting from an animated mascot. A Christmas romance movie · Bedtime stories · A New Year video.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Picker | Options | Changes the price? |
| Vibe | Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life | No |
| Length | 15, 30 or 60 seconds (one, two or four segments) | Yes |
| Format | Portrait 9:16 or Landscape 16:9 | No |
| Quality | 480P, 720P or 1080P | Yes |

## FAQ

### How do I make a Christmas ad with AI?

Upload one to four photos of the gift product, write one line with the Christmas angle and the offer, pick a tone and a length, and the spot is written, shot and finished with an end card.

### Can I put the price or the shipping date in the video?

Not in the picture; text is not drawn reliably. Put it in the spoken line, on the end card and in your caption.

### Can I reuse the ad for the January sale?

Yes. Edit the end card and rebuild the cut with the same shots; no reshoot.

### Which tone works for Christmas?

Warm for family and gifts, Twist for a reveal that ends on the offer, Minimal for premium goods, close-ups for food and drink.

### Does it add snow and a tree?

The setting is written from your line and tone: Christmas rooms, lights and snow are generated; the product stays as photographed.

### How long should a Christmas ad be?

30 seconds for one product, 15 for a teaser, 60 for a small collection with several photos.

### What does it cost?

SceneMixer quotes one figure before it starts, and it is a ceiling; length and quality set it, tone and format do not.

### Can the ad be in another language?

Yes. The spoken line and end card follow the language you choose; the same spot can be made again in a second language.

Updated November 3, 2026


---

# AI Christmas Movie, Hallmark-Style: Small Town, Snow, a Deadline

URL: https://scenemixer.com/guides/make-an-ai-christmas-romance-movie/
A Hallmark-style holiday romance with AI: a city career woman, a small town, a festival to save. Write it, light the snow, and post it in early December.

# How to Make an AI Christmas Romance Movie or Series

Quick answer: The holiday romance is a formula viewers love because it is a formula: someone from the city comes back to a small town at Christmas, a local business or festival needs saving, and the person who annoys them most turns out to be the one. To make one in SceneMixer, set the town in step 1 (American Small Town with Soft Everyday or Outdoor Gear wardrobe), write snow, string lights and the fire as light in each shot rather than as separate images, keep the festival crowd in the background line, and plan backwards from the date you want to post.

Brief for AI Create: A Christmas romance: a big-city developer is sent to her hometown to close the sale of the old inn before Christmas Eve; the innkeeper's son is running the tree-lighting festival to save it; she falls for the town before she falls for him; warm, funny, snowy.

## The beats of a holiday romance

- The arrival. A city person, a car that breaks down or a train that stops, and a town that is far too cheerful.
- The thing to save. An inn, a bakery, a tree farm, the festival itself. It gives the couple a reason to work together.
- The traditions. Decorating, baking, a sleigh ride, the tree lighting. Each one is a scene and a step closer.
- The secret. She was sent to close it down. The audience knows; he finds out at the worst moment.
- The choice on Christmas Eve. The job or the town. Snow, lights, a crowd, a kiss the camera leaves early.
It works as a single film of a few minutes or as a series that runs through December, one episode a day.

## A Hallmark-style holiday movie on a December schedule

The Hallmark-style Christmas movie has a formula people love to watch every year: a city career woman, a small town, a festival or an inn to save, and a choice on Christmas Eve — the beats above. Holiday movies are watched in December, so the aim is to post in the first days of the month. Starting in late October or November leaves room for it:

- Week 1 — script and cast. Write it yourself, or give AI Create the arrival, the thing to save, the secret and the Christmas Eve choice. Read the script, then generate the couple, the regulars, the inn, the main street and the tree farm on the default image tier, where your free sheets are drawn.
- Week 2 — episode 1. Generate its storyboard (free on your first project), add a light source to every evening shot, and render it on SceneMixer's default tier at 10 credits a second. Watch it before you go on.
- Week 3 — the rest. Episodes 2 to 5, then export each one. With Subtitles on your account, tick Add subtitles when you export.
- The poster. Cover art in step 2 with the Romance genre mood paints three posters with your title painted in; Make episode covers adds “EP 1” to “EP 5”. Promo cover posters.

Hallmark is a trademark of its owner; it is named here only to describe the genre.

## Settings to pick in step 1

- Type: Film for one holiday movie; Short Drama for a December series with a hook at the end of each episode.
- Aspect Ratio: 16:9 Landscape for a movie on YouTube or a TV; 9:16 Portrait for a daily series on phones.
- Visual Style: Digital HD or Soft Glow for the warm, glowing look; Family 3D Animation or Watercolor Storybook for a family version.
- Scene Style: American Small Town for the main street and the inn, Frozen Wilderness for the tree farm and the snowbound road, Nordic Minimal for a Scandinavian take.
- Costume Style: Soft Everyday for knitwear and coats, Outdoor Gear for the tree farm. Her city clothes go in her own description, or as a state if they change for good.

Visual Style: Soft Glow

Scene Style: American Small Town

Scene Style: Frozen Wilderness

Costume Style: Soft Everyday

Costume Style: Outdoor Gear
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Snow, lights and the fire: write them as light

The look of the genre is warm light against cold. A video model gives you that when the shot names the sources:

- String lights along the porch rail, warm against the blue dusk; snow falling slowly through them.
- The fire in the inn’s hearth lights their faces from the side; the window behind them is dark blue.
- The town tree lights up all at once; the crowd’s faces turn gold.
Snow and falling snow are written into the shot, not made into separate images: they ride on top of whatever the location image shows, like rain. Save state images for real changes — the inn after the storm damage, the main street fully decorated for the last episode. States · Describe the cause, not the result.

## The festival crowd and the town’s regulars

- Name only the people with lines. The baker, the mayor, the best friend: characters with cards. The crowd at the tree lighting is the segment’s background line. Background people.
- Santa is your own character. A generic bearded man in a red suit is fine; anything that copies a film or brand version of the character is not. Describe yours.
- Children in the scene: keep them to ordinary family moments, in the background or with their parents, as a holiday film would.
- A dog by the fire is a non-human character: breed, colour, size, no clothes unless you want the jumper. Non-human characters.

## Three shots to copy

| Beat | Shot |
| The arrival | [Shot 1] 5s: Opening: wide shot. A sleek city car stuck at the edge of a snowy main street strung with lights. @Claire climbs out in heels into the snow. @Claire (under her breath): Of course. Ambient: carols from a shop doorway. |
| The spark | [Shot 4] 4s: Medium shot, the inn kitchen. @Ben hands @Claire a rolling pin, flour on his sleeve. @Ben (dry): You said you could bake. @Claire lips closed, no lip movement. |
| The tree lighting | [Shot 9] 5s: Wide shot. The town tree lights up; the crowd cheers; @Claire and @Ben in the foreground, not looking at the tree. No dialogue. Ambient: cheering, bells. |
How to write one shot.

## Music and timing

- The score. An episode can carry its own score; the composite’s Music switch is on by default, and you can turn it off to lay your own track. Carols that are still under copyright are not yours to add unless you have a licence. Background music.
- Work backwards from the date. A series that should run through December needs its first episodes finished and its later ones storyboarded by the end of November. Generate the first three, post them, and keep going in the same project. Make the next episode.

## Make it in SceneMixer, start to finish

- Script: yours on Upload Script, or a brief in AI Create with the arrival, the thing to save, the secret and the Christmas Eve choice.
- Step 1: the settings above; read the script.
- Step 2: generate the couple, the regulars, the inn, the main street and the tree farm.
- Storyboard: episode 1 first (free on your first project). Add a light source to every evening shot.
- Video: generate on the default tier at 10 credits a second; composite; post.

| Episode | What happens | The hook |
| 1 | She arrives to close the deal on the inn. The car will not start. | The only room left is at the inn she is there to sell. |
| 2 | He drafts her into the festival committee. | She finds her own childhood drawing on the inn’s wall. |
| 3 | Baking night; the town adopts her. | Her boss calls: the deal closes on Christmas Eve. |
| 4 | He finds the contract. | The tree lights go out the night before the festival. |
| 5 | Christmas Eve. | She tears up the contract in front of the whole town — and he is not there to see it. |

## FAQ

### Can AI make a Hallmark-style Christmas movie?

Yes, as a film or a five-episode series: the couple, the town and the inn are drawn once and reused in every shot, and the characters speak their lines on screen.

### Can AI write a Hallmark-style Christmas movie script?

Yes. Give AI Create the arrival, the thing to save, the secret and the Christmas Eve choice, and it writes the script; SceneMixer then reads it into episodes, characters and places. Writing the one-line brief.

### Can I make a full Christmas movie with AI?

Yes. In SceneMixer, choose Film as the type, set a small-town scene style, and write or generate the script. The cast and locations stay consistent from the first scene to the last.

### How do I make snow and Christmas lights look right?

Write them into the shot as light: string lights against the dusk, the fire lighting faces, the tree lighting up. Snow is written into the shot, like rain, not made into a separate image.

### Can Santa Claus be in my film?

Yes, as your own character with your own description. Avoid copying a particular film's or brand's version of him.

### Can I add Christmas songs?

Only music you have the right to use. Episodes can carry their own generated score, and you can switch it off on the composite to add a licensed track yourself.

### How long should a Christmas movie be?

A few minutes works well as a single film; a December series can run one short episode a day.

### When should I start to post in December?

Finish the first episodes by late November so the series can run through December. Storyboard ahead while the first ones are out.

### Can I make it animated for families?

Yes. Pick Family 3D Animation or Watercolor Storybook as the visual style in step 1.

### Can the same story be made in another language?

Yes. Start a second project from the same script with another dialogue language and region.

Updated November 3, 2026


---

# AI Cinderella Story Drama: Rags to Riches, Then the Reveal

URL: https://scenemixer.com/guides/make-an-ai-cinderella-story-drama/
Make a Cinderella-style rags-to-riches drama with AI: the before and after looks as states, a ball or gala as the turning point, and a stepfamily who sees it too late.

# How to make an AI Cinderella story drama

Quick answer: A Cinderella story is a woman nobody sees, one night when everyone does, and the moment her stepfamily realises. In SceneMixer, her two looks are two states of one character: the plain clothes and the gown. The ball can be a royal one (European Palace, European Court Gown) or a modern gala (Urban Opulence, Haute Couture). The fairy tale itself is public domain; famous animated versions are not, so design your own gown, shoe and godmother.

Brief for AI Create: A modern Cinderella: a cleaner at a luxury hotel borrows a guest's lost invitation to the hotel's charity gala; the CEO she dances with is the guest who lost it, and her stepsisters are at the next table; 10 short episodes, each ends with someone almost recognising her.

## Royal or modern

| Version | Scene Style | Costume Style |
| Classic royal ball | European Palace | European Court Gown |
| Modern gala | Urban Opulence | Haute Couture |
| Office Cinderella | Corporate Cold | Power Tailoring |
| Fairy tale for kids | Medieval Castle | Medieval Nobility with Watercolor Storybook visuals |
Fairy tale videos · Royal dramas.

## Before and after are states

- Two looks, one face. The cleaner's uniform and the gala gown are states; the face stays hers. States.
- The lost object (a shoe, an earring, an invitation) is a prop. Props.
- The stepfamily each gets one clear mark; they will be in every episode. Same face.
- Original design. A blue gown and glass slippers in the famous cut will trigger the IP check; change the colour and the object. IP check.

## The gala, shot by shot

| Beat | Shot to adapt |
| The entrance | [Shot 1] 4s: Wide shot, a hotel ballroom under crystal chandeliers, guests in evening dress in the background. @Ella walks in wearing a deep green gown. |
| The table | [Shot 2] 3s: Medium shot. @Brianna and @Celeste at a table, turning to look. |
| The question | [Shot 3] 4s: Close-up on @Brianna. @Brianna (puzzled): Don't we know her? (voice: Brianna) |
| The hand | [Shot 4] 3s: Close-up. @Daniel offers his hand to Ella. |

## Story shapes that work

- The hotel cleaner.
- The secret heiress working as a maid.
- The office makeover.
- The reversed Cinderella: a poor man at a rich woman's ball.
- After midnight: what happens when he finds her.
Heiress dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, European Palace; Costume Style Haute Couture, European Court Gown; Visual Style Soft Glow, Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the heroine with an everyday state and a gown state; the stepfamily with one mark each; the lost object as a prop; the ballroom as a location. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The hotel's service corridor. | She finds the lost invitation. |
| 2 | The gala. | She dances with the guest who lost it. |
| 3 | Midnight. | She runs and leaves an earring. |
| 4 | Her stepmother's flat. | Her stepsister wears the matching earring. |
| 5 | The hotel lobby. | He asks the staff to line up. |

## FAQ

### How do I make a Cinderella story with AI?

Give the heroine two states (everyday and gown), choose European Palace or Urban Opulence in SceneMixer's Step 1, and make the lost object a prop.

### Can I use the famous animated Cinderella?

No. That design is someone's IP. The fairy tale is public domain; your gown, shoes and characters must be your own.

### How do I show the transformation?

Switch from her everyday state to her gown state between segments.

### Can it be modern?

Yes. A hotel gala in Urban Opulence with Haute Couture works well.

### Can I make a children's version?

Yes. Use Watercolor Storybook or Family 3D Animation as the visual style.

### Which visual style suits a romance version?

Soft Glow or Digital HD.

### Vertical or landscape?

9:16 Portrait for short-drama feeds.

### How do I show a whole ballroom?

Background guests in the background line and a few named people up close.

Updated November 3, 2026


---

# Make an AI Claymation Short: Handmade Clay Look, No Stop-Motion

URL: https://scenemixer.com/guides/make-an-ai-claymation-short/
Make a claymation-style short with AI: the Claymation visual style with clay figures and miniature sets, simple stories, a small cast and gentle comedy.

# How to make an AI claymation short

Quick answer: The Claymation visual style in SceneMixer draws everything as hand-pressed clay figures and miniature sets, fingerprints left in, without a single frame of stop-motion. Choose it in Step 1, keep the cast small and the sets simple (a kitchen, a garden, a shop), and lean into what clay does best: gentle comedy, small worlds and expressive faces.

Brief for AI Create: A claymation short: an elderly sheep who runs a tiny tea shop gets her first customer in a year, a nervous wolf who only wants a scone; a three-minute comedy, one room, one twist.

## Why clay works

- Small worlds. One room, a garden, a street of shops: miniature sets look their best close up.
- Expressive faces. Write the reaction (“her eyebrows shoot up”), not the feeling. Describe the cause.
- Gentle comedy. Pauses and small gestures land well; give them their own shots. One shot.

## Set it up

- Visual Style Claymation, fixed once you continue; scene and costume cards switch to drawn pictures to match. Visual styles.
- Scene Style for the world around the set: Pastoral Healing, American Small Town, Nordic Minimal.
- Animals are characters with a voice. Non-human characters.
- Type Film for a single short, TV Series for a run of episodes.

## The tea shop, shot by shot

| Beat | Shot to adapt |
| The shop | [Shot 1] 4s: Wide shot, a tiny clay tea shop, a bell over the door, three empty tables. |
| The bell | [Shot 2] 2s: Close-up on the bell jingling. |
| The wolf | [Shot 3] 3s: Medium shot. @Wolfram, a tall grey clay wolf in a knitted scarf, ducks under the door frame. |
| The line | [Shot 4] 4s: Close-up on @Wolfram. @Wolfram (very polite): Just one scone, please. I'm not here to eat anyone. (voice: Wolfram) |
| The sheep | [Shot 5] 3s: Close-up on @Mrs Woolly, eyebrows shooting up, teapot frozen in mid-pour. |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday; Visual Style Claymation; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: a small cast of clay characters with one mark and a voice each; the shop or the garden as a location. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Part | What happens | Shots |
| Set-up | The empty shop. | 2–3 |
| Arrival | The wolf comes in. | 3–4 |
| Misunderstanding | Everyone hides. | 4–5 |
| Twist | The wolf is the new baker next door. | 3 |
| End | Two cups, one scone. | 2 |

## FAQ

### How do I make a claymation video with AI?

Choose Claymation as the visual style in SceneMixer's Step 1, keep the cast and sets small, and write the story as short shots.

### Is it real stop-motion?

No. It is generated video in a clay look: hand-pressed figures and miniature sets, fingerprints included.

### Can animals be characters?

Yes. Give them a sheet and a voice.

### Which stories suit clay?

Gentle comedy, small worlds, one-room stories and children's tales.

### Can I make a series?

Yes. Choose TV Series as the type.

### Can I copy a famous clay character?

No. Known characters trigger the IP check.

### Vertical or landscape?

Either; 16:9 Landscape for a short film, 9:16 Portrait for feeds.

### Can the visual style change later?

No. It is fixed once you continue past Step 1.

Updated November 3, 2026


---

# AI Contract Marriage Drama: The Deal, the Rules and the Slip

URL: https://scenemixer.com/guides/make-an-ai-contract-marriage-drama/
Make a contract marriage drama with AI: the contract as a prop, Wedding and Power Tailoring states, one shared home as the stage, and the moment pretending stops.

# How to make an AI contract marriage drama

Quick answer: A contract marriage drama is a lie two people agree to tell, and the story of it becoming true. In SceneMixer, make the contract a prop that returns, give the leads a wedding state and an everyday state, set most scenes in the one home they share, and build each episode toward a moment where one of them forgets they are pretending.

Brief for AI Create: A contract marriage drama: an heir needs a wife for one year to inherit; a florist needs money for her father's surgery; rule three says no falling in love; twenty short episodes.

## The contract is a prop

Draw the contract once: the paper, the signatures, the clause everyone remembers. It appears at the signing, in the drawer, in the fight, and torn at the end. Props.

| Place | Scene Style | Costume Style |
| His home | Urban Opulence or Suburban Mansion | Minimal Luxury, Loungewear |
| Her world | A flower shop in Parisian Streets or American Small Town | Soft Everyday |
| The wedding | English Manor or Island Resort | Wedding |
| His family | Urban Opulence dining room | Power Tailoring |

## States for the performance

The wedding outfit, the public couple look, the at-home clothes: each a state of the same face. The difference between how they dress for others and for each other tells the audience how far the pretence has slipped. States.

## The slip, shot by shot

| Beat | Shot to adapt |
| The dinner | [Shot 1] 4s: Wide shot, a long dining table, his family on both sides. @Lily and @Ethan side by side. |
| The test | [Shot 2] 3s: @Grandmother: How did he propose, dear? |
| The answer | [Shot 3] 4s: Close-up on @Lily. @Lily (smiling): He didn't. I asked him. He said yes before I finished. |
| The slip | [Shot 4] 2s: Close-up on @Ethan, looking at her a second too long. |
Billionaire romance · Hooks.

## Rules that make episodes

- Write three or four contract rules; break one per arc.
- Give the deadline a date and show it on a calendar prop.
- Let a third person suspect the truth.
- End the season on the day the contract expires.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, English Manor; Costume Style Wedding, Minimal Luxury; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: a wedding, a public and an at-home state for each lead; the contract as a prop; the shared home as the main location. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Rule | The ending beat |
| 1 | The signing | Rule three: no falling in love |
| 4 | Separate rooms | A storm, one blanket |
| 8 | Public affection only | A kiss with no one watching |
| 14 | No questions about the past | His first fiancée returns |
| 20 | The contract ends | Who tears it up |

## FAQ

### How do I make a contract marriage drama with AI?

Make the contract a prop, give the leads wedding and everyday states, set it in one shared home, and break one rule per arc.

### How do I show the wedding?

A Wedding costume state for both leads and a wedding location such as English Manor or Island Resort.

### Which format suits it?

Short Drama, 9:16 Portrait, one to two minutes an episode.

### Which look suits it?

Digital HD for modern gloss; Soft Glow for a softer romance.

### How many episodes?

Twenty to forty short episodes is common.

### Can it be a period story?

Yes: Regency or Victorian costumes with English Manor.

### Can there be kissing?

Romance is fine; explicit content is not.

### Can I set it in another market?

Yes; set Target Region and dialogue language in Step 1.

Updated October 31, 2026


---

# AI Corporate Training Video: Short Dramas Staff Remember

URL: https://scenemixer.com/guides/make-an-ai-corporate-training-video-drama/
Make corporate training videos as short AI dramas: one workplace situation per episode, recurring staff characters, the wrong way then the right way, many languages.

# How to make an AI corporate training video as a short drama

Quick answer: People remember stories better than slides. A training drama shows one workplace situation per episode with the same small cast: the wrong way, what it costs, then the right way. In SceneMixer, write the scenario as a short script, dress the cast in Professional Uniforms or Smart Casual, set the office once, and make the same episode in other dialogue languages for other teams.

Brief for AI Create: A training drama for retail staff: a customer wants to return a broken kettle without a receipt; version one goes badly, version two follows the policy kindly; two minutes, three characters.

## The format that works

| Part | Length | What it does |
| The situation | 20–30 s | A real moment staff will recognise |
| The wrong way | 30–45 s | A plausible mistake, not a cartoon villain |
| The consequence | 10–20 s | A face, a lost customer, a risk |
| The right way | 30–45 s | The same moment, done well |
| The one line | 5–10 s | A narrator states the rule |

## A cast staff will recognise

Keep the same three or four characters across the series: the new hire, the experienced colleague, the manager, the customer. Recurring faces make it a series people follow. Set the workplace once in SceneMixer: Corporate Cold for offices, Hospital Ward for care settings, a shop or warehouse written as a location. Why a series stays consistent.

## Rewinding the scene

The wrong way and the right way happen in the same place with the same people, so the difference is only the behaviour. Because the location and cast are drawn once, both versions match exactly. A narrator's voice-over can say Let's try that again between them.

| Beat | Shot to adapt |
| The counter | [Shot 1] 3s: Medium shot. @Customer puts a kettle on the counter. @Customer: It broke on day two. |
| Wrong way | [Shot 2] 3s: @Sam (flat): No receipt, no return. |
| Rewind | [Shot 3] 2s: @Narrator (voice-over): Let's try that again. |
| Right way | [Shot 4] 4s: @Sam (warm): I'm sorry. Let me look it up by your card. |

## Languages and delivery

Make the same episode again with another dialogue language and region for another office. Subtitles can be burned in or exported as SRT for a learning platform (accounts that have made a purchase). Second language version · Using SceneMixer as a team.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Corporate Cold; Costume Style Professional Uniforms, Smart Casual; Visual Style Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: three or four recurring staff characters and a narrator; the counter or office as the single location. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Situation | Rule |
| 1 | A return without a receipt | Look for another way to verify |
| 2 | A customer who is upset | Acknowledge before you explain |
| 3 | A colleague cutting corners | Say it once, privately |
| 4 | A safety hazard | Stop, mark, report |
| 5 | A data request | Check identity first |

## FAQ

### Can AI make corporate training videos?

Yes, as short dramas: one situation per episode, with a recurring cast and a narrator for the rule.

### Can I show the wrong way and the right way?

Yes. Both happen in the same location with the same cast, so only the behaviour changes.

### Can I make it in several languages?

Yes. Make the episode again in SceneMixer with another dialogue language and region.

### Can I add subtitles for an LMS?

Yes: burned in or SRT, on SceneMixer accounts that have made a purchase.

### Can I use our logo and office?

SceneMixer does not draw brands and logos reliably, and well-known brands trigger the IP check; describe the space instead.

### Can several people work on it?

A SceneMixer project belongs to one account; others review through share links, which open without an account.

### How long should each video be?

Two to three minutes.

### Which look suits training?

Digital HD for realism, or Family 3D Animation for a lighter tone.

Updated November 3, 2026


---

# AI Courtroom Drama: Trials, Lawyers and the Witness Who Lies

URL: https://scenemixer.com/guides/make-an-ai-courtroom-drama/
Make a courtroom or legal drama with AI: Forensic Cold White court spaces, Power Tailoring for lawyers, evidence as props and a verdict per episode.

# How to make an AI courtroom drama

Quick answer: A courtroom drama is a question, a witness and a lawyer who already knows the answer. In SceneMixer, Forensic Cold White covers police and court spaces; Power Tailoring dresses the lawyers, Professional Uniforms the clerks and officers. Make the courtroom a location used in every episode, every exhibit a prop with its own picture, and cut between the witness box, the bench and the jury instead of framing the whole room.

Brief for AI Create: A courtroom drama: a public defender is assigned to defend the man accused of killing her mentor; the evidence is perfect, which is exactly what bothers her; 10 short episodes, each ends with an objection.

## Set the court

| Setting | Scene Style | Costume Style |
| A modern courtroom and police station | Forensic Cold White | Power Tailoring, Professional Uniforms |
| A big law firm | Corporate Cold | Power Tailoring, Minimal Luxury |
| A small-town trial | American Small Town | Smart Casual |
| A period trial | English Manor or Republican Era | Victorian, Republican Modern |

## Exhibits, the jury and the gallery

- Every exhibit is a prop. The knife in the evidence bag, the photograph, the phone: the same in every episode. Props.
- The jury and the gallery are background. Two or three named faces in each shot. Background people.
- Long speeches need room: a line needs time to be spoken, so split a closing argument across several shots. Shot timing.
- Legal words are fine in dialogue; write them exactly. Dialogue in your script.

## Cross-examination, shot by shot

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, a courtroom of pale wood and white walls, the jury in the background. @Ana stands before the witness box. |
| The exhibit | [Shot 2] 3s: Close-up. @Ana holds up a sealed photograph. |
| The question | [Shot 3] 4s: Close-up on @Ana. @Ana (quiet): Then who took this picture, Mr Hale? (voice: Ana) |
| The lie | [Shot 4] 3s: Close-up on @Mr Hale in the witness box, swallowing. |
| The bench | [Shot 5] 3s: Medium shot. @Judge Park leans forward. |

## Story shapes that work

- The case she cannot win.
- Rival lawyers who used to be married.
- The wrongful conviction: a second trial.
- The law firm: one client per episode, one partner with a secret.
- The juror: the one who does not agree.
Police and crime dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Forensic Cold White, Corporate Cold; Costume Style Power Tailoring, Professional Uniforms; Visual Style Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: lawyers, the accused and witnesses with one mark each; every exhibit as a prop; the courtroom and the office as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The public defender's office. | The case file has her mentor's name on it. |
| 2 | The jail. | Her client says he was there to warn him. |
| 3 | Court, day one. | The prosecution's evidence is perfect. |
| 4 | Her mentor's office. | A missing page from his calendar. |
| 5 | Court, day two. | Objection, and the judge allows it. |

## FAQ

### How do I make a courtroom drama with AI?

Choose Forensic Cold White and Power Tailoring in SceneMixer's Step 1, make the courtroom a location and each exhibit a prop, and cut between the witness, the lawyer and the bench.

### Can I film a whole jury?

Keep the jury in the background line and show a few named faces up close.

### How do I fit a long closing speech?

Split it across several shots; each line needs time to be spoken.

### Can I base it on a real trial?

Do not depict real people. Invent the case and everyone in it.

### Which costume style suits lawyers?

Power Tailoring for lawyers, Professional Uniforms for court officers and police.

### Which visual style?

Digital HD for modern courts, Shot on Film for a period trial.

### Vertical or landscape?

16:9 Landscape suits a courtroom; 9:16 Portrait suits short-drama feeds.

### Can the dialogue be in another language?

Yes; choose the dialogue language in Step 1.

Updated November 3, 2026


---

# AI Cozy Farm Life Drama: Slow Living, Seasons and a House to Fix Up

URL: https://scenemixer.com/guides/make-an-ai-cozy-farm-life-drama/
A farm life drama with AI: Pastoral Healing, Chinese Village or Ancient Farmstead scenes, a field that changes with the seasons, and small stakes people keep watching.

# How to make an AI cozy farm life drama

Quick answer: A cozy farm life drama runs on small stakes and visible progress: a roof mended, a first harvest, a neighbour who finally says hello. In Step 1 the scene style sets the world. Pastoral Healing is the idealised slow country life — log cabins, flower fields, streams and handmade things. Chinese Village is today’s Chinese countryside. Ancient Farmstead is farming, famine and family-split stories in ancient China, with rammed-earth walls, grey tile, bamboo fences and terraced fields. Pair it with a soft costume style, a gentle visual style, and write the seasons as two looks of the same field.

Brief for AI Create: A cozy farm drama: a laid-off accountant inherits a run-down orchard from an aunt she barely knew and has one growing season to make it pay; the neighbour who keeps leaving tools on her fence is the only person who knows why the orchard was never sold; 8 short episodes, every episode ends on something small growing.

## Pick the farm

| Kind of farm story | Scene Style | Costume Style |
| The dream cottage and the orchard | Pastoral Healing | Soft Everyday, Loungewear or Outdoor Gear |
| Today’s Chinese countryside | Chinese Village | Loungewear, Outdoor Gear or Smart Casual |
| Farming in ancient China | Ancient Farmstead | Jianghu Travelwear for working clothes, with rough cloth written into each character’s clothing |
| A ranch | Wild West | Western Cowboy |
| A small American town farm stand | American Small Town | Soft Everyday |
There is no dedicated costume preset for ancient farm clothes, so take the closest one and describe the rough cloth, the patches and the straw hat in the character’s clothing. All of this is written into the descriptions when the script is read, so choose it first. Why they are fixed afterwards · Every scene and costume style.

## A gentle look, and a region to match

The visual style is fixed after Step 1. Soft Glow and Digital HD for live action; Watercolor Anime, Watercolor Storybook or Family 3D Animation for a drawn, warmer look. Visual style presets, one by one.

For Target Region, Western gives a Western cast and names, East Asia an East Asian one, and Follow script keeps whatever the script says. Target Region.

## Seasons, weather and the passing of time

A farm drama shows time passing in the land. Two things to know:

- The same field in two seasons is a place with two looks. A bare plot in spring and a gold field in autumn are different pictures of one place, so make the second one a state of the location in Step 2 and pick it for the later segments. Costume changes and scene states.
- Weather and time of day are settings on each segment. Pick Dawn to Late night and Clear, Cloudy, Overcast, Rain, Storm, Snow, Fog or Windy on the segment header. Rain is just a setting, not a state. The segment header, field by field.

## Small stakes, shot by shot

Cozy dramas are paced in small beats: a kettle, a hand on a gate, a long look at a field. Give each one its own shot of two or three seconds, and end the scene on something that changed.

| Beat | Shot to adapt |
| The first morning | [Shot 1] 4s: Wide shot, fixed camera. A run-down farmhouse at dawn, an orchard behind it in light mist, one window with the shutters hanging open. @Nora stands in the yard with a mug in both hands. |
| The gate | [Shot 2] 3s: Close-up on @Nora’s hand lifting the latch of a weathered gate. It sticks, then gives. |
| The neighbour | [Shot 3] 3s: Medium shot. @Walt leans on the fence with a coil of rope over his shoulder, not quite looking at her. |
| The line | [Shot 4] 4s: Close-up on @Walt. @Walt (gruff, kind): That gate hasn’t opened since March. (voice: Walt) |
| The change | [Shot 5] 3s: Close-up on a single small apple blossom on the old tree, newly open. |
How to write one shot · Hooks and cliffhangers.

## Animals, neighbours and a crowd at the market

- Animals that appear often (a dog, a cat, chickens) are characters in Step 2 with their own sheet. Cat and pet dramas.
- The market or the fair is a crowd: write it as a background line in the shot and give names only to the few faces the story needs. Background people in a scene.
- Tools, baskets and jars that come back are props with a reference image. The same prop in every hand.

## Story shapes that work

- The inheritance: a run-down farm and one growing season.
- The move to the country: someone from the city learning what a neighbour is.
- The first harvest: every episode a stage of one crop, the last a table.
- The small business: jam, bread, cheese, tea; a stall, then two.
- The land dispute: a family, a boundary stone and a long-kept promise.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief in AI Create.
- Step 1: Type Short Drama or TV Series; scene and costume styles from the table; a gentle visual style; Target Region; dialogue language; 9:16 Portrait or 16:9. Read the script.
- Step 2: a specific face and one mark per character; the farmhouse, the field and the market as locations, and a second look for the field in the later season. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). One place per segment.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | The farmhouse at dawn. | The old gate opens for the first time since March. |
| 2 | The orchard in mist. | A rope on the fence, left there on purpose. |
| 3 | The weekly market. | She sells the first jar and nobody buys the second. |
| 4 | A storm over the field. | The neighbour’s tools are in her shed, and he is not. |
| 5 | Late summer. | The orchard has its first full row. |

## FAQ

### How do I make a cozy farm life drama with AI?

In SceneMixer's Step 1, choose a scene style such as Pastoral Healing, Chinese Village or Ancient Farmstead, a soft costume style and a gentle visual style, then write or generate the script and render episode by episode.

### What is the Ancient Farmstead scene style?

Ancient Chinese villages and market towns for farming, famine and family-split stories: rammed-earth walls, grey tile roofs, bamboo fences and terraced fields.

### Is there a costume style for ancient farmers?

SceneMixer has no dedicated one. Take the closest, such as Jianghu Travelwear, and write the rough cloth and patches into each character's clothing.

### How do I show the seasons changing?

Make the later season a second look of the same location in Step 2 and pick it for the later segments. Weather and time of day are settings on each segment.

### Can animals be characters?

Yes. A dog, a cat or chickens that appear often get their own sheet in Step 2.

### Which visual style suits a cozy drama?

Soft Glow or Digital HD for live action; Watercolor Anime, Watercolor Storybook or Family 3D Animation for a drawn, warmer look.

### How do I film a market full of people?

Write the crowd as a background line in the shot and give names only to the few faces the story needs.

### Will the farmhouse look the same in every episode?

Yes. Each location is drawn once and that picture is used in every shot, with a second look where you add one.

Updated November 3, 2026


---

# AI Cyberpunk Drama: Neon Cities, Implants and Corporate Towers

URL: https://scenemixer.com/guides/make-an-ai-cyberpunk-drama/
Make a cyberpunk drama with AI: Cyber Metropolis and Neon Nights sets, Cyberpunk Augment costumes, implants as states, rain set per segment, and screens kept wordless.

# How to make an AI cyberpunk drama

Quick answer: Cyberpunk is high tech and low life: a glittering tower above, wet streets below. In SceneMixer, choose Cyber Metropolis and Neon Nights in Step 1 with Cyberpunk Augment or Near-Future Techwear, make implants and augmentations part of each character's sheet (or a state when they change), set rain on the segment, and let screens glow rather than show readable text.

Brief for AI Create: A cyberpunk drama: a debt collector with a borrowed arm is sent to repossess the memory chip of a girl who remembers his real name; eight episodes, one rainy week, one corporate tower.

## Two levels of the city

| Level | Scene Style | Costume Style |
| The streets | Neon Nights | Cyberpunk Augment, Streetwear |
| The towers | Cyber Metropolis, Corporate Cold | Near-Future Techwear, Power Tailoring |
| The clinic | Hospital Ward made grimy | Professional Uniforms |
| The retro district | Retro Hong Kong | Rock & Punk |
Looks: Game Engine Realism for a game feel, Digital HD for live action, Painterly Anime for anime cyberpunk. The sample Neon Debt plays in a neon noodle-bar street.

## Implants on the sheet

An augmentation is part of how a character looks: write it into the description (a matte black prosthetic right arm with a cracked panel at the elbow) so it is on every shot. If it is installed or damaged mid-story, that is a new state. States · Describe a face.

## Rain, neon and screens

Rain and night are set on each segment's Time and Weather, not written into every shot. Neon reflections come with the scene style. Screens and holograms glow well but text on them is unreliable; let a character read it aloud. Text on screen.

| Beat | Shot to adapt |
| The street | [Shot 1] 4s: Wide shot, night, rain. A street of noodle stalls under stacked neon signs; @Ryn walks against the crowd. |
| The arm | [Shot 2] 2s: Close-up. @Ryn's black prosthetic hand flexes; a panel at the elbow sparks. |
| The target | [Shot 3] 3s: Medium shot. @Mio at a stall counter, a glowing chip behind her ear. |
| The line | [Shot 4] 3s: @Mio (without looking up): You're late, Ryn. You always were. |

## Story shapes

- The job that goes wrong: a courier, a thief, a collector.
- The stolen memory: who you are is a file someone owns.
- The corporation: one tower, one secret, one person inside.
- The android who wants out: robot and android drama.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Cyber Metropolis, Neon Nights; Costume Style Cyberpunk Augment, Near-Future Techwear; Visual Style Game Engine Realism, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: implants written into each sheet; a damaged state where needed; the street, the clinic and the tower as locations; rain set per segment. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Level | The ending beat |
| 1 | Streets | The target knows his name |
| 2 | Clinic | His arm is registered to someone else |
| 3 | Tower lobby | His employer is her father |
| 4 | Rooftops | The chip is a copy |
| 5 | The top floor | The original is in his own head |

## FAQ

### How do I make a cyberpunk drama with AI?

Use the Cyber Metropolis and Neon Nights styles in SceneMixer, Cyberpunk Augment costumes, implants written into each sheet, and rain set on the segment.

### How do I keep a character's implant consistent?

Write it into the character's description; changes become states.

### Can I show screens and holograms?

Yes; their glow renders well, but readable text is unreliable, so have a character say what it shows.

### Which look suits cyberpunk?

Game Engine Realism, Digital HD, or Painterly Anime for anime cyberpunk.

### How do I make it rain?

Set Weather on the segment; it is not needed in every shot.

### Can I copy a famous cyberpunk film or game?

No. Famous characters and designs trigger the IP check.

### Can there be fights?

Yes, one move per shot, without wounds.

### Vertical or landscape?

Landscape for city scale; vertical for a fast short-drama cut.

Updated November 3, 2026


---

# AI Dark Romance Drama: A Dangerous Lead, Dark Rooms, Clear Lines

URL: https://scenemixer.com/guides/make-an-ai-dark-romance-drama/
Make a dark romance drama with AI: a morally grey or dangerous lead, gothic and noir places, mood written into each shot, and the lines the content rules draw.

# How to make an AI dark romance drama

Quick answer: Dark romance is a love story where one of the two is dangerous: a crime heir, an assassin, a vampire, a man who keeps secrets for a living. In SceneMixer the darkness comes from the places (Gothic Dark, English Manor, Neon Nights), from light written into each shot, and from power: a debt, a deal, a secret. Sexual content and sexual violence are not allowed, and graphic violence is blocked, so the danger lives in threat, silence and choice.

Brief for AI Create: A dark romance: to pay her brother's debt, an archivist agrees to catalogue the private library of the man who owns the city's night; she finds her own family's name in his ledgers; twelve episodes, one house.

## What makes it dark

| Element | How it shows on screen |
| The dangerous lead | Calm, polite, always the one who decides when a conversation ends |
| The imbalance | A debt, a contract, a house she cannot leave until the work is done |
| The secret | A locked room, a ledger, a scar, as a prop or a state |
| The turn | The moment he chooses her over his own rules |

Write the danger as behaviour, not as violence: who stands, who sits, who is allowed to speak. A personality the video shows.

## Dark rooms, written per shot

Location images are deliberately evenly lit, so the camera can read the room. The mood belongs to each shot: one lamp on the desk, the rest of the library in shadow, neon through rain on the window. Night and rain come from the script for the scene. Why sheets are plain · Describe the cause.

Scene Style: English Manor

Scene Style: Neon Nights

Visual Style: Shot on Film

Visual Style: Black & White Film

Costume Style: Gothic Court
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## The deal, shot by shot

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, a long library at night, one lamp lit. @Elena stands at the door. |
| The man | [Shot 2] 3s: Medium shot. @Victor at the desk, not looking up, turning a page. |
| The terms | [Shot 3] 4s: Close-up on @Victor. @Victor: Three months. You touch nothing you were not given. |
| The answer | [Shot 4] 3s: Close-up on @Elena, the lamp on half her face. @Elena: And if I do? |
| The silence | [Shot 5] 3s: Two-shot. @Victor closes the book. No answer. |

## Where the lines are

Sexual content cannot be generated, sexual violence is never acceptable, and anything that sexualises a minor is zero tolerance. Weapons can sit on a desk; the act of hurting someone is better left to before and after, which is also what keeps a shot from being blocked. One block is not a finding against you; rewrite and try again. Content rules · Blocked by moderation.

## Neighbouring genres

Mafia romance (loyalty and family), vampire romance (the gothic version), demon king dark fantasy, psychological thriller (when nobody can be trusted).

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Gothic Dark, English Manor; Costume Style Power Tailoring, Gothic Court; Visual Style Shot on Film, Black & White Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: two leads, each with one sharp detail on the sheet; the house as two or three locations; the ledger and the key as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Beat | The ending |
| 1 | The deal | She signs |
| 3 | The first rule broken | He sees she has opened the ledger |
| 6 | The secret | Her family's name in his handwriting |
| 9 | The choice | He burns the page |
| 12 | The ending | She stays, on her terms |

## FAQ

### How do I make a dark romance drama with AI?

Give one lead real power and a secret, set it in dark places, write the light into each shot, and keep danger as threat and choice.

### Which scene styles suit dark romance?

Gothic Dark, English Manor and Neon Nights; Gothic Court or Power Tailoring for costumes.

### How do I make scenes look dark if the sheets are evenly lit?

Write the light per shot (one lamp, neon, candles); time and weather come from the script.

### What is not allowed?

Sexual content, sexual violence, anything sexualising minors, and graphic violence.

### Which look suits it?

Shot on Film for warmth and grain, Black & White Film for noir.

### Film or series?

Short Drama in 9:16 for feeds; Film in 16:9 for one long story.

### Can the dangerous lead be a vampire or a demon?

Yes; give them a human state and a revealed state.

Updated October 31, 2026


---

# AI Demon King Drama: Dark Fantasy Courts, Bargains and Curses

URL: https://scenemixer.com/guides/make-an-ai-demon-king-dark-fantasy-drama/
Make a demon king or dark fantasy drama with AI: Dark Asura and Gothic Dark worlds, Dark Path Robes and Gothic Court costumes, demon forms as states.

# How to make an AI demon king or dark fantasy drama

Quick answer: A demon king drama is a bargain: a mortal who needs something, a ruler of the dark who wants something back. In SceneMixer, Step 1 builds the underworld: Dark Asura (black and gold, obsidian, bone totems, fire and lava as the light) for an Eastern demon realm, Gothic Dark for a Western one, with Dark Path Robes or Gothic Court to dress the court. The demon king's true form is a state of the character, revealed when the story needs it.

Brief for AI Create: A demon king drama: a temple girl offers herself to the demon king to end a plague in her village; he accepts and then refuses to touch her, because she is the reincarnation of the queen who sealed him; 10 short episodes, each ends with a crack in the seal.

## Choose the dark realm

| Kind of darkness | Scene Style | Costume Style |
| An Eastern demon realm | Dark Asura | Dark Path Robes |
| A Western gothic hell court | Gothic Dark | Gothic Court |
| Heaven against hell | Celestial Colossus and Dark Asura | Immortal Sect Robes and Dark Path Robes |
| A demon in the modern city | Neon Nights or Urban Opulence | Power Tailoring |
A SceneMixer project has one scene style; for two realms, pick the one where most of the story happens and describe the other in the locations. Scene and costume styles apply at parse · Xianxia and wuxia.

## Horns, forms and curses

- The true form is a state. Horns, black eyes, the fire behind him: one character, two states. States.
- A curse mark that spreads is a state per stage, or a prop if it is an object (a cursed ring).
- Write the light. “Lava light from below, gold edges on black pillars” gives a dark throne room something to see by. Describe the cause.
- Fear without gore. Graphic injuries are blocked; menace lives in the voice, the silence and the size of the room. Content rules.

## The bargain, shot by shot

| Beat | Shot to adapt |
| The hall | [Shot 1] 4s: Wide shot, an obsidian throne hall lit by a river of lava, bone totems along the walls. @Lian kneels on the black floor. |
| The throne | [Shot 2] 3s: Low angle. @Demon King Xuan sits in shadow, gold embroidery catching the lava light. |
| The line | [Shot 3] 4s: Close-up on @Demon King Xuan. @Demon King Xuan (low): Mortals do not offer. They bargain. What do you want? (voice: Xuan) |
| The answer | [Shot 4] 3s: Close-up on @Lian, not looking away. |

## Story shapes that work

- The sacrifice bride.
- The sealed king: the girl who carries the seal.
- Reincarnation: enemies in one life, lovers in the next.
- The demon in the boardroom: a dark fantasy CEO.
- The exorcist and the demon who need each other.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Dark Asura, Gothic Dark; Costume Style Dark Path Robes, Gothic Court; Visual Style Digital HD, Painterly Anime; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the demon king with a true-form state, the mortal with one mark; the throne hall and the village as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The plague village. | She walks into the demon realm alone. |
| 2 | The throne hall. | He accepts, and sends her to a room alone. |
| 3 | The sealed gate. | Her touch makes the seal glow. |
| 4 | The village. | The plague is gone, and so is her name. |
| 5 | The throne hall. | He shows her his true form. |

## FAQ

### How do I make a demon king drama with AI?

Choose Dark Asura or Gothic Dark in SceneMixer's Step 1 with Dark Path Robes or Gothic Court, give the demon king a true-form state in Step 2, and write a light source into every dark shot.

### What does the Dark Asura style look like?

Demon realms in black and gold: matte iron, obsidian, bone totems and dark gilding, with fire and lava as the light.

### How do I show the demon's true form?

Add it as a state of the same character and pick it in the segments where he reveals himself.

### Can it be scary?

Yes, through the room, the voice and the silence. SceneMixer blocks graphic injury and gore.

### Can I mix heaven and the demon realm?

Choose one scene style for the project and describe the other realm in its locations.

### Which visual style suits dark fantasy?

Digital HD or Game Engine Realism for live action, Painterly Anime or Painted Manhua for a drawn look.

### Can I copy a famous demon character?

No. Known characters trigger the IP check. Design your own.

### Vertical or landscape?

9:16 Portrait for short-drama feeds, 16:9 Landscape for a fantasy series.

Updated November 3, 2026


---

# AI Desert Romance: Gulf Cities, Oasis Palaces and Arabic Dialogue

URL: https://scenemixer.com/guides/make-an-ai-desert-sheikh-romance/
Make a desert romance with AI: Gulf Metropolis and Silk Road City scenes, Middle Eastern Robes and Arab Everyday Wear, Target Region Middle East, and Arabic dialogue.

# How to make an AI desert or sheikh romance

Quick answer: A desert romance is a marriage of convenience, a family that expects an heir and a love that arrives late. In SceneMixer, set the world in Step 1: Gulf Metropolis (supertall towers, waterfront avenues, marble and brass) for the modern Gulf, Silk Road City or Frontier Vastness for the desert, Middle Eastern Robes and Arab Everyday Wear for the family, and Target Region Middle East. Arabic is one of the dialogue languages, so the lines can be spoken in Arabic.

Brief for AI Create: A desert romance: an architect hired to restore a ruler's old desert palace is asked to marry his son for one year to settle a family dispute; the son agrees only because he plans to leave the country when the year ends; 12 short episodes, each ends at the palace at dusk.

## The Gulf in Step 1

| Setting | Scene Style | Costume Style |
| A modern Gulf city | Gulf Metropolis | Middle Eastern Robes, Arab Everyday Wear |
| An old desert palace or oasis | Silk Road City | Middle Eastern Robes |
| The open desert | Frontier Vastness | Middle Eastern Robes |
| The old souq | Middle Eastern Souq | Arab Everyday Wear |
Target Region Middle East for faces and names. Target Region.

## Arabic dialogue and respectful detail

Choose Arabic as the dialogue language in Step 1; it is fixed once the script is read. SceneMixer's interface and these guides are also available in Arabic. Dialogue language.

Write family roles, titles and modest dress plainly on each sheet; romance in this genre lives in looks, distance and words, which also keeps you well inside content review. Real ruling families and real people cannot be depicted; invent the family and the country if you need one. Content rules.

## Dusk at the palace, shot by shot

| Beat | Shot to adapt |
| The courtyard | [Shot 1] 4s: Wide shot, a mud-brick palace courtyard at dusk, a stone well, blue-tiled arches. |
| The plans | [Shot 2] 3s: Close-up. @Layla's hand smooths restoration drawings on the well's edge. |
| The arrival | [Shot 3] 3s: Medium shot. @Faisal crosses the courtyard in a white thobe. |
| The line | [Shot 4] 4s: Close-up on @Faisal. @Faisal (dry): My father thinks you can restore anything. Even this family. (voice: Faisal) |

## Story shapes that work

- The marriage of convenience.
- The foreign architect and the old palace.
- The runaway heir.
- Two families, one oasis.
- The business rival who is also the fiancé.
Billionaire romance.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Gulf Metropolis, Silk Road City; Costume Style Middle Eastern Robes, Arab Everyday Wear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the couple and the family with one mark each; Target Region Middle East; the palace courtyard, the office and the souq as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The ruler's office. | A contract for a palace, and a marriage. |
| 2 | The desert palace. | The son refuses to look at the plans. |
| 3 | The wedding. | He tells her he will leave in a year. |
| 4 | The souq. | She finds his old sketchbook of the palace. |
| 5 | The courtyard at dusk. | He asks her to change one wall back. |

## FAQ

### How do I make a desert romance with AI?

Choose Gulf Metropolis or Silk Road City in SceneMixer's Step 1, Middle Eastern Robes or Arab Everyday Wear, Target Region Middle East, and Arabic or another dialogue language.

### Can the characters speak Arabic?

Yes. Arabic is one of SceneMixer's dialogue languages.

### What does the Gulf Metropolis style look like?

Gulf Metropolis on SceneMixer is a Dubai- or Doha-like city: supertall glass towers, wide waterfront avenues, marble-and-brass interiors and strong sun.

### Can I depict a real royal family?

No. SceneMixer cannot depict real people; invent the family.

### How do I get Middle Eastern faces and names?

Set SceneMixer’s Target Region to Middle East.

### Which visual style suits it?

Digital HD or Soft Glow.

### Vertical or landscape?

9:16 Portrait for short-drama feeds.

### Can it be set in the past?

Yes. Silk Road City and Frontier Vastness suit a period desert story.

Updated November 3, 2026


---

# AI Dinosaur Adventure: Expeditions, Scale and Real Animals

URL: https://scenemixer.com/guides/make-an-ai-dinosaur-adventure/
Make a dinosaur adventure with AI: each dinosaur that matters as a character, scale against trees and vehicles, realistic or cartoon looks, and jungles drawn once.

# How to make an AI dinosaur adventure

Quick answer: A dinosaur adventure needs dinosaurs that behave like animals and people who look small next to them. In SceneMixer, make each important dinosaur a character with its own sheet (species, colour, size, a scar), describe its size against a tree or a jeep, set the story in Tropical Rainforest or African Savanna, and choose a live-action or 3D cartoon look depending on the audience.

Brief for AI Create: A dinosaur adventure: a park ranger's twelve-year-old son and a young triceratops get lost together in the valley the night the fences fail; eight episodes, one storm, one long walk home.

## Live action or cartoon

| Audience | Visual Style | Dinosaurs as |
| Teens and adults | Digital HD, Shot on Film | Realistic animals on four or two legs |
| Families | Family 3D Animation | Expressive, rounded, can show feelings |
| Young children | 3D Toy Cartoon, Watercolor Storybook | Friendly, simple shapes, may talk |
Visual styles.

## Dinosaurs as characters

A dinosaur that appears more than once needs a sheet: species, colour pattern, size, one mark (a torn frill, a chipped horn). Herds in the distance are background. Realistic dinosaurs behave as animals; only cartoon ones should talk or use hands. Non-human characters.

Scale is a comparison: its head is level with the treetops, taller than the jeep. Say what is there.

## A first encounter

| Beat | Shot to adapt |
| The ferns | [Shot 1] 3s: Wide shot, rainforest at dawn, mist. @Theo pushes through giant ferns. Ambient: birds, dripping water. |
| The sound | [Shot 2] 2s: Close-up on @Theo freezing. A low snort from somewhere close. |
| The eye | [Shot 3] 3s: Close-up. A large amber eye blinks between the leaves. |
| The friend | [Shot 4] 4s: Medium shot. @Bramble, a young triceratops with a chipped left horn, steps out and sniffs @Theo's backpack. |

## Story shapes

- Lost together: a child and a young dinosaur find the way home.
- The expedition: scientists, a guide, and a valley no one mapped.
- Time travel: a modern character in the age of dinosaurs.
- The educational series: one species, one fact, one adventure per episode. Educational cartoons.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Tropical Rainforest, African Savanna; Costume Style Outdoor Gear; Visual Style Digital HD, Family 3D Animation; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: each named dinosaur as a character with one clear mark; the ranger station, the river and the cave as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Place | The ending beat |
| 1 | The ranger station | The fence alarm goes off |
| 2 | The ferns | A young triceratops follows Theo |
| 3 | The river | Something big drinks upstream |
| 4 | The cave | The storm and the flood |
| 5 | The valley edge | Bramble's herd, and the station lights |

## FAQ

### How do I make a dinosaur adventure with AI?

Make each important dinosaur a character with its own sheet, describe scale against familiar things, and choose a realistic or cartoon look for your audience.

### Will the same dinosaur look the same in every shot?

Yes, if it has a character sheet; give it one clear mark.

### Can dinosaurs talk?

Cartoon dinosaurs can; realistic ones should behave as animals.

### Can I show dinosaurs attacking?

Show the chase and the escape; graphic injury is blocked.

### Which setting works?

Tropical Rainforest, African Savanna or Frozen Wilderness for an ice age.

### Is it suitable for kids?

Yes, with a cartoon look and gentle writing.

### Can I copy a famous dinosaur film?

No. Famous franchises trigger the IP check; dinosaurs themselves are fine.

### Vertical or landscape?

Landscape suits scale; vertical works for short episodes.

Updated October 31, 2026


---

# AI Divorce Revenge Drama: The Glow-Up and the Comeback

URL: https://scenemixer.com/guides/make-an-ai-divorce-revenge-drama/
Make a divorce revenge drama with AI: the before and after as states of one heroine, the divorce papers as a prop, the ex's new life as contrast, and the payoff scene.

# How to make an AI divorce revenge drama

Quick answer: A divorce revenge drama is a comeback: the wife who was dismissed returns as the person no one can ignore. In SceneMixer, draw the heroine's before and after as two states of the same face, make the divorce papers and the wedding ring props, set the ex's world in Urban Opulence, and build each episode toward a public moment where the tables turn.

Brief for AI Create: A divorce revenge drama: on their fifth anniversary, a man hands his wife divorce papers and moves his assistant into the house; a year later she walks into his company's board meeting as the new investor; fifteen episodes.

## Before and after

The glow-up is the genre's promise. Draw the before state (plain clothes, tired light) and the after state (tailored, confident), keeping the face identical so the change is the clothes, the posture and how she is lit. States.

| Stage | Scene Style | Costume Style |
| The marriage | Suburban Mansion | Loungewear, Soft Everyday |
| The fall | A small flat, Nordic Minimal | Soft Everyday |
| The rise | Corporate Cold | Power Tailoring |
| The return | Urban Opulence ballroom | Haute Couture, Minimal Luxury |

## Props that carry the story

The divorce papers, the ring left on the table, the business card she keeps: make them props so they look the same when they return. The ring in episode 1 can be the ring she slides back across the table in episode 15. Same prop in every hand.

## The payoff scene

| Beat | Shot to adapt |
| The room | [Shot 1] 3s: Wide shot, a glass boardroom. @Adrian at the head of the table. |
| The door | [Shot 2] 3s: Medium shot. @Elena in a white tailored suit walks in, after state. |
| Recognition | [Shot 3] 2s: Close-up on @Adrian, the pen stopping. |
| The line | [Shot 4] 4s: @Elena (taking his chair): Let's begin. I believe you know my terms. |
Revenge dramas · Billionaire romance.

## Keeping the audience on her side

- Show the injustice clearly in episode 1; the comeback is only as good as the fall.
- Give her small wins every episode and a big one every five.
- Let the ex's new partner be a person, not a cartoon; the best revenge is not needing it.
- End episodes on her face, not his.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Corporate Cold; Costume Style Power Tailoring, Haute Couture; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: a before and an after state for the heroine; the ring and the papers as props; the mansion, the flat and the boardroom as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Her move | The ending beat |
| 1 | The papers | She signs without a word |
| 3 | A job interview | The interviewer knows her old work |
| 6 | The first deal | Her ex hears her name |
| 10 | The gala | She arrives on someone else's arm |
| 15 | The boardroom | She takes his chair |

## FAQ

### How do I make a divorce revenge drama with AI?

Draw the heroine's before and after as states of one face, make the divorce papers and ring props, and build to public turning points.

### How do I show the glow-up?

A second state of the same character with new clothes, hair and light; switch to it from the turning point on.

### Which format suits it?

Short Drama, 9:16 Portrait, one to two minutes an episode.

### Which look suits it?

Digital HD for the glossy modern feel.

### How many episodes?

Fifteen to thirty short episodes is common for the genre.

### Can I reuse the cast for a sequel?

Yes, from your SceneMixer library in a new project.

### Can the dialogue be in another language?

Yes; set the dialogue language and region in Step 1.

### Can there be a new love interest?

Yes; give him his own sheet and keep him out of the first episodes.

Updated November 3, 2026


---

# AI Documentary-Style Video: Narrator, Interviews and B-Roll

URL: https://scenemixer.com/guides/make-an-ai-documentary-style-video/
Make a documentary-style video with AI: a narrator over B-roll, people talking to camera, nature and history documentaries, and the limits on real people.

# How to make an AI documentary-style video

Quick answer: A documentary-style video is narration over pictures, a few people talking to camera, and B-roll that shows what they describe. SceneMixer makes it as a Film project: the narrator is a voice-only character, interviewees are characters seated to camera, and B-roll is shots without dialogue. Nature documentaries work with real animals in Digital HD; history documentaries use invented people, because real historical and public figures cannot be depicted.

Brief for AI Create: A nature documentary: one year in the life of a red fox family on the edge of a northern town, from the first snow to the cubs leaving; a calm narrator, eight minutes.

## Three kinds of shot

| Kind | How to write it |
| Narration over B-roll | Shots with no dialogue; the narrator's voice-over line under them |
| Interview | A person seated, medium close-up, looking just off camera, speaking |
| Establishing | A wide shot of the place, held for four or five seconds |

Voice-over lines · Voice-only characters.

## Nature

Real animals are drawn as animals, on four legs, no clothes; give each named animal a sheet (a young red fox, thin, a white tip on the tail) so the same fox comes back. Frozen Wilderness, Tropical Rainforest and African Savanna cover the classic settings. Animals as characters.

Scene Style: Frozen Wilderness

Scene Style: Tropical Rainforest

Scene Style: African Savanna
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## History

Real historical and public figures cannot be depicted, so tell history through invented witnesses (a soldier, a clerk, a child in the crowd) and the places themselves. Period settings are built in: WWII, history drama, medieval. Real people and the IP check.

## One look per project

The visual style is fixed once you continue, so a documentary that cuts between colour interviews and black-and-white archive needs two projects, joined in your editor. Black & White Film for the archive, Digital HD or Shot on Film for the present. Captions and titles are added afterwards. Visual style.

## Neighbours

Mockumentary (the comedy version), true crime style, training video.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Frozen Wilderness, American Small Town; Costume Style Outdoor Gear, Soft Everyday; Visual Style Digital HD, Black & White Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the narrator as a voice-only character; each named animal or witness as a character; the main places as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Minute | Part |
| 0–1 | Establishing: the town edge in snow, narration |
| 1–3 | Winter: hunting, the den |
| 3–5 | Spring: the cubs |
| 5–7 | Summer: the road |
| 7–8 | Autumn: leaving |

## FAQ

### How do I make a documentary-style video with AI?

Make a SceneMixer Film project with a voice-only narrator, B-roll shots without dialogue and interviewees framed to camera.

### Can I make a nature documentary?

Yes. Real animals are drawn as animals; give each named animal a sheet.

### Can I make a documentary about a real person?

Real public and historical figures cannot be depicted; use invented witnesses.

### Can I mix black-and-white archive with colour?

Not in one project; make two and join them in your editor.

### Can I add titles and captions?

Add them after export; generated video does not spell reliably.

### How long can it be?

As long as you need, built from segments and joined.

### Which aspect ratio?

16:9 Landscape for YouTube and screens.

Updated November 3, 2026


---

# AI Dragon Fantasy Drama: Dragon Riders, Dragon Kings and Castles

URL: https://scenemixer.com/guides/make-an-ai-dragon-fantasy-drama/
Make a dragon fantasy series with AI: the dragon as a character with its own sheet, Medieval Castle and Epic Scale worlds, shifter forms as states, battles in cuts.

# How to make an AI dragon fantasy drama

Quick answer: In a dragon drama the dragon is a character, not an effect. Give it a card in Step 2 with its own sheet, colour and one mark you will recognise in every shot (a torn wing, a scar across the jaw), and it stays the same dragon for the whole series. Build the world in Step 1: Medieval Castle or Epic Scale scenes, Medieval Nobility, Plate Armor or Fantasy Adventurer costumes. A dragon king who takes human form is one character with two states.

Brief for AI Create: A dragon fantasy drama: a stable girl bonds with the last dragon egg the night the king orders all dragons destroyed; the king's exiled brother is the dragon's true rider and wants it back; 10 episodes, each ends with the dragon growing and someone noticing.

## Choose the kind of dragon world

| Kind of story | Scene Style | Costume Style |
| Knights, castles and a dragon in the hills | Medieval Castle | Medieval Nobility, Plate Armor |
| Wars between kingdoms | Epic Scale | Plate Armor, Fantasy Adventurer |
| A dragon-shifter king in a dark court | Gothic Dark or European Palace | Gothic Court, European Court Gown |
| Eastern dragons and immortal sects | Celestial Colossus | Immortal Sect Robes |
| Dragons in the far north | Frozen Wilderness or Viking Fjord | Nordic Viking |
Every scene and costume style · Xianxia and wuxia dramas.

## The dragon as a character

- Describe what it is, not how impressive it is. “A bronze-scaled dragon the size of a barn, a torn left wing, smoke at the nostrils” beats “majestic and terrifying”. Describe the cause, not the result.
- Growing up is a state. Hatchling, young dragon and full-grown dragon are three pictures of one character; add them as states and pick the right one per segment.
- A shifter is two states. The king in human form and the king as a dragon share one card and one voice. States.
- It can speak. A talking dragon gets a voice from the preset library like anyone else. Non-human characters.

## Flight and fire, shot by shot

Big moments land better as several short shots than as one long one: the shadow first, then the wing, then the face.

| Beat | Shot to adapt |
| The shadow | [Shot 1] 3s: Wide shot, castle courtyard at dusk. A huge shadow passes over the stones and the soldiers look up. |
| The landing | [Shot 2] 4s: Low angle. @Ember lands on the battlements, torn left wing folding, smoke curling from her nostrils. |
| The rider | [Shot 3] 3s: Medium shot. @Wren climbs down from the saddle, hands shaking. |
| The line | [Shot 4] 4s: Close-up on @Aldric. @Aldric (cold): So the egg did hatch. (voice: Aldric) |
How to write one shot. Battles: one blow or one reaction per shot, and stay off the wounds. Fights and crowds.

## Story shapes that work

- The last egg: a forbidden dragon raised in secret.
- The dragon king's bride: a bargain marriage to a ruler who is not quite human.
- The rider academy: rivals, trials and a dragon that picks the wrong student.
- The war: two kingdoms, one with dragons and one with a weapon against them.
- The hidden heir: only the true heir can command the dragon.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Medieval Castle, Epic Scale; Costume Style Medieval Nobility, Plate Armor; Visual Style Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the dragon as a character with its own sheet and a state for each age; riders and royals with one mark each; the castle, the hills and the hall as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The stables at night. | The egg cracks in the straw. |
| 2 | The castle kitchens. | The hatchling eats a whole ham and is seen. |
| 3 | The king's hall. | An order: every dragon to be killed. |
| 4 | The hills. | A stranger whistles and the dragon answers him. |
| 5 | The battlements. | She flies for the first time, in front of everyone. |

## FAQ

### How do I make a dragon fantasy drama with AI?

Pick a scene style such as Medieval Castle or Epic Scale in SceneMixer's Step 1, create the dragon as a character with its own sheet in Step 2, and write flight and fire as short shots.

### Will the dragon look the same in every episode?

Yes, if it is a character with a sheet. Give it one clear mark, such as a torn wing, so it reads at a glance.

### How do I show a dragon growing up?

Add the hatchling, young and grown dragon as states of the same character and pick the right one in each segment.

### How do I make a dragon shifter?

One character, two states: the human form and the dragon form. Both share the same voice.

### Can the dragon talk?

Yes. Give it a voice from the preset library like any other character.

### Can I show dragon fire burning people?

Show the fire, the light and the aftermath, not burning bodies. SceneMixer's content review blocks graphic injury.

### Which visual style suits dragons?

Digital HD or Shot on Film for live action; Realistic 3D, Game Engine Realism or Painterly Anime for a drawn epic.

### Can I copy dragons from a famous show?

No. Names, designs and iconic looks from known films and books trigger the IP check. Design your own dragon.

Updated November 3, 2026


---

# Make an AI Educational Cartoon for Kids: Science and Stories

URL: https://scenemixer.com/guides/make-an-ai-educational-cartoon-for-kids/
Make an educational cartoon for children with AI: a recurring guide character, one idea per episode, 3D Toy Cartoon or Watercolor Storybook, and short, clear lines.

# How to make an AI educational cartoon for kids

Quick answer: An educational cartoon teaches one idea per episode through a character children want to follow. In SceneMixer, create a guide (a curious robot, an owl, a girl with a magnifying glass) and a sidekick, draw them once in a child-friendly style (3D Toy Cartoon, Family 3D Animation or Watercolor Storybook), and write each episode as a small story that ends with the idea said plainly. You check the facts; the cartoon carries them.

Brief for AI Create: An educational cartoon for ages 5 to 8: a curious robot named Bolt and a sleepy owl investigate one question per episode; episode 1: why the moon changes shape; three minutes, one experiment with a torch and a ball, one sentence to remember.

## One idea, one story

- Ask a question children ask: why is the sky blue, where does rain go, how do bees talk.
- Show an experiment with everyday props (a torch, a ball, a glass of water); props get reference pictures. Props.
- Say the idea plainly at the end, in one sentence the guide repeats. Dialogue.
- Check the facts yourself. The script is what the characters say; get it right before you render.

## Characters and look

- A guide and a sidekick, drawn once. A robot or an animal works well. Non-human characters.
- Style: 3D Toy Cartoon for preschool, Family 3D Animation for older kids, Watercolor Storybook for nature topics. Visual styles.
- A recurring lab or treehouse as a location. The same room.
- Subtitles help early readers (accounts with a purchase). Subtitles.

## The moon, shot by shot

| Beat | Shot to adapt |
| The question | [Shot 1] 4s: Medium shot, a treehouse at night. @Bolt, a small round robot, points at a half moon through the window. @Bolt (curious): Who keeps cutting the moon? (voice: Bolt) |
| The experiment | [Shot 2] 4s: Medium shot. @Hoot the owl holds a torch; @Bolt holds a white ball at arm's length. |
| The result | [Shot 3] 3s: Close-up on the ball, half lit and half dark. |
| The idea | [Shot 4] 4s: Close-up on @Hoot. @Hoot (sleepy): The moon is always whole. We just see the sunny part. (voice: Hoot) |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Pastoral Healing; Costume Style Soft Everyday; Visual Style 3D Toy Cartoon, Watercolor Storybook; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the guide and the sidekick with one mark and a voice each; experiment objects as props; the treehouse or lab as a location. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Question | The one sentence |
| 1 | Why does the moon change shape? | We see the sunny part. |
| 2 | Where does rain come from? | Water goes up, then comes down. |
| 3 | Why do leaves change colour? | Green fades and other colours show. |
| 4 | How do bees find flowers? | They dance directions. |
| 5 | Why is the sea salty? | Rivers carry salt from rocks. |

## FAQ

### How do I make an educational cartoon with AI?

Create a guide character and a sidekick in a child-friendly style, write one question per episode, show a simple experiment, and end with one clear sentence.

### Who checks the facts?

You do. The characters say exactly what your script says.

### Which style suits young children?

3D Toy Cartoon or Watercolor Storybook.

### Can a robot or an animal be the teacher?

Yes. Non-human characters have sheets and voices.

### Can I add subtitles?

Yes, on SceneMixer accounts that have made a purchase.

### Can it be in another language?

Yes. In SceneMixer, choose the dialogue language in Step 1.

### How long should an episode be?

Two to five minutes, one idea each.

### Do I need to label AI content?

Follow each platform's rules; see our answer on AI labels.

Updated November 3, 2026


---

# AI Elf and Fae Fantasy Drama: Forest Kingdoms and Fae Bargains

URL: https://scenemixer.com/guides/make-an-ai-elf-or-fae-fantasy-drama/
Make an elf or fae fantasy romance with AI: the Elven Forest scene style, Elven Attire, pointed ears and glamour as states, and a bargain that cannot be undone.

# How to make an AI elf or fae fantasy drama

Quick answer: An elf or fae drama is a mortal in a kingdom that does not follow human rules: a promise is binding, a name is power, a gift is a debt. SceneMixer has the forest built in: Elven Forest (a kingdom grown into giant trees, rope bridges in the canopy, glowing fungi, light through mist) with Elven Attire (slim lines, leaf and vine patterns, silver with green). A fae prince disguised as a human is one character with two states.

Brief for AI Create: A fae fantasy romance: a botanist who follows a lost child into the woods ends up in the fae court, where eating one berry has bound her for a year to the prince who found her; 10 episodes, each ends with a bargain she did not mean to make.

## Build the realm

| Kind of fae story | Scene Style | Costume Style |
| An elven kingdom in the trees | Elven Forest | Elven Attire |
| A dark fae court | Gothic Dark | Gothic Court |
| Fae among humans | Pastoral Healing or American Small Town | Soft Everyday |
| Elves, dwarves and orcs at war | Epic Scale | Elven Attire, Dwarven Forge, Orc Tribal as characters' clothing |
Dragon fantasy · Mermaid fantasy.

## Ears, wings and glamour

- Put the fae features on the sheet: pointed ears, silver eyes, wings if they have them. Describe a face.
- A glamour is a state. The prince as a human hiker and as a fae prince. States.
- Magic is light. “A ring of pale mushrooms lights up one by one”. Describe the cause.
- Bargains are lines. Write the exact words of every promise; the audience will hold them to it. Dialogue.

## The first bargain, shot by shot

| Beat | Shot to adapt |
| The canopy | [Shot 1] 4s: Wide shot, a forest kingdom high in giant trees, rope bridges, glowing fungi, mist. @June stands on a platform in a muddy raincoat. |
| The berry | [Shot 2] 3s: Close-up on a silver berry in @June's palm, half eaten. |
| The prince | [Shot 3] 3s: Medium shot. @Prince Ilvar, pointed ears, silver-green coat, watches her from a branch. |
| The line | [Shot 4] 4s: Close-up on @Prince Ilvar. @Prince Ilvar (amused): You have eaten at my table. You are mine for a year. (voice: Ilvar) |

## Story shapes that work

- The mortal bound by a bargain.
- The changeling: a child swapped at birth.
- The fae prince exiled among humans.
- The war between courts.
- The forest healer and the dying king.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Elven Forest; Costume Style Elven Attire; Visual Style Soft Glow, Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the mortal and the fae with features on their sheets and glamour states; the canopy court and the woods as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The woods. | She follows a child through a ring of mushrooms. |
| 2 | The canopy court. | She eats one berry. |
| 3 | The prince's hall. | A year, and no lies allowed. |
| 4 | The river. | The lost child is the prince's brother. |
| 5 | The court feast. | She makes a second bargain to save him. |

## FAQ

### How do I make an elf or fae drama with AI?

Choose Elven Forest and Elven Attire in SceneMixer's Step 1, put the fae features on each sheet, and give disguised characters a human state and a fae state.

### What does the Elven Forest style look like?

A forest kingdom grown into giant trees: timber platforms and rope bridges in the canopy, vines and glowing fungi, light spots through mist.

### How do I show pointed ears and wings?

Write them on the character's sheet description so they appear in every shot.

### How do I show a glamour?

Two states of one character: the human disguise and the fae form.

### Can I add dwarves and orcs?

Yes. Describe them as characters; Dwarven Forge and Orc Tribal describe their clothing.

### Which visual style suits fae fantasy?

Soft Glow or Digital HD; Watercolor Anime or Painterly Anime for a drawn look.

### Can I copy a famous elf saga?

No. Known names and designs trigger the IP check.

### Vertical or landscape?

9:16 Portrait for romance feeds, 16:9 Landscape for a fantasy series.

Updated November 3, 2026


---

# AI Enemies to Lovers Romance: Rivals, Banter and the Turn

URL: https://scenemixer.com/guides/make-an-ai-enemies-to-lovers-romance/
Make an enemies to lovers romance with AI: two rivals staged as mirrors, banter timed in shot seconds, a forced-proximity setting and the episode where it turns.

# How to make an AI enemies to lovers romance

Quick answer: Enemies to lovers runs on friction: two equals who cannot stop arguing until they cannot stop looking. In SceneMixer, stage the rivals as mirrors (same framing, opposite sides of the frame), time the banter in the shot seconds so each comeback lands, trap them together in one location, and keep the turn for the midpoint.

Brief for AI Create: An enemies to lovers romance: two rival wedding planners are hired by the same couple and forced to plan one wedding together; ten episodes, one country house, one storm that cuts the power.

## Mirror the rivals

Give them equal weight: matched framing, opposite sides of the frame, the same number of close-ups. Positions are tracked across cuts, so she stays left and he stays right until the story moves them. Everyone stays put · Why characters swap sides.

| Rivals | Scene Style | Costume Style |
| Office rivals | Corporate Cold | Power Tailoring |
| Fashion rivals | Parisian Streets | Haute Couture |
| School rivals | Elite Boarding School | Prep School Uniform |
| Royal rivals | European Palace | European Court Gown |

## Banter is timing

A comeback needs its own shot and a beat before it. Write the line, then the reaction, then the reply, each with seconds. The sample L’Héritier du Défilé plays a rivalry in a fashion studio. Shot seconds and timing.

| Beat | Shot to adapt |
| The jab | [Shot 1] 3s: Medium shot, a marquee half set up. @Clara: Peonies. In October. Bold choice. |
| The beat | [Shot 2] 2s: Close-up on @Jonah, unimpressed. No dialogue. |
| The reply | [Shot 3] 3s: @Jonah: Says the woman who booked a string quartet for a beach. |
| The crack | [Shot 4] 2s: Close-up on @Clara trying not to smile. |

## Forced proximity

One location they cannot leave: a country house, a delayed train, a shared office, a snowed-in lodge. It is cheaper too: one set drawn once, used in every episode. Same room in every episode.

## The turn

- Midpoint: one sees the other being kind when they think no one is watching.
- The almost: interrupted at the worst moment.
- The setback: the old rivalry returns with a reason.
- The admission: said badly, which is how it should be said.
Hooks and cliffhangers.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Corporate Cold, English Manor; Costume Style Power Tailoring, Smart Casual; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: two leads with specific faces and equal weight; one shared location; a storm set on the midpoint segments. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Friction | The ending beat |
| 1 | Same client, same day | They have to share the venue |
| 3 | Sabotaged flowers | Neither of them did it |
| 5 | The storm and the power cut | A candle and an honest answer |
| 7 | The old rivalry | One of them takes another job |
| 10 | The wedding | Not theirs, yet |

## FAQ

### How do I make an enemies to lovers romance with AI?

Stage the rivals as mirrors, time the banter in shot seconds, trap them in one location and save the turn for the midpoint.

### How do I make banter land?

Give each line and each reaction its own short shot.

### Will the two leads stay consistent?

Yes. Each has one sheet used in every shot.

### Which format suits it?

Short Drama in 9:16 for feeds, or TV Series for slower episodes.

### Which look suits it?

Digital HD for modern, Soft Glow for period, Korean Webcomic for a webtoon feel.

### Can there be kissing?

Romance is fine; explicit content is not.

### How many episodes?

Eight to twenty short episodes is typical.

### Can I set it in a palace?

Yes: European Palace and European Court Gown.

Updated October 31, 2026


---

# AI Fable Video: Make an Animal Fable Cartoon With a Moral

URL: https://scenemixer.com/guides/make-an-ai-fable-cartoon/
Make an animal fable cartoon with AI: the five parts of a fable, 16 original fables with morals, a worked storyboard, a gentle look and a narrator to say the moral.

# How to make an AI fable cartoon

Quick answer: A fable is a very short story in which animals with one trait each make a choice and live with the result, and it ends on one line of moral. In SceneMixer, write it in five parts (cast, wish, choice, consequence, moral), draw each animal once with one mark and a voice, pick a gentle look (3D Toy Cartoon, Claymation or Watercolor Storybook) and let an Off-screen narrator say the moral last. Sixty to 150 seconds is enough. The sixteen fables below are original, so the characters and designs are yours.

Brief for AI Create: A two-minute animal fable for ages 5 to 8: a fox who wants the whole meadow to listen to him buys a bigger and bigger drum until nobody comes to hear him; a quiet mouse is listened to without one; a warm narrator, one meadow, the moral said once at the end.

## The five parts of a fable

| Part | Its job | Seconds in a 90 s fable | The Fox and the Louder Drum |
| 1. Cast | Two or three animals, one trait each | 0 to 15 | Fox, who is vain; Mouse, who is quiet |
| 2. Wish | What the main animal wants | 15 to 30 | Fox wants the whole meadow to listen |
| 3. Choice | What they do about it | 30 to 55 | He buys a bigger drum, then a bigger one |
| 4. Consequence | What the choice costs | 55 to 80 | The meadow covers its ears; no one comes to his stories |
| 5. Moral | One plain line | 80 to 90 | The loudest voice is not the one that is heard |

## Sixteen original fables

Each has a cast of two or three, one place and a moral you can say in one breath. Change the animals and the place and they are yours.

| Fable | Animals | The choice | Moral |
| The Fox and the Louder Drum | Fox, mouse | Buys ever bigger drums to be heard | The loudest voice is not the one that is heard. |
| The Heron Who Waited for Clear Water | Heron, frogs | Refuses every pond that is not perfect | Waiting for perfect means missing good. |
| The Badger's Borrowed Boots | Badger, hedgehog | Borrows boots and never returns them | Trust is borrowed too, and it wears out. |
| The Sparrow Who Followed the Wind | Sparrow, crows | Agrees with whichever bird is loudest | If you agree with everyone, no one knows you. |
| The Owl Who Counted Stars Aloud | Owl, chicks | Counts stars through her chicks' first flight | Some things can only be seen now. |
| The Mole and the Window | Mole, squirrel | Is so proud of his tunnels he never looks up | Pride in one thing can hide the rest. |
| The Beaver's Last Dam | Beaver, otter | Builds a dam so strong it floods the neighbours | A wall for you can be a flood for someone else. |
| The Magpie's Shiny Pebble | Magpie, crow | Trades a friend's gift for a shinier pebble | Shiny is not the same as valuable. |
| The Rabbit Who Said Yes | Rabbit, tortoise | Promises the same afternoon to six friends | A yes to everyone is a no to someone. |
| The Glow-worm and the Lantern | Glow-worm, lamplighter mouse | Envies the lantern and goes dark trying to match it | Your small light is enough for your own path. |
| The Ox and the Bell | Ox, hen | Ploughs by the bell and never checks the field | A bell tells the time, not what is ready. |
| The Bee Who Kept the Flowers Secret | Bee, hive | Finds a meadow and tells no one | A secret found is only half found. |
| The Lizard's Longer Tail | Lizard, hawk | Glues on a grander tail and cannot run | Pretending to be bigger makes you slower. |
| The Mouse and the Dry Day | Mouse, voles | Mends the roof while the others sunbathe | Prepare while it is easy. |
| The Swan on the Path | Swan, ducklings | Laughs at the waddling ducklings, then crosses land | Everyone is clumsy somewhere; be kind about it. |
| The Wolf Who Practised Sorry | Wolf, sheep | Learns to apologise beautifully and changes nothing | A sorry that changes nothing is only a sentence. |

## Draw the animals so the moral can be funny

- One trait, one mark. A drum-loving fox with a red scarf; a mouse with a patched ear. The mark is what keeps each animal recognisable from shot to shot.
- Say what they have. Standing upright, hands, clothes, a bag? A real fox on four legs and a fox in a waistcoat are different sheets. Non-human characters.
- Two or three animals, one place. A meadow drawn once is steady and cheap to reuse.
- A voice each. A loud, plummy voice for the fox, a small clear one for the mouse. Character voices · Giving a character an accent.

## Who says the moral

| Option | How it plays | Use it when |
| The narrator, last | One plain sentence over the final image | Ages 4 to 8; the simplest and the clearest |
| A character realises | The mouse says it, quietly, to the fox | The moral is a lesson one animal teaches another |
| Show it, say nothing | The final image carries it | Older children, or when the picture is enough |
Whichever you choose, say it once, in the last 5 to 8 seconds, after the consequence has been shown. A moral said before the story is a lecture.

## The Fox and the Louder Drum, shot by shot

| Beat | Shot |
| The meadow | [Shot 1] 5s: Wide shot, slow push-in. A sunny meadow, a flat rock in the middle. @Narrator (voice-over, warm and amused): Fox loved to be heard. Ambient: birdsong. |
| The wish | [Shot 2] 4s: Medium shot. @Fox in a red scarf beats a small drum on the rock and looks around. @Fox (proud): Everybody listen to me! (voice: Fox) |
| The mouse | [Shot 3] 3s: Close-up. @Mouse, patched ear tilted, sitting very still on a clover leaf. No dialogue. Ambient: a faint drum, far off. |
| The bigger drum | [Shot 4] 4s: Wide shot. @Fox wheels a huge drum onto the rock with both paws, the meadow animals walking away behind him. Ambient: a loud, hollow boom. |
| The empty rock | [Shot 5] 3s: Wide shot. The meadow empty, @Fox alone beside his drum, the sunlight long. Ambient: one distant bird. |
| The moral | [Shot 6] 5s: Medium shot. @Mouse speaks softly to a circle of animals gathered close around her. @Narrator (voice-over, gentle): The loudest voice is not the one that is heard. |
That is 24 seconds of a 90-second fable; the middle two minutes repeat shot 4 with a bigger drum each time. How to write one shot.

## Make it a series

A fable a week is an easy channel: one cast of recurring animals in one meadow, a new moral each time. Draw the meadow and the animals once and reuse them. Reuse a cast across projects · A kids cartoon channel · Fairy tale videos, with public-domain tales · Talking animal videos · AI cartoon story generator.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Pastoral Healing; Costume Style Soft Everyday; Visual Style 3D Toy Cartoon, Watercolor Storybook; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the fox, the mouse and the meadow rock; the narrator is an Off-screen character with a voice and no drawing. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Length | Segments of 15 s | Segments of 30 s | Fits |
| 60 seconds | 4 | 2 | A very short fable |
| 90 seconds | 6 | 3 | The shape above |
| 2 minutes | 8 | 4 | A fable with a second try |
| 2.5 minutes | 10 | 5 | A fable with a twist at the end |

## FAQ

### How do I make a fable cartoon with AI?

Write a fable in five parts (cast, wish, choice, consequence, moral), draw each animal once, choose a gentle look and add an off-screen narrator to say the moral.

### What makes a fable different from a story?

It is very short, uses animals with one trait each, and ends on one line of moral.

### Can the animals talk?

Yes. Animals are characters with their own sheet and voice from the library.

### Which look suits a fable?

3D Toy Cartoon for ages 3 to 6, Claymation for a handmade feel, Watercolor Storybook for a picture-book fable.

### Can I retell Aesop's fables?

The classic fables are very old and widely retold. The designs of famous film versions are another matter and trigger the IP check; the guide to fairy tale videos covers it. The sixteen on this page are original.

### How long should a fable video be?

60 to 150 seconds. Longer fables need a second choice or a twist.

### Where does the moral go?

In the last five to eight seconds, said once by the narrator or by a character.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; a 90-second fable is 90 seconds of that.

Updated November 3, 2026


---

# AI Family Inheritance Drama: Wills, Heirs and Dinner-Table Wars

URL: https://scenemixer.com/guides/make-an-ai-family-inheritance-drama/
Make a family drama about a will or an inheritance with AI: Urban Opulence and Suburban Mansion homes, the will as a prop, dinner-table scenes and one secret heir.

# How to make an AI family inheritance drama

Quick answer: A family inheritance drama is a will, a dinner table and too many people with a claim. In SceneMixer, build the family's world in Step 1: Urban Opulence for old money in the city, Suburban Mansion for the house everyone grew up in, Corporate Cold for the company they fight over. Minimal Luxury and Power Tailoring dress them. Make the will, the ring or the deed a prop with its own picture, and play every reading of it at the same table.

Brief for AI Create: A family inheritance drama: the patriarch of a hotel empire dies and leaves everything to a waitress nobody has met; his three children have 30 days to contest the will, and she has 30 days to find out why he chose her; 10 short episodes, each ends at the family dinner table.

## Set the family

| Kind of family | Scene Style | Costume Style |
| Old money in the city | Urban Opulence | Minimal Luxury, Haute Couture |
| The family home | Suburban Mansion | Smart Casual |
| The family company | Corporate Cold | Power Tailoring |
| A country estate | English Manor | Minimal Luxury |
| A Gulf business family | Gulf Metropolis | Middle Eastern Robes or Arab Everyday Wear |
Billionaire romance · Heiress dramas.

## The table is the stage

Family drama happens at the dinner table, so make the dining room a location and use it again and again; the audience learns who sits where. Six people at a table is a group scene: two or three per shot, cut between them. Group scenes.

| Beat | Shot to adapt |
| The table | [Shot 1] 4s: Wide shot, a long dining table under a chandelier, one empty chair at the head. The family eats in silence. |
| The arrival | [Shot 2] 3s: Medium shot. @Rosa stands in the doorway in a waitress's black dress. |
| The reaction | [Shot 3] 3s: Close-up on @Victoria setting down her glass. |
| The line | [Shot 4] 4s: Close-up on @Victoria. @Victoria (icy): You're sitting in my father's chair. (voice: Victoria) |

## Props, ranks and secret heirs

- The will and the ring are props with a reference picture. Props.
- The rags-to-riches turn is a state: the waitress's uniform and the gown she wears in episode 9. States.
- Flashbacks to the father when he was young are another character or a state with a younger face. Age and looks.

## Story shapes that work

- The stranger in the will.
- Three siblings, one company.
- The second family: the father's secret children arrive at the funeral.
- The prodigal son: the disowned heir comes back richer than all of them.
- The fake will: someone at the table forged it.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Suburban Mansion; Costume Style Minimal Luxury, Power Tailoring; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: every family member with one mark, the heir with a before-and-after state; the will and the ring as props; the dining room and the study as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The funeral. | The lawyer reads a name nobody knows. |
| 2 | The family dinner. | She takes the chair at the head of the table. |
| 3 | The company board. | The eldest son calls for a vote. |
| 4 | The father's study. | A photo of her mother on his desk. |
| 5 | The dinner table. | The youngest sister changes sides. |

## FAQ

### How do I make a family inheritance drama with AI?

Choose Urban Opulence, Suburban Mansion or Corporate Cold in SceneMixer's Step 1, make the will and the family heirlooms props, and set the key scenes at one recurring dinner table.

### How do I film a whole family at dinner?

Show two or three people per shot and cut between them; keep the table and the chandelier as a location used in every episode.

### How do I keep the will the same?

Make it a prop with a reference picture in Step 2.

### Can a poor character dress rich later?

Yes. Add the new look as a state of that character and switch to it from the episode where it changes.

### Which costume style suits old money?

Minimal Luxury for quiet wealth, Haute Couture for display, Power Tailoring for the boardroom.

### Can I show flashbacks?

Yes. A younger version is a separate state or character with a younger face.

### Vertical or landscape?

9:16 Portrait suits short-drama feeds; family sagas also work in 16:9 Landscape.

### Which visual style?

Digital HD for modern families, Soft Glow for warmth, Shot on Film for a period saga.

Updated November 3, 2026


---

# AI Filipino Teleserye: Family Secrets, Switched Babies and Big Love

URL: https://scenemixer.com/guides/make-an-ai-filipino-teleserye/
Make a Filipino teleserye-style drama with AI: Southeast Asian Town and Island Resort scenes, Target Region Southeast Asia, and dialogue in English or another language.

# How to make an AI Filipino teleserye-style drama

Quick answer: A teleserye runs on family: switched babies, a rich family and a poor one, a love that crosses the line between them. In SceneMixer, set the world in Step 1 with Southeast Asian Town (shophouses, coloured shutters, palms, stalls and motorbikes), Island Resort for the province by the sea and Urban Opulence for the rich family, and Target Region Southeast Asia for faces and names. Tagalog is not one of the dialogue languages, so lines are spoken in English or another available language.

Brief for AI Create: A teleserye: two girls born the same night in a provincial hospital were switched by a nurse; twenty years later one is a fisherman's daughter, the other the heiress of a resort empire, and they fall for the same man; 12 short episodes, each ends with the nurse getting closer to confessing.

## Two families in Step 1

| Family | Scene Style | Costume Style |
| The poor family in the province | Southeast Asian Town or Island Resort | Soft Everyday |
| The rich family in the city | Urban Opulence | Minimal Luxury |
| The family business | Corporate Cold | Power Tailoring |
A SceneMixer project has one scene style: choose the side of the story you spend most time on and build the other through its locations. Styles apply at parse.

## Dialogue language, honestly

Tagalog is not one of the dialogue languages. Teleserye-style creators on global feeds mostly use English; you can add subtitles. Names, places and endearments (“Anak”, “Lola”) can still be written into the script exactly as you want them said. Dialogue language · Subtitles.

## The confrontation, shot by shot

| Beat | Shot to adapt |
| The resort | [Shot 1] 4s: Wide shot, a beach resort restaurant at sunset, white tablecloths. @Isabel in silk sits with her mother @Doña Carmen. |
| The waitress | [Shot 2] 3s: Medium shot. @Rosa sets down a plate and freezes. |
| The resemblance | [Shot 3] 3s: Close-up on @Doña Carmen looking from Rosa's face to Isabel's. |
| The line | [Shot 4] 4s: Close-up on @Doña Carmen. @Doña Carmen (shaken): What is your mother's name, iha? (voice: Carmen) |
Slaps and fights: the moment before and after. Fights.

## Story shapes that work

- Switched at birth.
- The mistress and the wife.
- The prodigal daughter returns from abroad.
- Two brothers, one inheritance.
- The amnesia that hides the truth.
Telenovelas and soap operas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Southeast Asian Town, Island Resort; Costume Style Soft Everyday, Minimal Luxury; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: both families with one mark each; Target Region Southeast Asia; the resort, the village and the hospital as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The provincial hospital, twenty years ago. | A nurse swaps two name bands. |
| 2 | The fishing village. | Rosa gets a job at the resort. |
| 3 | The resort restaurant. | Doña Carmen sees Rosa's face. |
| 4 | The village church. | The nurse lights a candle every year. |
| 5 | The resort's anniversary party. | Both girls wear the same birthmark. |

## FAQ

### How do I make a teleserye-style drama with AI?

Choose Southeast Asian Town, Island Resort or Urban Opulence in SceneMixer's Step 1, set Target Region to Southeast Asia, and build the story around two families.

### Can the characters speak Tagalog?

No. Tagalog is not one of SceneMixer's dialogue languages. Use English or another available language, with subtitles if needed.

### How do I get Filipino faces and names?

Set SceneMixer’s Target Region to Southeast Asia, or write the names into the script and choose Follow script.

### How do I show a birthmark both girls share?

Write it as one mark on each character's sheet.

### Which visual style suits a teleserye?

Digital HD or Soft Glow.

### Can I copy a famous teleserye?

No. The IP check on SceneMixer catches known shows and actors.

### Vertical or landscape?

9:16 Portrait for short-drama feeds.

### How long should the episodes be?

One to two minutes with one reveal or near-reveal each.

Updated November 3, 2026


---

# AI Folk Horror Drama: Old Houses, Rituals and a Village That Smiles

URL: https://scenemixer.com/guides/make-an-ai-folk-horror-drama/
Folk horror with AI: old family halls, village rituals and tradition turned on a newcomer. Which scene, costume and visual styles to pick, and how to keep fear off gore.

# How to make an AI folk horror drama

Quick answer: Folk horror is fear that comes from a place and its customs: an old house with rules, a village that is very welcoming, a ritual everyone knows and nobody explains. In Step 1 the scene style does most of the work. Chinese Folk Occult gives old family halls, Taoist temples, mourning halls and ancient tombs under dark courtyards, ancestral tablets, oil lamps and incense smoke; Miao Mountain Village gives a stilt-house village in mist. For a sunlit European-style village that is wrong in ways the audience can feel, pick Pastoral Healing or Frozen Wilderness and let the script make it wrong. Invent the ritual, show the moment before and the morning after, and keep the cruelty off screen.

Brief for AI Create: A folk horror drama: a woman returns to her grandmother’s village to sell the house and finds the neighbours have been keeping her room ready for twenty years; the harvest festival starts on Friday and everyone is being very kind; 8 short episodes, every episode ends on a door that was open and is now shut.

## Choose the place first

| Kind of folk horror | Scene Style | Costume Style |
| Chinese old house, temple or tomb | Chinese Folk Occult | Taoist & Monk Robes for priests and monks; everyone else in everyday clothes |
| Mountain village in southwest China | Miao Mountain Village | Miao Silver |
| A bright, too-welcoming village | Pastoral Healing | Eastern European Folk or Celtic Tribal |
| A cold northern settlement | Frozen Wilderness | Nordic Viking or Inuit Arctic |
| An old house with a secret | Gothic Dark | Victorian or Gothic Court |
Chinese Folk Occult is described in the style list as dark courtyards, black-lacquered ancestral tablets, oil lamps that never go out and incense smoke. Every scene and costume style is described in Scene and costume styles, one by one. These are written into the descriptions when the script is read, so choose them first: why they are fixed afterwards.

## Choose how it looks

The visual style is fixed after Step 1. Shot on Film gives grain and soft blacks, Black & White Film makes everything older and colder, Digital HD is the clean option for a modern frame. For an illustrated horror, Ink Wash Animation and Classical Oil Painting carry dread well. Visual style presets, one by one. For the general rules of making horror without gore, see How to make an AI horror short.

## Invent the ritual

The strongest folk horror is built on a custom that is believable and made up: a lantern left in a window on one night a year, a seat at the table nobody takes, a song sung once, a gift that has to be accepted. Three reasons to invent it:

- It can be exactly as frightening as your story needs. No real tradition to get wrong.
- It keeps real religions out of it. A real faith’s rites played as horror are better left alone; your own invented festival is not anyone’s belief.
- It stays on screen as a rule. A rule is easy to write and easy to break: the lantern must be lit before the third bell.

## The fear, without the gore

- The procession: torches and lanterns through the village at dusk, from a long way off, and then the empty street.
- The offering: what is carried in, and what comes out.
- The mask: a face you cannot read, turned toward the lead and held a beat too long.
- The hospitality: a table, too many dishes, everyone watching the newcomer eat.
- The morning after: the cruelty happens in the gap; the proof is a chair, a sleeve, a quiet door.
Gore is blocked on every generation; the dread above is not. If a clip is blocked. Crowds are better written as a background line than as named characters: Background people in a scene.

## A ritual scene, shot by shot

| Beat | Shot to adapt |
| The village | [Shot 1] 4s: Wide shot, fixed camera. A mountain village at dusk, timber houses stepping up the slope, lanterns being lit in every window one by one. @Mei stands at the end of the lane with her suitcase. |
| The welcome | [Shot 2] 3s: Medium shot. An old woman, @Lin, takes @Mei’s bag from her hand, smiling, and does not let go. |
| The line | [Shot 3] 4s: Close-up on @Lin. @Lin (warm, rehearsed): We kept your room just as you left it. (voice: Aunt Lin) |
| The room | [Shot 4] 3s: Medium shot through a doorway. A small room with the bed made, a lantern lit on the sill, a child’s shoes by the bed. @Mei stops in the door. |
| The window | [Shot 5] 3s: Wide shot from the room, looking out. The whole lane is quietly watching the window. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- The return: a granddaughter, a house to sell, a village that has been waiting.
- The outsider: a researcher, a nurse, a teacher, welcomed too well.
- The rule of the house: three things you must not do after dark, and one reason for each.
- The festival: a date on the calendar and everyone counting down to it.
- The missing: who left, who was sent, who never existed.
A character changed by something that happened to them (a mark, a transformation, a new face) is one character with a second look, not a second character. Costume changes and scene states.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief in AI Create.
- Step 1: Type Short Drama, TV Series or Film; the scene and costume styles from the table; a visual style; Target Region East Asia, Eastern Europe or Follow script; dialogue language; 9:16 Portrait or 16:9. Read the script.
- Step 2: a specific face and one mark per character; the lantern, the mask and the key as props. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). Set Night and Fog on the headers of the dark scenes.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | A lane of lit windows. | Her room has been kept ready for twenty years. |
| 2 | A dinner with too many dishes. | Everyone stops eating when she stops. |
| 3 | The attic and the child’s shoes. | The shoes are her size. |
| 4 | The festival rehearsal. | Her name is in the song. |
| 5 | The road out of the village. | The lantern is lit at the other end of it too. |

## FAQ

### How do I make a folk horror drama with AI?

In SceneMixer's Step 1, choose a scene style that carries the place (Chinese Folk Occult, Miao Mountain Village, Pastoral Healing, Frozen Wilderness or Gothic Dark), a matching costume style and a dark visual style, then write or generate the script.

### What does the Chinese Folk Occult scene style look like?

Old family halls, Taoist temples, mourning halls and ancient tombs: dark courtyards behind tall brick walls, black-lacquered ancestral tablets, oil lamps and incense smoke.

### Which costume style fits a priest or a monk?

Taoist & Monk Robes covers Taoist priests, monks, feng shui masters and folk exorcists, with crossed-collar wide-sleeved robes and embroidered ritual robes.

### Can I use a real religion's rites?

Better not. Invent the custom, which can be as frightening as your story needs and is not anyone's belief.

### Can I show blood or violence?

No. Show the procession, the offering and the morning after. SceneMixer blocks gore on every generation.

### How do I show a crowd of villagers?

Write them as a background line in the shot rather than as named characters, and give the lead and a few faces a sheet each.

### Can a character change after something happens to them?

Yes, as a second look of the same character, picked for the segments where the change shows.

### Will the villagers look the same in every episode?

The ones you draw do. Each character is drawn once and SceneMixer uses that sheet in every shot.

Updated November 3, 2026


---

# AI Fruit Drama: How to Make a Fruit Soap Opera Series, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-fruit-drama/
A strawberry, a banana and a betrayal in the villa: how to cast fruit characters that stay recognisable, write confessional shots and make a fruit soap opera.

# How to Make an AI Fruit Drama

Quick answer: A fruit drama is a soap opera or a villa dating show where every contestant is a piece of fruit with a face: a strawberry in love, a banana who cheats, a lemon who tells everyone. It works because the format is familiar and the cast is absurd, so the writing can be completely serious. In SceneMixer, each fruit is a character with its own card, image and voice; pick an animation style in step 1 (3D Toy Cartoon or Family 3D Animation), give every fruit one colour, one shape and one accessory nobody else has, keep each shot to three named fruit or fewer, and cut between the drama and the confessional, where one fruit talks straight to camera.

Brief for AI Create: A fruit soap opera: in a villa by the sea, Strawberry falls for Mango on the first night, but Mango's ex, Pineapple, arrives on day three with a secret; the Banana host announces a recoupling; Strawberry has to choose; dramatic, petty, very serious.

## Two formats that work

- The villa dating show. Fruit arrive at a villa, pair off, get tested, recouple. Every episode has an arrival, a betrayal or a vote. The confessional — one fruit on a sofa talking to camera about the others — is half the comedy.
- The kitchen soap opera. A fruit-bowl family, a forbidden romance, a rival from the next bowl. Played as straight melodrama, with slaps that stop before contact and dramatic zooms on a gasp.
Both run on the same rule as any short drama: a hook in the first seconds, one conflict per scene, a cliffhanger at the end. Hooks and cliffhangers.

## Settings to pick in step 1

- Visual Style: 3D Toy Cartoon for the glossy, bright look most fruit dramas use; Family 3D Animation for a softer, family-film one; Anime for a flat cartoon. Live action is the wrong choice: a fruit with a face is a cartoon by definition.
- Scene Style: Island Resort for the villa and the pool, Suburban Mansion for a family soap, Soft Romance for a sweet one.
- Type and Aspect Ratio: Short Drama and 9:16 Portrait — this is a feed format.
- Costume Style: optional. Fruit wear what you write on their cards; a project-wide style is not needed.
Visual style presets.

Visual Style: Family 3D Animation

Visual Style: Anime

Scene Style: Island Resort

Scene Style: Suburban Mansion
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Cast the fruit so nobody gets mixed up

Ten fruit in a villa can blur together. Give each one a card in step 2 and three things no other fruit shares:

- A colour and a shape that read at a glance. A strawberry and a cherry are both small and red; do not cast both.
- A face. Say it has one, and how much of a person it is: large round eyes, a small mouth, thin stick arms, short legs, walks upright.
- One accessory. Mango’s gold chain, Lemon’s sunglasses, Strawberry’s bow. It is the fastest way for viewers to tell two similar fruit apart.
A full example for a card: A plump ripe strawberry the size of a hand, glossy red with yellow seeds, a leafy green top like a little crown, large dark eyes with long lashes, a small pink mouth, thin arms and short legs, a white silk bow on the top leaf; walks upright. The card description is what the sheet is drawn from, so the bow will be there in every shot. Non-human characters: the general rules · Regenerate one sheet.

Character sheets from the sample Rocket Pup: one reference image per character, used in every shot they appear in. Non-human leads get sheets the same way: what the thing is, its face, its one mark.

## Write the confessional and the drama

Keep each shot to three named fruit or fewer. A wide shot of the whole villa is a crowd, and a crowd loses faces; put the other contestants in the segment’s background line instead.

| Beat | Shot to adapt |
| The arrival | [Shot 1] 4s: Opening: wide shot, the villa pool at sunset. @Pineapple walks in through the gate, sunglasses on. The others at the pool turn. No dialogue. Ambient: a record scratch, then silence. |
| The confessional | [Shot 2] 5s: Medium shot, fixed camera, a pink sofa against a palm-leaf wall. In frame only: @Strawberry, facing the camera. @Strawberry (whispering, furious): She said she was a fruit salad. She is not a fruit salad. |
| The reaction | [Shot 3] 3s: Close-up. In frame only: @Mango. He drops his smoothie. No dialogue. Ambient: the glass shattering. |
Play it completely straight: the delivery notes are the comedy. (through her teeth), (trying not to cry), (very slowly). How to write one shot · Background people.

## Voices: the fastest way to make fruit funny

Every fruit has a Voice description like any character. Cast against type: a tiny blueberry with a deep, slow, gravelly voice; a huge watermelon who is breathy and shy; the banana host bright, game-show smooth, always one beat too cheerful. A fruit with a mouth moves it when it speaks; one without a mouth does not, so give your speaking fruit a mouth. How character voices work.

## Make it in SceneMixer, start to finish

- Script: write it with the fruit as named characters, or put a brief into AI Create with the format, the couples, the twist and the tone.
- Step 1: an animation style and the villa set; read the script.
- Step 2: check every fruit has a colour, a shape and an accessory of its own; regenerate any two that look alike. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). Split any shot with more than three named fruit.
- Video: generate on the default tier at 10 credits a second; composite; post the first episode and keep the series going in the same project.

| Episode | What happens | The hook |
| 1 | Arrivals at the villa. Strawberry and Mango couple up on the first night. | A new arrival is announced — Mango goes pale. |
| 2 | Pineapple arrives and sits next to Mango at dinner. | Lemon saw them by the pool at midnight. |
| 3 | Lemon tells everyone in the confessional, then tells Strawberry to her face. | Strawberry finds Pineapple’s sunglasses in Mango’s room. |
| 4 | The Banana host calls a recoupling. | Mango chooses — and it is neither of them. |
| 5 | The finale vote. | Strawberry walks out of the villa with the one fruit nobody noticed. |

## FAQ

### Can I make a fruit drama on SceneMixer?

Yes. Write the fruit as characters, pick an animation style in step 1, and follow the same steps as any drama. Each fruit gets its own card, image and voice.

### How do I stop two fruit looking the same?

Give each one a colour and shape no other fruit has, a face, and one accessory of its own. If two sheets still look alike, regenerate one with a sharper description.

### Which style suits a fruit drama?

3D Toy Cartoon for the glossy, bright look, Family 3D Animation for a softer one, or Anime for a flat cartoon. The style is fixed for the project once you continue.

### How many fruit can be in one shot?

Three named fruit or fewer works best. Put the rest of the villa in the segment's background line instead of naming them.

### How do the fruit talk?

Give each one a voice description and a mouth. Lines are written the same way as for people; a fruit without a mouth does not move one.

### Can I make it vertical for TikTok?

Yes. In SceneMixer, pick 9:16 Portrait in step 1. The aspect ratio is fixed once you continue.

### Can I use characters from a famous show or game?

No. Known characters and franchise designs trigger the IP check. Invent your own fruit and their names.

### Does a fruit drama cost less because it is animated?

No. Video on SceneMixer is priced by the second at the tier you choose, whatever the style. Images are priced the same for every character.

Updated November 3, 2026


---

# AI Game Trailer or Cutscene: Your World in Game Engine Realism

URL: https://scenemixer.com/guides/make-an-ai-game-trailer-or-cutscene/
Make a game trailer or cutscene with AI: Game Engine Realism or Cel-shaded 3D, Cyber Metropolis, Wasteland or Dungeon sets, your own characters, 60–120 seconds.

# How to make an AI game trailer or cutscene

Quick answer: A game trailer is a 60- to 120-second film in three acts: the world, the hero, the threat, ending on the title. A cutscene is one scene with dialogue. SceneMixer makes both from a script: choose Game Engine Realism (sharp materials, no grain, like a next-generation engine) or Cel-shaded 3D for a stylised game, pick a set such as Cyber Metropolis, Wasteland, Starship or Dungeon, and your characters are designed once and kept in every shot. It is for your own game and your own characters; published game characters are refused by the IP check. No HUD or on-screen text is drawn; titles and logos go on a poster and in your edit.

Brief for AI Create: A 90-second trailer for a game about a salvage crew on a drowned Earth: the world (flooded towers), the hero (a diver who hears the city's old radio), the threat (a company that wants the signal); three acts, Game Engine Realism, ends on a black frame for the title.

## Trailer or cutscene

| You want | Project | Shape | Guide |
| An announcement or launch trailer | Film type, 16:9, 60–120 s | World, hero, threat; a narrator or none; rising cuts; the title last | AI trailer generator |
| A story cutscene | Film type, 16:9, one scene | Two or three characters, a place, a decision; dialogue performed | Dialogue in the storyboard |
| A pitch for a publisher | A trailer plus one cutscene plus a poster | What the game feels like, before it exists | A pitch trailer |
| A campaign recap | TV Series, episodes | Your tabletop or MMO guild's story as a series | A tabletop campaign as a series |

## The look of a game

| Visual Style | What it looks like | Suits |
| Game Engine Realism | Next-gen engine graphics: sharp materials, clean edges, no grain | Shooters, RPGs, sci-fi, open worlds |
| Cel-shaded 3D | Anime characters built in 3D, flat colour bands | Stylised action, JRPG-style |
| Realistic 3D | Near-human CG characters | Cinematic story games |
| 3D Toy Cartoon | Rounded, matte, bright | Party and family games |

Visual Style: Cel-shaded 3D

Visual Style: Realistic 3D

Scene Style: Cyber Metropolis

Scene Style: Wasteland

Scene Style: Starship

Scene Style: Dungeon

Costume Style: Cyberpunk Augment
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

Scene presets set the world (Cyber Metropolis, Wasteland, Starship, Mars Base, Dungeon, Medieval Castle, Magic Academy, Alien World), costume presets the faction (Cyberpunk Augment, Plate Armor, Fantasy Adventurer, Space Explorer, Starship Uniform). Every scene and costume preset.

## Your characters, designed once

Describe the hero as a sheet would: build, face, scars, armour, the thing they carry. The sheet is drawn once and every shot is generated from it, so the diver in the first act is the diver in the last. Members can upload concept art as the reference and it is redrawn in the look. Factions get a costume style and a mark; vehicles and weapons are props with their own reference. Character sheet generator · Props and references.

## The trailer, shot by shot

| Act | Shot to adapt |
| The world | [Shot 1] 4s: Wide shot, slow push in. Flooded skyscrapers at dawn, water lapping at the twentieth floor. Ambient: wind, a distant radio tone. |
| The hero | [Shot 2] 3s: Medium shot. @Ines, the diver, surfaces beside a rusted sign, pulls off her mask. @Ines (quietly): It's still broadcasting. |
| The threat | [Shot 3] 3s: Low angle. A corporate skiff with a searchlight sweeps the water. Ambient: an engine, a loudspeaker. |
| The turn | [Shot 4] 2s: Close-up on @Ines. She dives. |
| The title | [Shot 5] 3s: Wide shot, black water, a single light descending. Cut to black for the title card in your edit. |

Camera language works when written into the shot: slow push in, low angle, orbit. Keep each shot to one move, and show the fight before and after, not the wound. Cinematic camera moves · Fights and crowds.

## What it will not do

- Published characters and franchises. Refused by the IP check. This is for your game.
- HUD, menus, on-screen text. Not drawn reliably; add them in your edit, and put the title on a poster. Movie poster generator.
- Gameplay footage. It generates cinematics from text, not captures from a build.
- Gore. Violence is shown before and after; graphic injury is blocked. Content rules.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Cyber Metropolis, Wasteland; Costume Style Cyberpunk Augment, Near-Future Techwear; Visual Style Game Engine Realism, Cel-shaded 3D; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the hero and the antagonist with one mark each; the flooded city and the skiff as locations; the radio and the diving rig as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post. Export the trailer, then add the title card and logo in your editor.

| Act | Seconds | What it does |
| World | 15–25 | Where we are and what is wrong with it |
| Hero | 20–30 | Who, what they want, the thing they carry |
| Threat | 20–30 | Who is against them; the stakes |
| Turn and title | 10–15 | The decision; black frame for your title |

## FAQ

### Can AI make a game trailer?

Yes. Write the trailer as a short script in three acts, choose Game Engine Realism or another game look and a set, and the shots are generated with your own characters kept throughout.

### Can it make cutscenes?

Yes. A cutscene is one scene with dialogue performed by your characters in the look you chose.

### Can I use my game's characters?

Your own, yes; SceneMixer members can upload concept art as the reference. Published game characters are refused.

### Does it show a HUD or health bars?

No. On-screen text is not drawn reliably; add interface elements in your edit.

### Which look is closest to a modern game?

Game Engine Realism. Cel-shaded 3D for a stylised game, Realistic 3D for cinematic story games.

### How long should a trailer be?

60 to 120 seconds in three acts, ending on a black frame for your title.

### Can it generate gameplay?

No. SceneMixer makes cinematics from text, not captures from a build.

### What does a 90-second trailer cost?

About 900 credits of video on SceneMixer's default tier plus the join, before retakes; raise the keepers to 1080P or 2K afterwards.

Updated November 3, 2026


---

# How to Make an AI Game World Trailer From Your Lore

URL: https://scenemixer.com/guides/make-an-ai-game-worldbuilding-trailer/
Make a game world trailer from your lore: pick three facts you can see, write places and factions as locations and marks, show the world, not the lore. Not gameplay.

# How to make an AI game worldbuilding trailer

Quick answer: A worldbuilding trailer shows what a world is like to be in, rather than reading out its lore. Choose three facts about your world that can be seen — a rule, a place, a faction's mark — and write each as a scene. In SceneMixer a place is drawn once and every shot is set in it, a faction is a costume style plus a mark, and the narrator covers history. It makes cinematics from text, not gameplay, and it does not draw readable maps or interface text.

Brief for AI Create: A 75-second world trailer for a fantasy RPG set on a dry river basin: the river has been gone forty years, a toll-bell hangs on the bridge over nothing, a ferrywoman still keeps her boat on the cracked bed; a narrator, slow, ending on the bell ringing once.

## Lore is not a trailer

Lore is what you know; a trailer is what a viewer sees. Turn each lore line into a picture:

| Lore line | What the viewer sees |
| The river has been gone for forty years | A boat on cracked mud, the waterline stained on the bridge pier |
| The Ferrymen tax the dead | A toll-bell on a bridge over nothing, a rope worn smooth |
| Salt is currency | A child pays for bread with a pinch of white from a twist of paper |
| Nobody speaks the old king's name | A statue with the face chipped away, flowers at its base |
Pick three. A trailer that tries to carry the whole lore book carries none of it.

## Write a short world bible first

Five headings are enough: places (two or three), factions (one mark each), rules (what is scarce or forbidden), a hero, a threat. A story bible template · A character sheet template.

| Lore element | Where it goes in SceneMixer |
| A place | A scene: drawn once, used by every shot set there |
| A faction | A costume style, plus one mark on each member |
| A creature or machine | A character, human or not (non-human characters) |
| An artefact | A prop with its own reference image |
| History | A narrator's voice-over, a line at a time |
| A rule of the world | Written into a shot's action and ambience |

## How a place is drawn once

Before the storyboard is written, each place in the episode gets a scene image and, where the account and the space allow it, a 360° panorama; the room is measured and the cast placed. That is why a shot that says over her shoulder, the toll-bell behind her can be framed from the place instead of guessed, and why the bridge is the same bridge in every shot. A project holds up to 100 scenes and 100 characters. Building the world · Your scene image becomes the set · What a world model is.

## A 75-second world tour

| Act | Seconds | Shows |
| The rule | 20 | The dry basin; the boat on the cracked bed |
| The place | 20 | The bridge over nothing; the toll-bell |
| The people | 20 | The ferrywoman; the Ferrymen's mark on a cuff |
| The line | 15 | The bell rings once; cut to black for the title |
[Shot 1] 4s: Wide shot, slow push in. A dry river basin at dawn, a wooden boat resting on cracked mud, a waterline stained on a bridge pier. Ambient: wind, a distant bell. No dialogue.
[Shot 2] 4s: Low angle. A stone bridge over nothing; a bronze bell on a rope at its midpoint. @Narrator (voice-over, slow): The river left forty years ago. The toll did not. (voice: Narrator)

Every shot is one move and one place; the title goes on in your edit. Game trailer or cutscene: the three-act shape · Cinematic camera moves.

## One world, several films

If your world has regions, make one short per region as episodes of a TV Series project in the same look, or as separate projects that share characters and scenes through the asset library. Reuse a character, scene or voice · A tabletop campaign as a series.

## What it will not do

- Gameplay. It makes cinematics from text, not captures from a build.
- Readable maps, names on signs, interface text. Not drawn reliably; add them in your edit.
- Published game characters and franchises. Refused by the IP check; this is for your own world.
- Gore. Violence is shown before and after; graphic injury is blocked. Content rules.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Wasteland, Medieval Castle; Costume Style Fantasy Adventurer; Visual Style Game Engine Realism, Painterly Anime; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the ferrywoman and one Ferryman, the bridge and the basin as places, and the toll-bell as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| You have | Make | Do not expect |
| A lore document | A 60 to 120 second trailer showing three facts | Every fact in the document |
| A place map | Scenes for the two or three places in the trailer | A readable map on screen |
| Faction designs | A costume style and a mark per faction | Logos drawn reliably |
| A finished game | A cinematic or a pitch trailer | Gameplay footage |

## FAQ

### Can AI make a game world trailer?

Yes. Write three facts of the world as scenes, draw the places and cast once, and make a 60 to 120 second film in a game look.

### Can I use my lore document?

Paste it as the script if it reads as scenes, or write a brief; a trailer needs three seeable facts, not the whole document.

### Does it make gameplay?

No. SceneMixer makes cinematics from text, not captures from a build.

### Can it show our map or faction logos?

Not reliably: readable text and logos are not drawn well. Add them in your edit.

### Which look is closest to a game?

Game Engine Realism for next-generation graphics, Cel-shaded 3D for a stylised game, Painterly Anime for painterly worlds.

### Can I use characters from another game?

No. Published characters and franchises are refused; this is for your own world.

### How many places can a world have?

A SceneMixer project holds up to 100 scenes; a trailer needs two or three.

### What does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join, so 75 seconds is 75 seconds of each, before retakes.

Updated November 3, 2026


---

# AI Gangster or Yakuza Movie: Families, Codes and Quiet Threats

URL: https://scenemixer.com/guides/make-an-ai-gangster-or-yakuza-movie/
Make a gangster, yakuza or triad movie with AI: the era and city in Step 1, tattoos and tailoring on the sheet, codes shown in rituals, and threat instead of gunfire.

# How to make an AI gangster or yakuza movie

Quick answer: A gangster film is about rules: who bows, who sits first, who is allowed to leave. In SceneMixer, pick the world (a 1920s American city, a Tokyo night district, a 1990s Hong Kong street), put the tailoring and the tattoos on each sheet, film the codes as rituals (the tea, the bow, the envelope) and keep violence as threat and aftermath, which is what the genre does best anyway.

Brief for AI Create: A yakuza film: the adopted son of an ageing boss is told to choose between the family's new partners and the old man who raised him; one night in Kabukicho, one dinner, one decision.

## Choose the world

| Version | Scene Style | Costume Style | Region and language |
| Yakuza | Neon Nights, Japanese Machiya | Power Tailoring, Kimono | East Asia, Japanese |
| Hong Kong triad | Retro Hong Kong | Retro Hong Kong | East Asia, Chinese |
| Prohibition gangster | New York Metropolis | Jazz Age | Western, English |
| Modern crime family | Urban Opulence | Power Tailoring | Any |

Scene Style: Japanese Machiya

Scene Style: Retro Hong Kong

Scene Style: New York Metropolis

Costume Style: Jazz Age

Costume Style: Power Tailoring
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

For a period city, write the decade into the script and the location descriptions. Dialogue language.

## Marks on the sheet

Rank is what the audience reads first. Write it into each character: full back tattoo of a carp, visible when the shirt is off, a missing joint on the left little finger, a three-piece suit, a pocket watch. A boss who wears the same suit in every scene is easier for the model and for the viewer. Describe a face.

## Codes as rituals

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, a tatami room, low table, men kneeling in two rows. |
| The tea | [Shot 2] 3s: Close-up. @Oyabun pours one cup and pushes it across the table. |
| The choice | [Shot 3] 3s: Close-up on @Kenji, looking at the cup, not drinking. |
| The rows | [Shot 4] 2s: Medium shot, the kneeling men glancing at each other. |
| The answer | [Shot 5] 3s: @Kenji bows until his forehead almost touches the mat. @Kenji: Forgive me, father. |

A cup not drunk carries more threat than a drawn gun. Props in a shot.

## Violence off the frame

Guns and swords can be props on the table; shootings, stabbings and wounds are blocked as graphic violence, and even ordinary fight shots are sometimes blocked and pass after a rewording. Cut away before, come back after: the door closing, the car leaving, the empty chair. Content rules. Real people cannot be depicted; invent the family and its name.

## Neighbouring genres

Mafia romance (the love story inside the family), police and crime (the other side), heist, noir detective.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Neon Nights, Japanese Machiya; Costume Style Power Tailoring, Kimono; Visual Style Shot on Film, Black & White Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the boss, the heir and the rival, each with one mark on the sheet; the meeting room and the street as locations; the cup and the envelope as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Act | Scenes |
| The family | A ritual that shows who is who |
| The offer | Outsiders with money |
| The split | Two loyalties, one dinner |
| The night | Everything happens off screen; we see the morning |
| The code | The last bow |

## FAQ

### How do I make a gangster movie with AI?

Pick the city and era in SceneMixer's Step 1, write rank and tattoos into each sheet, film the codes as rituals and keep violence off the frame.

### Can I make a yakuza film in Japanese?

Yes. Set East Asia and Japanese dialogue; Neon Nights and Japanese Machiya for the places.

### Can there be guns?

As props. SceneMixer blocks shootings and wounds; show the before and the after.

### Can I set it in 1920s Chicago or New York?

Yes: Jazz Age costumes, New York Metropolis, and the decade written into the script.

### Can I recreate a famous gangster film?

No. Famous films and characters trigger the IP check; invent your family.

### Which look suits it?

Shot on Film for the seventies feel, Black & White Film for noir, Digital HD for modern.

### Film or series?

Film in 16:9 for one story; Short Drama in 9:16 for a feed series.

Updated November 3, 2026


---

# AI Ghost and Haunted House Drama: Hauntings Without Gore

URL: https://scenemixer.com/guides/make-an-ai-ghost-or-haunted-house-drama/
Make a ghost story or haunted house series with AI: the ghost as a character with a state, Gothic Dark and Suburban Mansion houses, sound before sight, a rule to break.

# How to make an AI ghost or haunted house drama

Quick answer: A ghost story is a house with a rule: the door that must stay shut, the room nobody sleeps in. The ghost is a character with its own card in Step 2, and its pale, changed look is a state of that character, so it is the same ghost every time it appears. Build the house in Step 1 with Gothic Dark, English Manor or Suburban Mansion, write a light source into every dark shot, and put the creak and the whisper in the shot's ambient line before anyone sees anything.

Brief for AI Create: A haunted house drama: a family buys a lakeside house far below its value; the youngest daughter starts talking to a girl in the attic who died there in 1962, and the girl says the family is not the first to be lied to by the estate agent; 8 episodes, each ends with a sound at 3 a.m.

## Choose the house

| Kind of haunting | Scene Style | Costume Style |
| A gothic mansion | Gothic Dark | Victorian, Gothic Court |
| A country house with a past | English Manor | Soft Everyday for the living, Regency or Victorian for the dead |
| The new house in the suburbs | Suburban Mansion | Soft Everyday |
| An abandoned hospital | Hospital Ward | Professional Uniforms |
| A Japanese house | Japanese Machiya | Kimono |
| Chinese folk spirits | Chinese Folk Occult | Taoist & Monk Robes |
Folk horror · Horror shorts.

## The ghost as a character

- Give it a sheet. A ghost seen five times must be the same face five times. Describe the person it was, with one mark (a ribbon, a burn on the wrist).
- Its ghostly look is a state. The living girl in a 1962 photo and the pale figure in the attic are two states of one character. States.
- A voice from nowhere. A ghost heard but not seen speaks as an off-screen voice in the shot. Voice-only characters.
- Write the light. “One candle on the landing, moonlight through the attic window” gives a dark shot something to see by. Describe the cause.

## The 3 a.m. scene, shot by shot

| Beat | Shot to adapt |
| The sound | [Shot 1] 3s: Close-up on @Lily asleep, a night light on the table. Ambient: a slow knock from above. |
| The stairs | [Shot 2] 4s: Medium shot from behind. @Lily climbs the attic stairs with a torch, the beam shaking. |
| The room | [Shot 3] 3s: Wide shot, the attic, moonlight through one round window, an old rocking horse moving by itself. |
| The ghost | [Shot 4] 4s: Medium shot. @June stands by the window in a 1960s dress, pale, a ribbon in her hair. @June (softly): You can see me. (voice: June) |
Fear comes from the rule and the sound; graphic horror is blocked by content review. Content rules.

## Story shapes that work

- The family that moved in: the house remembers the last owners.
- The ghost romance: the only person who can see him.
- The investigator: a sceptic paid to prove the house is not haunted.
- The unfinished business: a ghost who needs one wrong put right.
- The game: teenagers, a dare and a rule they break.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Gothic Dark, Suburban Mansion; Costume Style Soft Everyday, Victorian; Visual Style Shot on Film, Black & White Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the family, and the ghost with a living state and a ghostly state; the house's rooms as locations, the attic with a night look. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The lake house, moving day. | A child's voice from a sealed attic. |
| 2 | The kitchen. | The youngest says she has a new friend. |
| 3 | The town library. | A 1962 photo of the friend. |
| 4 | The attic at 3 a.m. | The ghost asks her to find something. |
| 5 | The estate agent's office. | A drawer of keys from every family before them. |

## FAQ

### How do I make a ghost story with AI?

Make the ghost a SceneMixer character with its own sheet and a ghostly state, choose a house scene style such as Gothic Dark or English Manor, and build each scene on sound first.

### How do I keep the ghost looking the same?

Give it a sheet in Step 2 and one mark you can recognise, and pick the same state each time it appears.

### Can a ghost speak without being seen?

Yes. Write it as an off-screen voice in the shot.

### How do I make dark scenes readable?

Write a light source into each shot: a candle, a torch, moonlight through a window.

### Can I show bloody ghosts?

No. SceneMixer blocks graphic injury. Pale skin, wet hair, an old dress and the wrong place at the wrong time are enough.

### Which visual style suits a ghost story?

Black & White Film or Shot on Film for dread, Digital HD for a modern haunting, Watercolor Anime for a gentle ghost romance.

### Can I make a ghost romance instead of a horror?

Yes. The same tools work: a ghost character with a state, a house, and a rule that keeps them apart.

### Can it be vertical?

Yes. 9:16 Portrait suits a short-drama feed; 16:9 Landscape suits a film or YouTube series.

Updated November 3, 2026


---

# AI Greek Mythology Drama: Gods, Heroes and Marble Worlds

URL: https://scenemixer.com/guides/make-an-ai-greek-mythology-drama/
Make a Greek mythology drama with AI: Greco-Roman sets and attire, gods as characters with a divine state, monsters described plainly, and public-domain myths retold.

# How to make an AI Greek mythology drama

Quick answer: Greek myths are public domain and built for drama: proud heroes, jealous gods, monsters with rules. In SceneMixer, set the world with Greco-Roman or Mediterranean Coast and Greco-Roman Attire, make each god a character with a mortal disguise and a divine state (light, scale, presence), describe monsters as creatures rather than by famous film designs, and retell the myth in your own words.

Brief for AI Create: A retelling of a Greek myth: the weaver who challenged a goddess to a contest, told from the goddess's side; six episodes, one loom, one transformation.

## The ancient world in Step 1

| Place | Scene Style | Costume Style |
| Temples and cities | Greco-Roman | Greco-Roman Attire |
| Islands and coasts | Mediterranean Coast | Greco-Roman Attire |
| Mount Olympus | Celestial Colossus used sparingly | Divine Regalia |
| The underworld | Dungeon, Gothic Dark | Greco-Roman Attire in dark colours |
Looks: Classical Oil Painting for a painted epic, Digital HD for live action, Realistic 3D for film CG.

## Gods have two forms

Myths turn on gods in disguise. Give each god a mortal state (an old woman, a shepherd) and a divine state (taller, lit from within, robes that move without wind). Pick the state in each segment; the reveal is switching it. States.

Transformations (a woman into a spider, a hunter into a stag) are a new character or a final state, with the moment itself shown through light and a cut.

## A contest, shot by shot

| Beat | Shot to adapt |
| The challenge | [Shot 1] 4s: Wide shot, a marble courtyard, two looms facing each other. Villagers watch from the colonnade. |
| The weaver | [Shot 2] 3s: Close-up on @Arachne's hands flying across the threads. |
| The goddess | [Shot 3] 3s: Medium shot. @Athena in her mortal state, grey cloak, watching. |
| The reveal | [Shot 4] 4s: @Athena in her divine state, light spilling across the courtyard. @Athena: You asked for me. I came. |

## Myths that adapt well

- Contests: mortals who challenge gods.
- Quests: a hero, a task, a monster per episode.
- Love and jealousy: gods who fall for mortals, and what it costs.
- The underworld journey: one rule, broken at the end.
Retell in your own words; avoid the designs of any recent film or game. IP check.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Greco-Roman, Mediterranean Coast; Costume Style Greco-Roman Attire, Divine Regalia; Visual Style Classical Oil Painting, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: a mortal and a divine state for each god; the courtyard and the temple as locations; a narrator if you want one. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Myth beat | The ending beat |
| 1 | The weaver's fame | A stranger asks to see her work |
| 2 | The boast | The stranger warns her |
| 3 | The contest | The goddess reveals herself |
| 4 | The tapestries | Arachne's shows the gods' crimes |
| 5 | The judgement | The transformation |

## FAQ

### How do I make a Greek mythology drama with AI?

Use SceneMixer's Greco-Roman sets and attire, give gods a mortal and a divine state, and retell public-domain myths in your own words.

### Are Greek myths free to use?

The ancient myths are public domain; recent films' and games' designs and characters are not.

### How do I show a god's power?

A divine state of the character: light, presence, robes that move, and the reaction of mortals.

### Can I show monsters like Medusa?

Yes, described plainly as your own design; violence stays off screen.

### Which look suits mythology?

Classical Oil Painting for an epic painting, Digital HD for live action, Realistic 3D for film CG.

### Can I make it for kids?

Yes, with a gentle retelling and a look such as Watercolor Storybook.

### Vertical or landscape?

Landscape suits temples and seas; vertical for short-form.

### Can it be narrated?

Yes. A narrator's voice-over suits myth retellings.

Updated November 3, 2026


---

# AI Heiress Drama: Hidden Identity, a Swapped Daughter and the Reveal

URL: https://scenemixer.com/guides/make-an-ai-heiress-drama/
A hidden-identity heiress drama with AI: who knows what, one woman under two names and two looks, a swapped-at-birth pair who must look related, and the reveal.

# How to make an AI heiress drama

Quick answer: A hidden-identity story runs on a gap: the audience knows who the heroine is and the people around her do not, and each episode closes a little of it. In SceneMixer that comes down to a few choices. In Step 1 pick a luxury set and wardrobe — Urban Opulence or Suburban Mansion with Haute Couture or Minimal Luxury. Tie her two names together once in the script so the card keeps both under Also called. Draw the disguise and the reveal as two states of one character, not two characters. Then stage the reveal as reaction shots — the proof, the face that recognises it, the silence — instead of a crowd.

Brief for AI Create: A heiress drama: a hotel kitchen hand named Nora is the missing granddaughter of the family whose gala she serves at; the family’s adopted daughter has worn her place for twenty years; 10 short episodes, English dialogue, one rung of the secret revealed at the end of each.

## What the genre runs on

Plan the secret as a ladder before you write a line. Each rung is one episode’s ending, and each one changes who knows what.

| Rung | What happens | Who learns it |
| A clue | Something small does not fit: a locket, a scar, a surname. | The audience, and one watcher |
| An ally | One person works it out and keeps quiet. | The ally |
| A near miss | She is almost exposed and the wrong person takes the blame. | Nobody on screen |
| The public reveal | The proof is shown in front of the people who wronged her. | Everyone |
Write who knows what into the script before episode 1. It is the first thing a long series loses. The same reversal drives billionaire romance and revenge dramas; this page is about the hiding.

## Settings to pick in Step 1

| Step 1 setting | Choose | Why |
| Scene Style | Urban Opulence, Suburban Mansion or Corporate Cold | The family’s town house and gala hall; the estate; the company floors. |
| Costume Style | Haute Couture or Minimal Luxury | The world she was born to: gala gowns, or old money that does not show. |
| Visual Style | Digital HD, Soft Glow or Shot on Film | Live-action looks suit the genre. |
| Target Region | Follow script, Western or East Asia | Names and faces for the family you write. |
| Dialogue language | Your choice | Spoken as each clip renders. |
The costume style is the wardrobe of the whole film. Her work clothes belong in her own description, written plainly, so that her two looks are different on purpose. Why these are fixed before the script is read · Every scene and costume style.

## Two names, one person

She is Nora in the kitchen and Eleanor in the family’s old letters. If the script names two characters, the storyboard will cast two people. Say once, in plain words, that they are one person — Nora, whose real name is Eleanor Hartwell — and keep using whichever name the scene needs.

When the script is read, her card keeps the other names your text uses for her on its grey Also called line. That line is read-only: it is filled by the read and by renames, and there is nothing to type into it. A word that could mean anyone (“the girl”) is not kept. Also called.

## Two looks, one face

The disguise and the reveal must be the same woman. Do not make two characters. Give her one sheet and draw the second look as a state — a second sheet for the same character. After the storyboard is generated SceneMixer assigns states to segments from the story, and you can override any segment in the editor; only a state that has an image is used. Using a state in one scene · Costume changes and scene states.

What carries the likeness is the face. Describe it once in Step 2, specifically, and let the clothes, hair and make-up change around it. The same face in every shot.

## The swapped daughters

A swapped-at-birth story needs two young women who read as relatives without being copies. Write the shared features into both descriptions — the same brow, the same hands, the same way of standing — and let age, grooming and clothes separate them. Do not write “twin”, and do not compare either to a real person: a named celebrity can trigger the IP check.

The proof is a prop. A locket, a bracelet or a letter is made once with its own image and named in the shots with @, so it is the same object in the clue, in the near miss and in the reveal. The same prop in every hand.

## The reveal, written as shots

Do not film the crowd. Film the proof, the face that recognises it and the silence. The guests are background people; two or three named faces carry the scene. In this example @Margaret is the family’s matriarch and @Celeste is the adopted daughter. Background people in a scene.

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, fixed camera. A hotel ballroom at a charity gala, a long table of guests, waiters moving between the chairs. @Nora carries a tray of glasses along the wall in a plain black server’s uniform. |
| The clue | [Shot 2] 3s: Close-up on @Nora’s hand setting a glass down. The silver chain of @Locket slips out of her collar and swings in the light. |
| The recognition | [Shot 3] 3s: Close-up on @Margaret at the head table. Her smile stops. Her eyes are on @Locket. |
| The line | [Shot 4] 4s: Medium shot. @Margaret rises from her chair and the room goes quiet. @Margaret (barely audible): Where did you get that? (voice: Margaret) |
| The silence | [Shot 5] 3s: Close-up on @Celeste beside her, a glass held halfway to her mouth. She does not move. |
How to write one shot · Hooks and cliffhangers.

## Keep the humiliation on the safe side

- Describe wealth, do not brand it. A named luxury label can trip the IP check; a black sedan, a grey silk gown and a long dining table do not. If an IP check blocks a character.
- Cruelty stays small and stays on screen as a look or a spilled glass. Anything harsher is better as the moment before and the moment after. Content rules and moderation · If a clip is blocked.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief in AI Create. Put the “who knows what” ladder in the first lines.
- Step 1: Type Short Drama or TV Series; scene and costume styles from the table; a visual style; Target Region; dialogue language; 9:16 Portrait or 16:9. Read the script.
- Step 2: give each woman a specific face and age; generate the second look as a state; make the locket a prop. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). One place and one stretch of time per segment: the kitchen is one, the ballroom another.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | The hotel kitchen, the night before the gala. | A chain slips out of her collar in the mirror. |
| 2 | The gala; a guest notices the chain. | The hostess asks to see it, then is called away. |
| 3 | The adopted daughter finds the matching mark. | She says nothing and takes the photograph. |
| 4 | A lawyer’s office on a rainy morning. | A name on a file: Eleanor Hartwell. |
| 5 | The family’s anniversary dinner. | The locket is placed on the table. |

## FAQ

### How do I make an AI heiress drama?

In SceneMixer's Step 1, pick a luxury scene style and costume style, a visual style, the region and dialogue language, then write the script around a ladder of who-knows-what and render episode by episode.

### Can one woman have two names?

Yes. Say once in the script that the two names are one person. When the script is read, her card keeps the other names under Also called, which is read-only.

### How do I keep her face the same in disguise and in the reveal?

Make one character and draw the second look as a state of that character. The face stays on the sheet; the clothes and hair change.

### How do two daughters look related?

Write the shared features into both descriptions and let age, grooming and clothes separate them. Do not write twin, and do not compare them to real people.

### Can I name a luxury brand?

Avoid it. A named brand can trigger the IP check. Describe the look instead: a black sedan, a grey silk gown.

### How do I stage a reveal in front of a crowd?

Make the guests background people and give two or three named faces the reaction shots: the proof, the face that recognises it, the silence.

### Is this the same as a billionaire or revenge drama?

They share the reversal. This page is about the hiding: who knows what, two names, two looks and the reveal.

### Will she look the same in every episode?

Yes. Each character is drawn once and that sheet is used in every shot; her second look is a state of the same character.

Updated November 3, 2026


---

# AI Heist Movie: The Crew, the Plan and the Twist

URL: https://scenemixer.com/guides/make-an-ai-heist-movie/
Make a heist movie with AI: a crew with one look each, the plan told then shown, the target as a location drawn once, props that matter, and a twist planted early.

# How to make an AI heist movie

Quick answer: A heist is two films at once: the plan as they tell it and the job as it happens. In SceneMixer, give each member of the crew one strong look and one skill, draw the target (a vault, a gala, a train) once as a location, make the things that matter (the key card, the painting, the case) props, and plant the twist in a shot the audience will only understand later.

Brief for AI Create: A heist film: five people who each lost money to the same art dealer plan to steal one painting from his own gala; one night, one floor, one of them is lying.

## A crew you can read at a glance

| Role | Look written into the sheet |
| The planner | Grey suit, reading glasses, always holding a notebook |
| The inside person | Catering uniform, a name badge |
| The tech | Oversized hoodie, earpiece, laptop prop |
| The driver | Leather jacket, a toothpick |
| The face | Evening gown or tuxedo, a state for the gala |
Describe a face · Leads and supporting characters.

## Tell it, then show it

The plan scene is voice-over over the locations: the planner's voice while the camera moves through the empty gallery. The job is the same locations, now full, and something different happens. Because each location is drawn once, the two versions match. Voice-over lines · The set.

| Target | Scene Style | Costume Style |
| Gala and gallery | Urban Opulence | Haute Couture, Professional Uniforms |
| Bank or tower | Corporate Cold | Power Tailoring |
| Casino city | Neon Nights, Gulf Metropolis | Minimal Luxury |
| Period heist | Republican Era | Jazz Age |

## The twist, planted

| Beat | Shot to adapt |
| The plan | [Shot 1] 4s: Wide shot, the empty gallery. @Iris (voice-over): The painting comes off the wall at 9:40. Not 9:41. |
| The plant | [Shot 2] 2s: Close-up. @Leo pockets a second key card when no one is looking. |
| The job | [Shot 3] 4s: The same gallery, full of guests. @Iris checks her watch: 9:40. |
| The surprise | [Shot 4] 3s: The frame on the wall is empty already. |
Same prop in every hand.

## Keep it clean

Heists work without violence: misdirection, timing, a disguise. Guns can be props, but shooting and wounds are blocked. Content rules.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Urban Opulence, Corporate Cold; Costume Style Haute Couture, Professional Uniforms; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: one strong look per crew member and a gala state for the face; the gallery and the vault as locations; the key card and the painting as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Act | Scenes | Purpose |
| Setup | Why each member joins | Give everyone a reason |
| The plan | Voice-over through the empty target | Teach the audience the layout |
| The job | The same rooms, full | Show what goes wrong |
| The twist | The planted shot, explained | Reward attention |
| The getaway | The car, the dawn | A last look between two of them |

## FAQ

### How do I make a heist movie with AI?

Give each crew member one strong look, draw the target once as a location, make key objects props, and plant the twist early.

### How do I show the plan and the job in the same place?

The location is drawn once, so the empty version and the full version match; change the people and the time.

### Can I use voice-over for the plan?

Yes. Write the planner's lines as voice-over over the locations.

### Can there be guns?

As props; SceneMixer blocks shooting and wounds.

### Which look suits a heist?

Digital HD for modern, Shot on Film for a seventies caper, Black & White Film for noir.

### Film or series?

Both: a film for one job, a series for one job per episode.

### How many characters can the crew have?

Four to six named members works well; extras are background.

### Can I copy a famous heist film?

No. Famous films and characters trigger the IP check.

Updated November 3, 2026


---

# AI High School Drama: Make a School Series with AI, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-high-school-drama/
A school series with AI: Korean, Japanese, American, boarding-school or Chinese campuses with the right uniforms and lockers, and how to keep the romance innocent.

# How to make an AI high school drama

Quick answer: A high school drama only works when the school looks like a school from that country: the shoe lockers by the Japanese entrance, the Korean classroom with a row of lockers at the back, the American hallway and the English stone quadrangle. In SceneMixer the country comes from two choices in Step 1: a scene style — Korean High School, Japanese High School, American High School, Elite Boarding School or Campus Youth (a Chinese campus) — and a costume style to match, such as Korean School Uniform, Japanese School Uniform or Prep School Uniform. Target Region only says East Asia for Korea, Japan and China alike, so the scene style is what tells the schools apart. Keep the story to ordinary school life: rivalries, exams, clubs, a first crush that stays at a glance and a shared umbrella.

Brief for AI Create: A Japanese high school drama: a transfer student joins the photography club the week before the culture festival and finds one photo in the darkroom that should not exist; 8 short episodes, school-festival deadline, every episode ends on a new photo.

## Pick the school

Set these in Step 1, before the script is read. They are written into every scene and every character’s clothes, so they cannot be changed afterwards. Why they are chosen first.

| School | Scene Style | Costume Style |
| Korea — Seoul, cram schools at night | Korean High School | Korean School Uniform |
| Japan — shoe lockers, sliding doors, cherry blossom | Japanese High School | Japanese School Uniform |
| China — terrazzo corridors, camphor trees, running track | Campus Youth | School Uniform |
| United States — lockers, gym, playing field | American High School | Streetwear, Sportswear or Smart Casual |
| England or New England boarding school | Elite Boarding School | Prep School Uniform |
| A school of magic | Magic Academy | Academy Robes |
American public high schools mostly have no uniform, which is why that row points to everyday clothes instead. Every scene and costume style is described in Scene and costume styles, one by one.

## Choose how it is drawn

The visual style is the other half of the look, and SceneMixer fixes it after Step 1. Digital HD for a live-action school drama; Anime or Cel-shaded 3D for an anime school series; Korean Webcomic for a polished webtoon feel; Watercolor Anime for a gentle slice-of-life. Visual style presets, one by one.

## Region and dialogue language

- Target Region. East Asia gives an East Asian cast and names, and Western gives a Western one. Follow script keeps whatever your script already says. Target Region.
- Dialogue language. Korean, Japanese, English and the other supported languages are spoken on screen as the clip renders. Native dialogue vs dubbing.
- A second language version can be made later from the finished project. Second language version.

## Keep it ordinary, and keep it safe

School stories are ordinary content: classes, clubs, exams, friendships, rivalries, a secret, a rumour. Two things to watch.

- Romance stays innocent. A note in a locker, two people under one umbrella, a hand almost taken. Anything that sexualises a school-age character is refused whatever the context. Content rules and moderation.
- Violence stays off screen. A fight, a fall or a prank gone wrong works better as the moment before and the moment after. If a clip is blocked.
Teenagers can come out looking older than they are. Put the age and the school year in the first sentence of each character’s appearance in Step 2 — a fifteen-year-old Korean girl in her first year of high school — and regenerate the sheet. When a character looks the wrong age.

## A school scene, shot by shot

The beats are small, so give each one its own shot and its own seconds. Write what happens and who speaks; SceneMixer works out who is in frame.

| Beat | Shot to adapt |
| The hallway | [Shot 1] 3s: Wide shot, fixed camera. A school hallway at eight in the morning, students streaming past the lockers. @Hana stands at her locker holding a folded note. |
| The note | [Shot 2] 3s: Close-up on the note in @Hana’s hand. It reads: Rooftop. 4 pm. |
| The line | [Shot 3] 4s: Medium shot, @Hana looks up and across the hallway at @Ren, who is looking the other way. @Hana (whispering): He wrote this? (voice: Hana) |
| The look | [Shot 4] 3s: Close-up on @Ren. He glances back, meets her eyes for a second and turns away fast. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- The transfer student: everyone knows everyone, and one new face changes the order.
- The club with a deadline: a festival, a tournament, a showcase. The deadline gives every episode a clock.
- The rumour: one message in a group chat; each episode is who has heard it.
- Exam week: a rivalry that is really a friendship.
- The secret room: a rooftop, a darkroom, a library stack that nobody else uses.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write a brief in AI Create.
- Step 1: Type Short Drama or TV Series; the scene and costume styles from the table; a visual style; Target Region; dialogue language; 9:16 Portrait for feeds or 16:9 for YouTube. Read the script.
- Step 2: give each student a specific face and age, and one mark that is easy to see (glasses, a scarf, a hair clip). The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). Keep every segment to one place and one stretch of time: the hallway is one segment, the rooftop another.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | The new student walks into the photography club. | In the darkroom, a photo of the club room taken last night. |
| 2 | Nobody admits to taking it. | The negative strip has one frame missing. |
| 3 | The rooftop, 4 pm. | Nobody is there; the note is on the railing. |
| 4 | The culture-festival schedule is posted. | The club’s slot has been given to someone else. |
| 5 | A rival club offers a deal. | The missing frame is in their display case. |

## FAQ

### How do I make an AI high school drama?

In SceneMixer's Step 1, pick the scene style and costume style of the school you mean, a visual style, the region and the dialogue language; then write or generate the script and render episode by episode.

### Is there a style for a Korean or Japanese school?

Yes. Scene styles Korean High School and Japanese High School, with the costume styles Korean School Uniform and Japanese School Uniform.

### What about an American high school?

Pick American High School for the scene in SceneMixer. Most American public high schools have no uniform, so use Streetwear, Sportswear or Smart Casual for the clothes.

### Can I make a boarding-school drama?

Yes. Elite Boarding School for the stone quadrangles, chapel and dining hall, and Prep School Uniform for the crested blazers and striped ties.

### Can the characters speak Korean or Japanese?

Yes. Choose the dialogue language in SceneMixer’s Step 1 and the lines are spoken in it on screen. English lines in a Korean or Japanese school work too.

### Can I include a romance?

Yes, as long as it stays innocent. SceneMixer refuses anything that sexualises a school-age character, whatever the context.

### Can I change the school style after the script is read?

SceneMixer does not let you change the setting. Edit the scene and clothing descriptions in Step 2 and regenerate those sheets, or start a new project with the same script.

### Will the students look the same in every episode?

Yes. SceneMixer draws each character once and uses that sheet in every shot. A second look, such as a sports uniform, is a state of the same character.

Updated November 3, 2026


---

# AI History Videos: How to Make a Historical Drama Series, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-history-drama/
AI history videos that feel like a period drama: an invented witness inside a real event, narration or dialogue, the right era for sets and costumes, what gets blocked.

# How to Make an AI History Drama or History Video Series

Quick answer: The AI history videos that hold attention are not slideshows with a narrator; they put the viewer next to someone inside the event — a messenger at Waterloo, a stonemason on the pyramid, a telegraph girl in 1912. In SceneMixer you write that person’s story as a script, choose the era in step 1 (scene style, costume style, a film-like visual style), and every episode is rendered with the same cast and sets. Invented characters in a real period work best: named real political figures are refused by the content checks, so leaders stay in the background, in dialogue or off screen.

Brief for AI Create: The sinking of the Titanic seen by a young wireless operator, in five short episodes from departure to the lifeboats; documentary narration between scenes; each episode ends on the clock.

## Documentary or drama: pick the telling

| | Narrated history | Historical drama |
| Who carries it | A narrator, over scenes | Characters, in dialogue |
| Episode shape | One fact or turning point, about a minute | One scene with a cliffhanger, two to four minutes |
| Best for | History channels, Shorts, explainers | Series with returning characters |
| Visual Style | Shot on Film, Black & White Film, Classical Oil Painting | Digital HD, Shot on Film, Soft Glow |
Many channels mix both: a narrator opens and closes, the scene in between is acted. The narrator is an Off-screen character with a voice and no image. Voice-only characters.

## Pick the era in step 1

| Period | Scene Style | Costume Style |
| Ancient Egypt | Ancient Egypt | Ancient Egyptian Attire |
| Greece and Rome | Greco-Roman | Greco-Roman Attire |
| Middle Ages | Medieval Castle | Medieval Nobility, Plate Armor |
| Vikings | Frozen Wilderness | Nordic Viking |
| Regency and Victorian | European Palace, Parisian Streets | Regency, Victorian |
| American frontier | Wild West | Western Cowboy |
| 1920s | New York Metropolis | Jazz Age |
| Twentieth-century wars | Military Base, Epic Scale | Modern Military |
The styles set the direction; the exact year still belongs in each location’s and character’s description (a London street in 1888, gas lamps, horse-drawn cabs). Scene style and costume style.

Visual Style: Black & White Film

Visual Style: Classical Oil Painting

Scene Style: Medieval Castle

Scene Style: Greco-Roman

Scene Style: Wild West

Costume Style: Medieval Nobility

Costume Style: Victorian
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Real events, invented witnesses

The content checks refuse real public figures by name or likeness, and real political figures are not made here at all. That sounds limiting for history and is mostly not: the strongest history videos are told by people nobody has heard of.

- Put an invented character at the centre: the general’s aide, the cook, the soldier’s wife.
- Let the famous figure stay off camera — a voice behind a door, a rider seen from far away, a name spoken by others. A name in a line can still be checked; if a line is refused, describe instead (the emperor, the president).
- Keep the facts in the narration, where you control them, and the feeling in the scene.
Content rules and checks · IP check: celebrities, brands, franchises.

## Battles and disasters

Big events are a run of short shots: a wide shot that shows the scale, the witness reacting, one detail (a dropped musket, the water at the stairs), back to the witness. Show violence at a distance or by its result; detailed gore is refused. Crowds are declared as background people, not cast one by one. Writing fights, crowds and group scenes · Background people.

| Beat | Shot to adapt |
| Scale | [Shot 1] 4s: Extreme wide shot, slow pan. A muddy field at dawn, two lines of soldiers in red and blue coats, smoke drifting low. @Narrator (voice-over, measured): By nine in the morning, the rain had stopped. Ambient: drums, distant shouts. |
| The witness | [Shot 2] 3s: Close-up. @Tom, a drummer boy, swallows hard, rain on his face. No dialogue. Ambient: drums. |
| The detail | [Shot 3] 2s: Insert. A letter in Tom’s hand, the ink running. No dialogue. Ambient: a cannon, far off. |

## One person across the years

A series that follows someone from youth to old age keeps one character card and adds states for the later looks: grey hair, a uniform, a wound. The face comes from the main sheet each time, so the audience knows it is the same person. Costume changes and states · Character looks the wrong age.

Character sheets from the sample 浪人の誓い: one reference image per character, used in every shot they appear in.

## Make it in SceneMixer, start to finish

- Script: your own on Upload Script, or a brief in AI Create naming the event, the witness and how many episodes.
- Step 1: Type Short Drama or TV Series; the visual, scene and costume styles for the period; 16:9 for YouTube or 9:16 Portrait for Shorts; the dialogue language and target region. Then read the script.
- Step 2: check each character’s clothing names the period; regenerate any sheet that looks modern. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project); write the narrator as voice-over.
- Video: generate on the default tier at 10 credits a second; composite; post.
For a period look on film, watch 浪人の誓い, a samurai drama set in an estate courtyard, with a grainy, film-like live-action look.

| Episode | The moment | The hook |
| 1 | The ship leaves Southampton; the operator’s first night shift. | An ice warning he is told to ignore. |
| 2 | The passengers above, the operator below. | Another warning. 11:40 pm. |
| 3 | The impact, felt as a shudder in the wireless room. | The captain’s order: “Send the call.” |
| 4 | Calling ships that do not answer. | One ship answers. Four hours away. |
| 5 | The lifeboats. | The narrator: who was saved, and who was not. |

## FAQ

### How do I make AI history videos for YouTube?

Write the event as a short script told by one witness, or give SceneMixer's AI Create a brief. Choose a film-like visual style and the period's scene and costume styles in step 1, 16:9 for YouTube, and add a narrator as an off-screen character.

### Can I show real historical figures?

Named real public and political figures are refused by the content checks. Tell the story through invented people, and keep famous figures off camera or described instead of named.

### How do I make sets and costumes look accurate?

Pick the matching scene and costume styles in step 1, then write the year and the details into each location's and character's description.

### Can a narrator explain the history?

Yes. Add an off-screen narrator character and write its lines as voice-over; they are heard over the shots.

### Can I show battles?

Yes, as a sequence of short shots with scale, reaction and detail. SceneMixer refuses violence shown in close detail.

### Can the same character age across episodes?

Yes. Keep one card and add states for the older looks; the face is taken from the main sheet.

### What style looks most like a documentary?

Shot on Film or Black & White Film for archive-like footage; Classical Oil Painting for a painted, museum feel.

### How long should a history episode be?

About a minute for one narrated turning point; two to four minutes for an acted scene with a cliffhanger.

Updated November 3, 2026


---

# AI History Video Generator: Narrated Scenes From the Past

URL: https://scenemixer.com/guides/make-an-ai-history-video/
Make a history video with AI: narrated scenes with invented witnesses, a facts-first script, an honest list of what it cannot draw, and period sets and costumes.

# How to make an AI history video

Quick answer: SceneMixer makes history videos as narrated, dramatised scenes: a voice-over states the facts and short scenes with invented witnesses show what a day in that place looked like. It does not make archive footage, legible maps and timelines, or the faces of real historical figures. So the honest recipe is: facts go in the narration, where you control them; feeling goes in the scene; and you check every date and object yourself before you render.

Brief for AI Create: A three-minute history explainer: a day in a Roman bakery from the first fire before dawn to the last loaf sold; an invented baker and apprentice, a calm narrator, four scenes, the facts said in the narration.

## What it can and cannot do

| It can | It cannot |
| Show invented people living in a real period | Show archive footage or real photographs |
| Build sets and costumes in the style of an era | Draw the face or voice of a real, named person; real public figures are refused |
| Narrate facts over scenes, in any of 15 languages | Draw legible dates, labels, maps or timelines (generated video does not write legible text) |
| Keep the same cast and places across a series | Guarantee that every object is correct for the year |
Put maps, dates and captions in with your own editor afterwards, and treat the pictures as an illustration of the narration, not as evidence. Content rules · The IP check.

## Three formats

| Format | Shape | Typical narration share | Best for |
| Narrated explainer | Voice-over over scenes with no dialogue | About 70% of the runtime | One topic in 1 to 3 minutes; a channel or a lesson |
| A day in the life | One invented person from morning to night | About 40%, the rest is a few lines of dialogue | A trade, a place, a daily routine; 3 to 5 minutes |
| The event through a witness | A series; each episode follows the same witness | About 30% | A siege, a voyage, a flood; episodes of 2 to 4 minutes |
A history drama series · A documentary-style video · A shot list for a historical video.

## Facts first: the workflow

- Write a fact sheet. Five to ten facts with sources you can cite. Keep them in your own notes.
- Put every fact in a narration line. The narrator says it; the scene does not need to prove it.
- Give scenes only what all your sources agree on. Everyday tasks, tools and rooms rather than disputed details.
- Say when historians disagree. One narration line is enough: “Sources differ on this.”
- Put the year in every location's description. The style card sets the direction; the year belongs in the text. Scene and costume styles.
- Read the storyboard for anachronisms. Clothes, tools, writing, glass, paper. Editing is free on SceneMixer; rendering is not.

## A three-minute script, by the second

English narration is planned at about three words a second plus a pause at every sentence end, so a minute of connected narration is about 140 words. How many words fit a minute. The rest of the runtime is pictures with no voice.

| Part | Seconds | Narration (words) | On screen |
| Hook | 0 to 15 | About 30 | A close detail: hands, fire, flour |
| The place | 15 to 45 | About 55 | Wide shot, then the room |
| The morning work | 45 to 85 | About 60 | Invented baker and apprentice at the oven |
| The midday trade | 85 to 125 | About 60 | The street, customers, a brief exchange |
| The turn | 125 to 160 | About 55 | Something goes wrong: a shortage, a delay |
| The close | 160 to 180 | About 40 | The last loaf, the empty shelves, the fire banked |

## Period sets and costumes

Pick the era in SceneMixer’s Step 1, before the script is read. These cards are all in the picker; the exact year still goes in each description.

| Period | Scene Style | Costume Style |
| Viking age | Viking Fjord | Nordic Viking |
| Silk Road | Silk Road City | Silk Road Robes |
| Ottoman Empire | Ottoman Palace | Ottoman Court |
| The age of sail | Age of Sail Harbor | Pirate (for a ship's crew) |
| Pre-Columbian Americas | Pre-Columbian Americas | Aztec, Andean Highlands |
| Republican-era China | Republican Era | Republican Modern |
| The Second World War | WWII Front | WWII Era |

Scene Style: WWII Front

Scene Style: Pre-Columbian Americas
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
For a look that suits the age of the footage you are imitating, the Visual Style row has Black & White Film, Shot on Film and Classical Oil Painting.

## Make it a channel

A history channel is a list of topics and a cast of recurring invented guides. Ten topics you could start with: a day in a bakery, a night watch on a wall, a morning in a caravanserai, a market day in a port, a printing shop, a lighthouse keeper's year, a railway station in its first year, a harvest that fails, a letter delivered across a sea, a school in a village. Draw the places once, reuse the guide, and put the facts for each topic in the narration. A faceless YouTube channel · Videos for your classroom · A World War II drama.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Greco-Roman; Costume Style Greco-Roman Attire; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the baker, the apprentice and the bakery; the narrator is an Off-screen character, with a voice and no drawing, who speaks the facts. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Question to ask of every shot | Why |
| Could this object exist in that year? | Generated scenes can include modern tools and materials |
| Does the narration carry the fact? | The picture illustrates; the narrator states |
| Is there text on screen? | Generated video does not write legible text; add it in an editor |
| Is a real person visible? | Real public figures are refused; use an invented witness |

## FAQ

### Can AI make a history video?

Yes, as narrated scenes with invented people in a real period. SceneMixer does not make archive footage or the faces of real historical figures.

### Can I show real historical figures?

No. Real public figures are refused by SceneMixer. Put an invented witness at the centre and let the figure stay off screen, in a voice behind a door or a name in a line.

### Can it draw maps and timelines?

No. Video made on SceneMixer does not write legible text. Add maps, dates and captions in your own editor afterwards.

### Is the history accurate?

Only as accurate as your script. Put the facts in the narration, check the dates yourself and read the storyboard for anachronisms before rendering.

### Which style should I use?

Pick the Scene Style and Costume Style for the era in SceneMixer’s Step 1, then Shot on Film or Digital HD for a live-action look or Classical Oil Painting for a painted one.

### How long should a history video be?

One to three minutes for an explainer, three to five for a day in the life.

### Can I narrate in another language?

Yes. Dialogue and voice-over on SceneMixer can be in any of 15 languages.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; the narration costs nothing extra because it is generated with the video.

Updated November 3, 2026


---

# AI Hockey Romance Drama: The Rink, the Team and the Off-Ice Story

URL: https://scenemixer.com/guides/make-an-ai-hockey-romance-drama/
Make a hockey romance drama with AI: Sportswear kits with numbers, the rink and the locker corridor as locations, games in one-action shots, and the romance off the ice.

# How to make an AI hockey romance drama

Quick answer: Hockey romance is mostly off the ice: the bar after the game, the small town, the player who is all edges until he is not. In SceneMixer, dress the team in Sportswear with colours and numbers written into each sheet, make the rink and the locker corridor locations, write game moments one action per shot (the faceoff, the check, the goal, the bench), and let the romance happen everywhere else.

Brief for AI Create: A hockey romance: the new physiotherapist for a small-town team refuses to clear its star player to return from injury; he is not used to being told no; twelve episodes, one season.

## Kit and number on the sheet

With helmets on, players are told apart by number and colour. Write both into each player's description, and give the leads a game state (kit, helmet) and an off-ice state (a hoodie, a coat). States.

| Place | Scene Style | Costume Style |
| The town | American Small Town | Outdoor Gear, Soft Everyday |
| The rink | A rink written as a location | Sportswear |
| The winter outside | Frozen Wilderness | Outdoor Gear |
| The clinic | Hospital Ward | Professional Uniforms |

## The game in short shots

| Beat | Shot to adapt |
| The faceoff | [Shot 1] 3s: Low angle on the ice. @Cole, number 19, crouches at the faceoff circle. Ambient: crowd, skates. |
| The rush | [Shot 2] 2s: Wide shot. @Cole breaks away up the left side. |
| The shot | [Shot 3] 2s: Medium shot. He shoots; the puck hits the net. |
| The look | [Shot 4] 3s: Close-up on @Maya behind the bench, arms crossed, not smiling. |
Crowds are background; teammates without lines too. Crowds and group scenes · Sports dramas.

## Off the ice

- The injury: a reason they must see each other every day.
- The town: everyone knows everyone, and everyone has an opinion.
- The rival team: an ex, a brother, an old teammate.
- The big game: the season's last episode.

## Rough play, kept clean

Checks and scuffles read as one move per shot; fights stay short and without wounds. An injury is told through the physio room and a limp, not the moment it happens. Describe the cause, not the result.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style American Small Town, Frozen Wilderness; Costume Style Sportswear, Outdoor Gear; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: a game state and an off-ice state for each lead; colours and numbers in every player's sheet; the rink and the clinic as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | On ice | Off ice |
| 1 | The injury | She refuses to clear him |
| 3 | He watches from the bench | Physio sessions |
| 6 | His return | The bar after the win |
| 9 | The rival team | His past arrives |
| 12 | The final | Who she was watching |

## FAQ

### How do I make a hockey romance drama with AI?

Dress the team in Sportswear with numbers, make the rink a location, write game moments one action per shot, and set most of the romance off the ice.

### How do I tell players apart in helmets?

Write each player's colour and number into the description.

### Can AI show hockey games?

Yes, as short one-action shots; crowds are background.

### Can there be fights on the ice?

Short scuffles without wounds.

### Which format suits it?

Short Drama in 9:16 for feeds, or TV Series for longer episodes.

### Which look suits it?

Digital HD for live action; Korean Webcomic or Anime for an animated version.

### Can I use a real team or league?

No. Invent the team, logo and league.

### Can I set it in a small town?

Yes: American Small Town with winter set on the segments.

Updated October 31, 2026


---

# AI Halloween Horror Short: How to Make One That Scares, No Gore

URL: https://scenemixer.com/guides/make-an-ai-horror-short/
Make an AI horror short for Halloween: dread, a light source and one thing that is wrong, dark scenes lit so they render, scares built in sound, no gore.

# How to Make an AI Horror Short Film or Series

Quick answer: Horror on a phone screen is built from three things: something ordinary that is slightly wrong, a sound before a sight, and a cut at the worst possible moment. None of them needs gore, which is fortunate, because gore is blocked. In SceneMixer, choose Film for a single short or Short Drama for an episodic scare series, pick a style that can hold darkness (Black & White Film, Digital HD with Gothic Dark or Hospital Ward sets), write a light source into every dark shot, put the creak and the breathing in the shot’s Ambient line, and make anything that changes a face — a possession, a reveal — a state of that character.

Brief for AI Create: A horror short: a night-shift security guard in an empty care home keeps hearing the call bell from room 14, which has been sealed for a year; the residents' records say the last patient never left; he goes in at 3 a.m.; slow dread, one shock at the end.

## What makes a short scary

- The rule of the place. A room nobody enters, a call that comes at the same minute, a reflection that is a beat late. Establish the rule early; break it late.
- Sound first. The viewer hears it before they see it: a door, a child humming, breathing on the other side of a wall. Sound costs nothing in a shot and does most of the work.
- The almost-reveal. Show the shape in the background, out of focus, then cut to the lead not noticing. The audience does the screaming.
- One real shock. A short needs one; a series needs one per episode, at the end, as the hook. More than that and nothing lands.
- No explanation. The scariest endings answer the question with a worse question.

## A Halloween short before October 31

Halloween is the week people go looking for something scary to watch and share. A one-minute horror short is the right size for it: one place, one rule, two escalations and one shock — the shape in the table further down. It fits into a few evenings:

- Evening 1 — the script. Write the minute yourself, or give AI Create a brief: the place, its rule, the lead, the night it goes wrong. Pick Film, and 9:16 Portrait if it is for TikTok, Reels or Shorts. Read the script.
- Evening 2 — the cast and the storyboard. Generate the lead, the place and the thing on the default image tier, where your free sheets are drawn. On your first project the storyboard for episode 1 is free.
- Evening 3 — the video. Generate on SceneMixer's default tier at 10 credits a second and regenerate only the shots that miss. A segment usually takes ten to twenty minutes to come back; you can close the page.
- The poster. In step 2, Cover art, set the genre mood to Horror; the title is painted into three poster options. Promo cover posters.
- Subtitles. Feeds are watched with the sound off. If your account has Subtitles (it does once you have bought credits or a plan), tick Add subtitles when you export. Add subtitles.

Post it before the 31st. The same project works next year: reuse the cast and the house in a second episode.

## Settings to pick in step 1

- Type: Film for a single short with room to breathe (no episode cap, slower cut); Short Drama for a series with a scare at the end of every episode.
- Aspect Ratio: 16:9 Landscape for a short film with corridors and wide empty rooms; 9:16 Portrait for a feed series, where tight framing hides what is just outside it.
- Visual Style: Black & White Film for classic dread, Digital HD or Shot on Film for found-in-the-real-world horror, Ink Wash Animation or American Comic for a stylised one.
- Scene Style: Gothic Dark for old houses and churches, Hospital Ward for clinical horror, American Small Town for a town with a secret, Frozen Wilderness for isolation, Japanese Machiya for a quiet, traditional haunting, Chinese Folk Occult for old family halls, temples and tombs, Miao Mountain Village for a stilt-house village in mist (folk horror guide).

Visual Style: Digital HD

Scene Style: Gothic Dark

Scene Style: Hospital Ward

Scene Style: American Small Town

Scene Style: Japanese Machiya
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Light the dark, or the model will not

The most common failure in AI horror is a frame that is simply dim and muddy. A video model lights a scene from what the shot names; “a dark corridor” names nothing. Give every dark shot a source and say where it falls:

- A single flashlight beam from his hand, sweeping left to right; everything outside it is black.
- A red exit sign at the far end of the corridor, its glow on the wet floor.
- Lightning through the window, a white flash, then the room is gone again.
- Her phone screen under her chin, the only light in the room.
The darkness is then whatever the source does not reach — which is exactly where the scare lives. Describe the cause, not the result · Say what is there, not what is not.

## Ghosts, possessions and things in the dark

- A ghost that is seen is a character with a card and an image like anyone else. Describe what is wrong about them precisely: a woman in a grey 1970s nurse’s uniform, skin with no colour, hair wet as if from rain, always barefoot.
- A ghost that is only heard — a voice in the baby monitor, a caller — set to Voice only (off-screen) in SceneMixer’s step 2: no image is made, the face is never attached, the voice is.
- A possession or a reveal that changes a face is a state: on the character’s card, Add a state, name it, and describe only the difference: eyes fully black, a fixed smile, head tilted too far. Pick it in the shot where it happens.
- A creature is a non-human character: describe what it is, its size against a person, how it moves, whether it has a face. Non-human characters.
Leads, supporting cast and voice-only characters · States.

## Scares without gore

Every generation passes the content check; gore, mutilation and extreme violence are blocked, and repeated blocks pause video generation on SceneMixer for 1 hour. Horror is stronger without them:

| Instead of | Write |
| A body | A shoe at the bottom of the stairs, laces still tied. |
| An attack | The flashlight drops and rolls, its beam spinning across the ceiling. Black. A scream from off-screen. |
| A wound | He keeps his hand pressed to his side and will not let her look. |
| A monster’s kill | Something drags across the floor above them. Then nothing. |
Blocked by moderation: what to change · Content rules.

## A scare, written as shots

| Beat | Shot |
| The sound | [Shot 3] 4s: Medium shot, fixed camera. In frame only: @Tom. He stops in the corridor, head turning toward the dark end. No dialogue. Ambient: a call bell rings once, far away. |
| The almost-reveal | [Shot 4] 4s: Wide shot down the corridor, his flashlight beam the only light. At the far end, the door of room 14 stands open a hand’s width. No dialogue. Ambient: the bell again, closer. |
| The cut | [Shot 5] 3s: Close-up. In frame only: @Tom. His breath fogs in the beam. @Tom (whisper): That door was sealed. Ambient: behind him, very close, a second breath. |
The scare in shot 5 is entirely in the Ambient line: the second breath is heard, not seen. How to write one shot · Dialogue and sound in one generation.

## Make it in SceneMixer, start to finish

- Script: your own on Upload Script, or a brief in AI Create with the place, its rule, the lead, the night it goes wrong and the tone. The 1k or 2.5k length is enough for a short. Writing the one-line brief.
- Step 1: Type, aspect ratio, style and set as above; read the script.
- Step 2: generate the lead, the place and the thing; add the possession or reveal state; set heard-only voices to Voice only (off-screen).
- Storyboard: read every dark shot and add its light source; move scares into the Ambient line where you can.
- Video: generate on the default tier at 10 credits a second, regenerate the shots that miss, composite and post.

| Length | Shape | Type in step 1 |
| One minute | Rule, two escalations, one shock | Film |
| Three to five minutes | Rule, investigation, the almost-reveal, the shock, the worse question | Film |
| A series | Each episode ends on a new sign; the shock is saved for the last | Short Drama |

## FAQ

### Can I make an AI Halloween video with SceneMixer?

Yes. Make it as a one-minute horror short: one place with a rule, two escalations and a shock, in 9:16 for TikTok, Reels and Shorts, with a Horror cover poster from step 2.

### How long does a Halloween horror short take to make?

A few evenings for a one-minute short: the script and read on one, the cast and storyboard on the next, the video on the third. Each segment on SceneMixer usually takes ten to twenty minutes to render, and you can close the page while it does.

### Can AI make a horror short that is actually scary?

Yes, if the scare is built from sound, light and timing rather than from gore. Put the noise in the shot's Ambient line, give each dark shot one light source, and cut before the reveal.

### Why are my dark scenes grey and muddy?

Nothing in the shot names a light. Write the source and where it falls: a flashlight beam, an exit sign, lightning, a phone screen. The darkness is then whatever it does not reach.

### Can a horror video show blood or bodies?

SceneMixer blocks gore and mutilation. Show the aftermath indirectly: a shoe at the bottom of the stairs, a dropped flashlight, a scream from off-screen.

### How do I make a ghost that is only heard?

Set the character to Voice only (off-screen) in SceneMixer's step 2. No image is generated for them and their face is never attached to a shot; their lines are still spoken in their voice.

### How do I show a possession?

Make the possessed look a state of the character with its own image and pick it in the shot where it happens. The rest of the episode keeps the normal face.

### Should a horror short be Film or Short Drama?

Film for a single piece with a slow build and no episode cap; Short Drama for a series with a hook at the end of every episode.

### Is black and white a good choice for horror?

Often. Black & White Film hides imperfections in the dark and gives a classic feel. Like every visual style, SceneMixer fixes it for the project once you continue past step 1.

### Can I remake a famous horror film?

No. Known films and characters trigger the IP check on SceneMixer and are not allowed without rights. Write your own place, rule and creature.

Updated November 3, 2026


---

# AI Isekai Anime: Two Worlds, One Hero, Consistent Looks

URL: https://scenemixer.com/guides/make-an-ai-isekai-anime/
Make an isekai anime with AI: the other world as the project's look, the old world through locations and states, the transfer moment, and a hero who stays on model.

# How to make an AI isekai anime

Quick answer: Isekai is a hero pulled from an ordinary life into another world. In SceneMixer, choose an anime look in Step 1, make the other world the main setting (Elven Forest, Medieval Castle, Dungeon), keep the ordinary world to a few locations, give the hero an everyday state and an adventurer state, and treat the transfer as one strong sequence. Menus and status screens are spoken, not drawn.

Brief for AI Create: An isekai anime: a night-shift convenience store clerk wakes up as the shopkeeper of a village in a fantasy world where nobody has invented money; ten episodes, cozy, one dragon customer.

## Two worlds in one project

The visual style is fixed for the whole project, so choose it for the other world: Anime, Painterly Anime or Watercolor Anime. The ordinary world appears through a few locations (the apartment, the shop) and the hero's everyday state.

| Kind of world | Scene Style | Costume Style |
| Sword and sorcery | Medieval Castle, Dungeon | Fantasy Adventurer |
| Forest kingdom | Elven Forest | Elven Attire |
| Magic school | Magic Academy | Academy Robes |
| Cozy village | Pastoral Healing | Eastern European Folk |
| Dark lord's realm | Dark Asura | Orc Tribal, Dwarven Forge |

## The transfer sequence

| Beat | Shot to adapt |
| The old life | [Shot 1] 3s: Medium shot, 3 a.m., a convenience store. @Haru stacks rice balls, exhausted. |
| The light | [Shot 2] 2s: The automatic doors open on white light instead of the street. |
| The new world | [Shot 3] 4s: Wide shot. @Haru, now in adventurer state, stands in a village square under two moons. |
| The first customer | [Shot 4] 3s: A small dragon lands on the counter. @Ember: One of those triangle things, please. |
States.

## Status screens and skills

Video models are unreliable at drawing readable menus. Let a system voice or the hero say the stats; show skills as light and effect around the character. Text on screen · Voice-over lines.

## Isekai shapes

- The overpowered hero: the joke is how easy it is.
- The cozy isekai: a shop, a farm, a kitchen in another world.
- Reincarnated as something odd: a slime, a sword, a vending machine; a non-human character with a voice.
- The villainess: waking up as the doomed rival in a story.
Anime series for YouTube · AI anime videos.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Pastoral Healing, Medieval Castle; Costume Style Fantasy Adventurer; Visual Style Anime, Painterly Anime; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: an everyday and an adventurer state for the hero; the old shop and the village as locations; non-human customers as characters. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Old world habit | The ending beat |
| 1 | Night shifts | The dragon pays in gold scales |
| 2 | Price tags | The village elder wants to ban them |
| 4 | Loyalty cards | The king hears of the shop |
| 7 | Two-for-one deals | A rival merchant from the capital |
| 10 | Closing time | A door back home opens |

## FAQ

### How do I make an isekai anime with AI?

Choose an anime look, set the other world as the main setting, keep the old world to a few locations, and give the hero two states.

### How do I show the two worlds in one style?

In SceneMixer, the style is fixed per project; the worlds differ by locations, costumes and light.

### Can AI draw status screens?

Not reliably. Have a system voice or the hero read the stats.

### Can the hero be reincarnated as a monster or object?

Yes, as a non-human character with its own sheet and voice.

### Can the dialogue be Japanese?

Yes; set it in Step 1.

### Can I copy a famous isekai?

No. Known characters and worlds trigger SceneMixer's IP check.

### Vertical or landscape?

Landscape for YouTube, vertical for Shorts.

### How long should episodes be?

Two to five minutes.

Updated November 3, 2026


---

# AI Joseon Palace Drama: Hanbok, Royal Court and Slow-Burn Intrigue

URL: https://scenemixer.com/guides/make-an-ai-joseon-palace-drama/
A Joseon palace drama with AI: the Joseon Korea scene style, hanbok worn by rank, Korean dialogue, and court intrigue told without showing the poison or the punishment.

# How to make an AI Joseon palace drama

Quick answer: A Joseon palace drama stands on two looks: the palace and the clothes. In Step 1 pick the scene style Joseon Korea — palaces and scholar houses of Hanyang, with grey tile roofs and gently upturned eaves, beams painted in red, green and blue, paper-screen doors and stone terraces — and the costume style Hanbok. Set the dialogue language to Korean, and Target Region to East Asia or Follow script. Then do what the genre does best: write rank into the clothes, keep each segment to one room or courtyard, and let the poison, the beating and the verdict happen just off screen.

Brief for AI Create: A Joseon palace drama: a young court lady who serves tea to the queen mother notices the cup has been changed; she tells no one, and the crown prince, who has been avoiding her, is the only one who sees her notice; 8 short episodes, Korean dialogue, every episode ends on a closed door.

## The choices that make it Joseon

| Step 1 setting | Choose | Why |
| Scene Style | Joseon Korea | Hanyang palaces and scholar houses: grey tile roofs with upturned eaves, painted beams, paper doors, stone terraces, a wide front courtyard. |
| Costume Style | Hanbok | Short jacket over a high-waist full skirt for women, wide robes for men, soft bright colours. |
| Visual Style | Shot on Film, Soft Glow or Digital HD | Live action with a film grade or a soft glow suits a court drama; Korean Webcomic or Watercolor Anime if you want a drawn look. |
| Target Region | East Asia or Follow script | East Asia gives a Korean cast and names; Follow script keeps exactly what your script says. |
| Dialogue language | Korean | The lines are spoken in Korean as each clip renders. English lines work too. |
These are read into the descriptions when the script is read, so choose them before that. Why they are fixed afterwards · Every scene and costume style.

## Write rank into the clothes

Hanbok is the wardrobe; rank is what you add to each character’s clothing description in Step 2. A few forms that read clearly on screen:

- The king: a crimson robe with gold dragon roundels on the chest, shoulders and back, and a black court hat.
- A civil official: a round-collared robe with a crane badge on the chest and a black hat with a pair of wings at the sides. A military official wears a tiger on the badge.
- A court lady: a short jacket over a long full skirt in the colour of her office, hair in a low braided chignon with a pin.
- A scholar: a plain pale robe and a black horsehair hat with a brim.
A prince who walks the city in disguise is one character with a second look, not a second character: make the commoner’s clothes a state of the same sheet so the face never changes. Costume changes and scene states.

Court titles pile up — Your Majesty, a rank, a given name. Tie each title to its person in the script the first time it appears (“the Queen Mother, Lady Han”). When the script is read, the card keeps the other names your text uses for the same person on its grey Also called line (read-only; there is nothing to type), so a later “the Queen Mother” still finds her. Also called.

## One room, one courtyard, one segment

Court drama is staged in a handful of places, and the system likes that: each segment happens in one place and one stretch of time. Give every place its own scene card in Step 2.

| Place | What the segment is for |
| The throne hall | Audiences, edicts, officials kneeling in two rows. |
| The queen mother’s quarters | Tea, quiet orders, someone being watched. |
| A garden pavilion | The meeting nobody is supposed to see. |
| The kitchen or the medicine office | Where the cup, the soup or the draught is made. |
| The back gate at night | Leaving, arriving, being followed. |
Moving from the hall to the garden is a new segment, not a cut inside one. A segment stays in one place and one timeline.

## Intrigue without the gore

Poison, flogging and execution are the engine of palace plots, and the part the video checks will stop. Tell them the way the genre does: the cup before, the empty cup after.

- The poison: the bowl set down, the quiet look, the closed door; then the morning after.
- The punishment: the sentence read in the courtyard, the figure walked out of frame, the empty mat.
- The verdict: a faint sound off screen, a face turning away, a page of the edict dropped.
Invent your court. Characters from a known drama can trigger the IP check, and an original king with an original name keeps the story yours. If a clip is blocked.

## A court scene, shot by shot

| Beat | Shot to adapt |
| The courtyard | [Shot 1] 4s: Wide shot, fixed camera. A palace courtyard at dusk, grey tile roofs and painted beams, lanterns being lit one by one. @Yeon-hwa crosses the stone terrace carrying a covered tray. |
| The door | [Shot 2] 3s: Medium shot. A court lady slides the paper door open for @Yeon-hwa without looking at her. |
| The tray | [Shot 3] 3s: Close-up on @Yeon-hwa’s hands setting the tray before @Dowager. The lid is lifted; steam rises. |
| The line | [Shot 4] 4s: Close-up on @Yeon-hwa, eyes lowered. @Yeon-hwa (steady, Korean): 전하, 탕약을 드시옵소서. (voice: Yeon-hwa) |
| The pause | [Shot 5] 3s: Close-up on the cup in front of @Dowager. She does not reach for it. The paper door behind her darkens as someone passes. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- The court lady who saw it: she knows something; the question is who else knows she knows.
- The crown prince and the one he cannot choose: duty on one side of the courtyard, a person on the other.
- The royal physician: a cure that is also a weapon.
- The disguised royal: a prince among commoners, a clock ticking on how long it lasts.
- A succession in doubt: two heirs, one court, every episode a quiet vote.
The wider Korean drama picture, including modern stories, is in How to make an AI K-drama. For other palaces: Ottoman palace drama · Period romance.

## Make it in SceneMixer, start to finish

- Script: paste your own (Korean or English) on Upload Script, or write the brief in AI Create.
- Step 1: Type Short Drama or TV Series; scene style Joseon Korea; costume style Hanbok; a visual style; Target Region; Korean dialogue; 9:16 Portrait or 16:9. Read the script.
- Step 2: give each character a specific face, an age and the clothes of their rank; add a second state for any disguise. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). One place per segment.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | The queen mother’s tea. | The cup is not the one she brought. |
| 2 | The court lady says nothing. | The crown prince has been watching from the garden. |
| 3 | The medicine office. | Someone has copied the physician’s seal. |
| 4 | An audience in the throne hall. | The prince speaks for her, and regrets it at once. |
| 5 | The back gate at night. | A second lantern is lit behind her. |

## FAQ

### How do I make a Joseon-era palace drama with AI?

In SceneMixer's Step 1, choose the scene style Joseon Korea and the costume style Hanbok, set the dialogue language to Korean, then write or generate the script and render episode by episode.

### Which scene style should I use for a Korean palace?

Joseon Korea: palaces and scholar houses of Hanyang with grey tile roofs, painted beams, paper doors and stone terraces.

### How do I show rank with hanbok?

Hanbok is the wardrobe. Write the rank into each character's clothing description: a king's crimson dragon-roundel robe, a civil official's crane badge, a court lady's jacket and skirt.

### Can the characters speak Korean?

Yes. Choose Korean as the dialogue language in Step 1 and the lines are spoken in Korean on screen.

### Can I show a poisoning or an execution?

Not the act itself. Show the moments before and after: the cup, the closed door, the empty mat. The video checks stop graphic violence.

### Can I use a real Joseon king or a famous drama's character?

Use invented characters. Characters from known dramas can trigger the IP check, and an original king keeps the story yours.

### How is this different from the K-drama guide?

The K-drama guide covers modern and Joseon stories in broad strokes. This one goes into the palace, rank in clothing and court intrigue.

### Will the court lady look the same in every episode?

Yes. Each character is drawn once on SceneMixer and that sheet is used in every shot. A disguise is a second look of the same character.

Updated November 3, 2026


---

# AI K-Drama: How to Make a Korean Drama Series with AI, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-k-drama/
A Korean drama made with AI: a modern romance or a Joseon-era story, Korean dialogue spoken on screen, hanbok and palace sets, and the slow-burn K-drama rhythm.

# How to Make an AI K-Drama, Modern or Joseon

Quick answer: An AI K-drama needs three things to feel Korean rather than generic: Korean faces, names and places; Korean dialogue spoken by the characters on screen; and the slow-burn rhythm — long looks, rain, a hand almost taken. In SceneMixer you set the target region to East Asia and the dialogue language to Korean in step 1, choose modern sets (Neon Nights, Corporate Cold, Korean High School) or a Joseon court (Hanbok costumes, the Joseon Korea set), and every episode is rendered with the same cast and joined into one video. Lines can also be in English for an international audience.

Brief for AI Create: A modern K-drama: a chaebol heir who has lost his memory works at a small Seoul bakery, unaware the owner was his first love; 10 short episodes, Korean dialogue, every episode ends on a look.

## Modern or Joseon

| | Modern Seoul | Joseon era |
| Scene Style | Neon Nights, Corporate Cold, Korean High School, Urban Opulence | Joseon Korea |
| Costume Style | Smart Casual, Power Tailoring, Korean School Uniform | Hanbok |
| Visual Style | Digital HD, Soft Glow | Shot on Film, Soft Glow |
| Typical story | Chaebol romance, office, campus, revenge | Palace intrigue, a disguised royal, forbidden love |
For a webtoon look rather than live action, pick Korean Webcomic as the visual style. The visual style is fixed after step 1. Visual style presets.

Visual Style: Soft Glow

Visual Style: Korean Webcomic

Scene Style: Neon Nights

Scene Style: Corporate Cold

Scene Style: Joseon Korea

Scene Style: Korean High School

Costume Style: Hanbok
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Korean faces, names and dialogue

- Target Region: East Asia. The reading casts names, faces and places for that audience. Target region.
- Dialogue language: Korean. The lines are performed in Korean as the shot renders, with no dubbing pass. Native dialogue vs dubbing.
- Or English lines with a Korean world. Keep East Asia as the region and choose English: a K-drama for an international audience.
- A second language version of the same drama can be made from the finished project. Second language version.

## The K-drama rhythm, shot by shot

K-dramas take their time where other short dramas rush: the look held a second longer, the rain, the umbrella offered, the first hand-hold as an episode’s climax. Write those beats as their own shots and give them seconds.

| Beat | Shot to adapt |
| The rain | [Shot 1] 4s: Wide shot, fixed camera. A narrow Seoul alley at night in the rain, a convenience store glowing. In frame only: @Ji-woo, no umbrella, holding a cake box. Ambient: heavy rain. |
| The umbrella | [Shot 2] 3s: Medium shot. An umbrella appears over her. @Min-jun beside her, looking straight ahead. No dialogue. Ambient: rain on the umbrella. |
| The line | [Shot 3] 3s: Close-up on @Ji-woo, turning to him. @Ji-woo (quiet, Korean): 왜 여기 있어요? Ambient: rain. |
| The look | [Shot 4] 3s: Close-up on @Min-jun. He does not answer. No dialogue. Ambient: rain softening. |
How to write one shot · Hooks and cliffhangers.

## Joseon: hanbok and palace

The palace itself is the Joseon Korea scene style (grey tile roofs with gently upturned eaves, painted beams, paper doors, stone terraces). For a school story use Korean High School with Korean School Uniform. Deeper guides: Joseon palace drama · High school drama.

Pick Hanbok as the costume style and write rank into each character’s clothing: a deep blue official’s robe with a square embroidered crane on the chest, a black gat hat for a scholar-official; a pale pink jeogori and a long red chima, hair in a braided bun with a jade pin for a court lady. A disguised royal is one character with a second state (commoner clothes), so the face stays the same. Costume changes and states · Period romance.

Character sheets from the sample 윈터 프라미스: one reference image per character, used in every shot they appear in.

## Story shapes that work

- Chaebol romance: the heir and the ordinary woman; every episode a class wall, every episode a crack in it.
- Amnesia and first love: one of them remembers; the audience knows.
- Contract marriage: the rules are written in episode 1 and broken one by one.
- Palace intrigue: a crown prince, a poison, a court lady who saw it.
- Revenge: she returns years later under a new name. Revenge drama.

## Make it in SceneMixer, start to finish

- Script: your own on Upload Script (Korean or English), or a brief in AI Create.
- Step 1: Type Short Drama or TV Series; region East Asia; dialogue Korean or English; the scene, costume and visual styles above; 9:16 Portrait for feeds or 16:9 for YouTube. Read the script.
- Step 2: give each lead a specific face and one mark; regenerate any generic sheet. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project); give the slow beats their own shots.
- Video: generate on the default tier at 10 credits a second; composite; post.
For a Korean-language sample, watch 윈터 프라미스.

| Episode | The scene | The ending beat |
| 1 | A stranger with no memory is hired at the bakery. | He knows how she takes her coffee. She never told him. |
| 2 | His family is searching for him. | His photo on the news; she turns the TV off. |
| 3 | The rain and the umbrella. | “Why are you here?” He does not answer. |
| 4 | His fiancée finds the bakery. | “You don’t remember me either?” |
| 5 | A box of old letters. | Her handwriting. From ten years ago. |

## FAQ

### How do I make an AI K-drama?

Set the target region to East Asia and the dialogue language to Korean in SceneMixer's step 1, choose modern or Joseon scene and costume styles, write or generate the script, and render episode by episode.

### Can the characters speak Korean?

Yes. Choose Korean as the dialogue language and the lines are performed in Korean on screen; English lines in a Korean setting work too.

### Can I make a Joseon-era historical K-drama?

Yes. Pick Hanbok as the costume style and a palace scene style, and write each character's rank into their clothing.

### Can it look like a webtoon?

Yes. Choose Korean Webcomic as the visual style in step 1; the style is fixed after that.

### Will the leads look the same in every episode?

Yes. Each character is drawn once on SceneMixer and that sheet is used in every shot; a disguise or a later look is a state of the same character.

### Can I make an English version too?

Yes. A second language version can be made from the finished project on SceneMixer.

### How long should an episode be?

One to three minutes for feeds; longer for YouTube. Give the slow beats their own shots.

### Can I use a real K-drama's characters or actors?

No. Known characters and real actors are refused by the IP check; invent your own cast.

Updated November 3, 2026


---

# AI K-Pop Idol Drama: Trainees, Debuts and Backstage Secrets

URL: https://scenemixer.com/guides/make-an-ai-k-pop-idol-drama/
Make a K-pop idol or trainee drama with AI: Neon Nights and Korean High School settings, Streetwear and stage looks as states, dance practice and debut-stage shots.

# How to make an AI K-pop idol drama

Quick answer: An idol drama is a trainee, a debut date and a secret that could end it. In SceneMixer, set Seoul in Step 1: Neon Nights for the city after dark, Korean High School for trainees still at school, Corporate Cold for the agency. Streetwear dresses them for practice; the stage outfit is a state of each member. Make the group's members easy to tell apart with one mark each, and invent your group and your agency: real idols and labels trigger the IP check.

Brief for AI Create: A K-pop idol drama: a trainee who failed her final evaluation is secretly hired as the ghost singer behind a new girl group's lead vocalist; the group's manager starts to suspect, and so does the lead; 12 short episodes, each ends one day closer to the debut stage.

## Set Seoul and the agency

| Where | Scene Style | Costume Style |
| The city at night | Neon Nights | Streetwear |
| Trainees at school | Korean High School | Korean School Uniform |
| The agency | Corporate Cold | Power Tailoring for the executives |
| The dorm | Soft Romance | Loungewear |
| Award shows and red carpets | Urban Opulence | Haute Couture |
Set Target Region to East Asia for Korean faces and names, or Follow script. Target Region · K-dramas.

## Members, stage looks and practice rooms

- One mark per member. Five people in matching outfits blur together; give each a hair colour, a piercing or a scar. The same face in every shot.
- The stage look is a state. Practice clothes and the debut-stage outfit are two looks of one member. States.
- The practice room (mirror wall, wooden floor) is a location you will use in every episode. The same room in every episode.
- Dance works best as short shots of one or two members, not the whole routine in one take.

## The debut stage, shot by shot

| Beat | Shot to adapt |
| The wings | [Shot 1] 3s: Medium shot backstage, coloured stage light spilling through a curtain. @Hana adjusts an in-ear monitor, hands shaking. |
| The stage | [Shot 2] 4s: Wide shot, the stage from the crowd, five members in silver outfits in formation, light sticks waving in the background. |
| The booth | [Shot 3] 3s: Close-up on @Jiwoo in a dark sound booth, headphones on, singing. |
| The line | [Shot 4] 4s: Close-up on @Manager Seo. @Manager Seo (whispering): That's not Hana's voice. (voice: Seo) |
Music: every episode is scored for you automatically, under the voices; there is no track picker, so plan the stage scenes around faces and reactions rather than a particular song. Background music.

## Story shapes that work

- The ghost singer.
- The dating ban: an idol and a staff member.
- The survival show: one member is cut each episode.
- The fallen idol: a scandal, a comeback.
- The manager: the one person who knows all their secrets.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Korean High School; Costume Style Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: each member with one mark and a stage-outfit state; the practice room, the dorm and the stage as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The final evaluation. | She is cut, then called back at midnight. |
| 2 | The recording studio. | Her voice, someone else's face. |
| 3 | The dorm. | The lead finds her lyric notebook. |
| 4 | The practice room. | The manager plays two recordings side by side. |
| 5 | Backstage at the debut. | The lead's microphone goes dead. |

## FAQ

### How do I make a K-pop idol drama with AI?

Pick Neon Nights, Korean High School or Corporate Cold in SceneMixer's Step 1, Streetwear for practice, give each member a stage-outfit state in Step 2, and render one story beat per segment.

### Can I use a real K-pop group?

No. Real people and real labels cannot be depicted; invent the group, the agency and the songs.

### How do I keep five members apart?

Give each one a distinct mark, such as a hair colour or a piercing, on their sheet.

### Can I use a real K-pop song?

No. The episode's music is chosen and mixed by SceneMixer, with no track picker or upload, and real songs are someone else's IP. Build the stage scenes on faces and reactions.

### How do I show the stage outfits?

Add them as a state of each member and pick that state for the stage segments.

### Can the dialogue be in Korean?

Yes. Korean is one of the 15 dialogue languages on SceneMixer, set in Step 1.

### Which visual style suits an idol drama?

Digital HD or Soft Glow for live action, Korean Webcomic or Anime for a drawn version.

### Vertical or landscape?

9:16 Portrait fits idol content on short-video feeds.

Updated November 3, 2026


---

# AI Bedtime Story Videos for Kids: Make a Series, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-kids-bedtime-story-series/
Make bedtime story videos for children with AI: a storybook or 3D look, a gentle narrator, the same characters every night, slow calm shots and a quiet ending.

# How to Make an AI Kids’ Bedtime Story Series

Quick answer: A bedtime story series is the opposite of a drama for feeds: slow, warm, the same friendly characters every night, and an ending that winds down instead of a cliffhanger. In SceneMixer you write or generate the story, pick a children’s look in step 1 — Watercolor Storybook, Family 3D Animation, 3D Toy Cartoon or Claymation — and let a narrator carry it as voice-over while the characters say a few short lines. Each character is drawn once and stays the same in every episode, so a child meets the same little fox every night.

Brief for AI Create: Bedtime stories about a small fox who cannot sleep, one short story per night: each night a different forest friend shows him one calm thing; a soft narrator, slow shots, every episode ends with the fox asleep.

## Pick a look that is calm

| Visual Style | What it feels like | Good for |
| Watercolor Storybook | A picture book that moves | Ages 2–6, very gentle stories |
| Family 3D Animation | A family animated film | Adventures with a lesson, ages 4–9 |
| 3D Toy Cartoon | Toys come to life | Funny stories, animal casts |
| Claymation | Handmade, tactile | Short, cosy stories |
Scene Style: Pastoral Healing for meadows and farms, Elven Forest for an enchanted wood, Soft Romance for a cosy home. SceneMixer fixes the visual style after step 1. Visual style presets.

Visual Style: Family 3D Animation

Visual Style: 3D Toy Cartoon

Visual Style: Claymation

Scene Style: Pastoral Healing

Scene Style: Elven Forest
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## The shape of a bedtime episode

- One to three minutes. Long enough to settle, short enough for one night.
- One small problem, one small kindness. The fox cannot sleep; the owl shows him the stars.
- A quiet ending, not a cliffhanger. The last shot is the character asleep, the light going down, the narrator saying goodnight.
- The same opening and closing every night. Children like the ritual: the same first line, the same last image.

## Characters children recognise

Animal and toy characters are cards like any other. Describe what the character is, its colours and one mark that never changes: a small red fox cub with a white tip on his tail, oversized ears, a blue knitted scarf; walks upright; big round amber eyes. The scarf and the white tail tip are what keep him recognisable from shot to shot and night to night. Non-human characters.

Keep the cast small — the hero and two or three friends — and reuse them. Add new friends one at a time, in new episodes of the same SceneMixer project, so the cast and the forest carry over. Reuse a character across projects.

Character sheets from the sample Rocket Pup: one reference image per character, used in every shot they appear in. An animal or toy lead is a card like any other, with its own sheet and voice.

## A gentle narrator and slow shots

Add the narrator as an Off-screen character: it has a voice and no image. In its voice description, write how it should sound (a warm, soft, slow female voice, like a parent reading aloud); the characters’ own voices are described the same way on their cards. You can also pick a voice from SceneMixer’s library or upload your own on the voice field. Voice-only characters · Character voices.

| Beat | Shot to adapt |
| The ritual opening | [Shot 1] 5s: Wide shot, slow push in. A little den under an old oak at dusk, fireflies. In frame only: @Fox, curled up, eyes open. @Narrator (voice-over, soft, slow): In the old oak wood, a little fox could not sleep. Ambient: crickets. |
| A friend arrives | [Shot 2] 4s: Medium shot. @Owl lands on a branch above the den. @Owl (gentle, slow): Shall I show you something? Ambient: wings, crickets. |
| The goodnight | [Shot 3] 6s: Close-up, the light fading. @Fox asleep, tail over his nose. @Narrator (voice-over, whisper): Goodnight, little fox. Ambient: crickets fading. |
Slow shots are written, not added later: say slow push in, fixed camera, and give each shot a few more seconds than a drama would. How to write one shot.

## Keep it safe and calm

- No real scares. Darkness is cosy, not threatening; no sudden loud sounds in the Ambient line.
- No known characters. Characters from films, games and books are refused by the IP check; invent your own. IP check.
- Language: choose the dialogue language in step 1 and the narrator reads in it; 15 languages are available.

## Make it in SceneMixer, start to finish

- Script: your own story on Upload Script, or a brief in AI Create naming the hero, the friends and the nightly ritual.
- Step 1: Type Short Drama or TV Series; a calm visual style; 16:9 for a tablet or TV, 9:16 Portrait for a phone; the dialogue language. Read the script.
- Step 2: check each character has its colours and one mark; add the narrator if the reading did not. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project); slow shots, a voice-over opening and goodnight.
- Video: generate on the default tier at 10 credits a second; composite.
For a family 3D look, watch Rocket Pup.

| Night | The small problem | How it ends |
| 1 | The fox cannot sleep. | The owl shows him the stars; he falls asleep counting. |
| 2 | The fox is afraid of the rain. | The badger shows him the drops on the leaves; asleep. |
| 3 | The fox misses the sun. | The moon is a night-time sun; asleep. |
| 4 | The fox lost his scarf. | A friend has kept it warm; asleep. |
| 5 | The fox wants to stay up. | He tells the story this time; asleep mid-sentence. |

## FAQ

### How do I make AI bedtime story videos?

Write a short story or give AI Create a brief, pick a calm style such as Watercolor Storybook or Family 3D Animation in SceneMixer's step 1, add an off-screen narrator, and generate one episode per night.

### Can I make kids' story videos for YouTube?

Yes. Pick 16:9 in SceneMixer's step 1 for regular videos or 9:16 for Shorts, and post each composited episode.

### Will the same character appear every night?

Yes. SceneMixer draws each character once as a reference sheet and uses it in every shot it appears in; add new nights to the same project and the cast carries over.

### Can the narrator sound like a parent reading?

Describe the voice on the narrator's card, for example a warm, soft, slow voice; you can also pick a library voice or upload your own.

### How long should a bedtime episode be?

One to three minutes: one small problem, one kindness, and a quiet ending.

### Can I use characters from my child's favourite show?

No. The IP check on SceneMixer refuses known characters. Invent your own: the colours and one mark make them memorable.

### Which language can the story be in?

Any of the 15 dialogue languages on SceneMixer; choose it in step 1.

### Is it free to try?

The first project includes free allowances, and the first 15 setup images on a SceneMixer account are free; the generate button shows the price before anything is charged.

Updated November 3, 2026


---

# AI Korean Webtoon Video: Make a Webtoon-Style Series With Voices

URL: https://scenemixer.com/guides/make-an-ai-korean-webtoon-style-series/
Make a Korean webtoon-style video series with AI: the Korean Webcomic look, vertical 9:16, panel-to-shot rules, a closing panel per chapter, Korean sets and wardrobes.

# How to make an AI Korean webtoon-style series

Quick answer: SceneMixer has a visual style called Korean Webcomic: slender, polished characters with soft airbrushed shading. Pick it in Step 1 with 9:16 Portrait, write each chapter as an episode that ends on a closing panel, and the cast is drawn once and speaks in any of 15 languages, Korean included. It does not animate drawn comic pages and it is not a copy of any artist's style: you write the story as a script and SceneMixer makes the moving, voiced version in that look.

Brief for AI Create: A webtoon-style romance fantasy: a palace seamstress in a Joseon-inspired kingdom is mistaken for a noblewoman at the king's banquet and has three days to keep the secret; eight episodes, each ends on a closing panel; soft, airy, a little funny.

## The look and where it fits

Korean Webcomic is one of the 2D animation styles: slender, polished characters, soft airbrushed shading. It is a look, not a script format, so it works for any genre a webtoon covers. The pairing below uses real Scene Style and Costume Style cards:

| Webtoon genre | Scene Style | Costume Style | A chapter engine |
| School romance | Korean High School, Campus Youth | Korean School Uniform | A note, a misunderstanding, a rooftop |
| Elite academy | Elite Boarding School | Prep School Uniform | A ranking list and a rival |
| Office romance | Corporate Cold, Urban Opulence | Smart Casual, Power Tailoring | A late file and a shared lift |
| Palace romance fantasy | Joseon Korea | Hanbok | A secret identity and a banquet |
| City thriller romance | Neon Nights | Streetwear | A message that arrives one minute early |

Scene Style: Korean High School

Scene Style: Joseon Korea

Costume Style: Hanbok
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
All visual styles · Scene and costume styles.

## Panels become shots

Write the chapter the way a webtoon reads, then translate each device into a shot:

| Webtoon device | In the video | How to write it |
| Tall establishing panel | A wide shot held for 4 seconds | Wide shot, slow push-in. Name the place and the light. |
| Blank gap between panels | A held beat with no dialogue | A shot with ambient sound only, 2 to 3 seconds |
| Close-up on the eyes | A reaction shot | Close-up on @Name; say what the face does, not what it feels |
| Speech bubble | A spoken line | @Name (tone): the line (voice: Name) |
| Thought bubble | A voice-over or a line said aloud | Give it to an Off-screen narrator or let the character say it |
| Sound effect lettering | The ambient line | Ambient: a door bursting open |
| The last panel of the chapter | The final shot, held 2 to 3 seconds | End on the image that raises the question |
Plan two to four shots for each screen of scroll. Writing dialogue in the storyboard · How to write one shot.

## One chapter, one episode, one closing panel

A webtoon chapter is a short read that ends on a pull. A Short Drama episode on SceneMixer, up to 240 seconds, is the same shape: hook, a turn, a closing panel that makes the viewer swipe to the next. A 90-second chapter is six segments of 15 seconds, roughly 18 to 24 shots. Hooks and cliffhangers · AI hook generator.

| Part | Seconds (90 s chapter) | Does |
| Opening panel | 0 to 8 | The place and one wrong detail |
| First beat | 8 to 30 | The want, said or shown |
| Complication | 30 to 60 | Someone or something blocks it |
| Reaction | 60 to 80 | A close-up held without a line |
| Closing panel | 80 to 90 | The image or line that asks the next question |

## Vertical, Korean dialogue and voices

- Vertical. Choose 9:16 Portrait in Step 1 and the video fills a phone the way a webtoon does. Vertical or landscape.
- Korean dialogue. Korean is one of the 15 dialogue languages; pick it in Step 1 and write the lines in Korean. Keep one speech level between each pair of characters (polite or casual), so the voices stay consistent. Choosing the dialogue language.
- Voices. Each character gets a voice; play the library before choosing. Character voices.
- Subtitles. For viewers who do not read Korean, accounts with a purchase can add burned-in captions or an SRT file. Add subtitles.

## What this is not

Comic pages are not an input: SceneMixer reads a script, not artwork, so a finished webtoon has to be written out as scenes first. If you own a webtoon and want to keep its characters, upload your designs as reference sheets (a membership feature). Known webtoon characters are refused by the IP check. Turn a manga or webtoon into an anime series · An AI K-drama · A Joseon palace drama · A shojo romance anime.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Joseon Korea; Costume Style Hanbok; Visual Style Korean Webcomic; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the seamstress, the king and the noblewoman she is mistaken for, plus the palace and the sewing room; each is reused in every chapter. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Chapter | Closing panel |
| 1 | The seamstress is handed the noblewoman's invitation by mistake |
| 2 | At the banquet, the king asks for her name and she gives the other woman's |
| 3 | The real noblewoman is seen in the palace gate |
| 4 | A guard recognises the seamstress's needlework on the king's robe |
| 5 | The king asks her to dance |
| 6 | The two women meet in the sewing room, alone |
| 7 | The noblewoman offers a trade |
| 8 | The seamstress chooses her own name |

## FAQ

### Is there a Korean webtoon style in SceneMixer?

Yes. Korean Webcomic is one of the 21 visual styles: slender, polished characters with soft airbrushed shading.

### Can I make a webtoon-style video series with AI?

Yes. Write each chapter as an episode, choose Korean Webcomic and 9:16 Portrait in Step 1, draw the cast once and render the episodes in order.

### Can it animate my webtoon pages?

No. Comic pages are not an input; write the chapter as a script, with dialogue from your speech bubbles, and upload your own designs as reference sheets if you are a member.

### Can the characters speak Korean?

Yes. Korean is one of the 15 dialogue languages SceneMixer offers.

### Which format suits a webtoon-style video?

9:16 Portrait, with a Short Drama episode of 60 to 120 seconds ending on a closing panel.

### Can I copy a famous artist's style?

No. Korean Webcomic is a general look; known characters and franchises are refused by the IP check.

### How do I show thoughts?

Give them to an off-screen narrator as voice-over, or let the character say the line aloud.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; the first storyboard on your first drama is free.

Updated November 3, 2026


---

# AI Kung Fu Drama: Martial Arts Fights, Masters and Schools

URL: https://scenemixer.com/guides/make-an-ai-kung-fu-martial-arts-drama/
Make a kung fu drama with AI: Retro Hong Kong streets or Ink Jianghu bamboo, one technique per shot, training as states, and the master and student at the heart of it.

# How to make an AI kung fu or martial arts drama

Quick answer: A kung fu story is a student, a master and a fight that means something. In SceneMixer, choose the era in Step 1 (Retro Hong Kong or Republican Era for streets and tea houses, Ink Jianghu or Dark Wuxia for the old martial world), write every technique as its own shot, and show training as states of the student. Wounds stay off screen.

Brief for AI Create: A kung fu drama: a delivery boy in 1980s Hong Kong is taught by the old woman who runs the noodle shop, who was once the best fighter in the city; ten episodes, one rooftop duel.

## Choose the martial world

| Era | Scene Style | Costume Style |
| Neon city streets | Retro Hong Kong | Streetwear, Sportswear |
| Old Shanghai schools | Republican Era | Republican Modern |
| The poetic jianghu | Ink Jianghu | Jianghu Travelwear |
| The grim jianghu | Dark Wuxia | Jianghu Travelwear |
| Temple training | Tibetan Plateau or Classical Elegance | Taoist & Monk Robes |
For flying immortals and sects, see xianxia and wuxia.

## One technique per shot

Name the move and give it its own seconds: a low sweep, a palm strike, a block with the forearm, a roll. Then the reaction. The rhythm of a kung fu fight is in the cuts, not in one long take. Fights and group scenes.

| Beat | Shot to adapt |
| The stance | [Shot 1] 3s: Wide shot, a rooftop at dusk, laundry lines. @Ming and @Rival face each other in stance. |
| The attack | [Shot 2] 2s: Medium shot. @Rival steps in with a straight punch. |
| The block | [Shot 3] 2s: Close-up. @Ming's forearm turns the punch aside. |
| The counter | [Shot 4] 2s: Medium shot. @Ming's palm stops an inch from @Rival's chest. |
| The lesson | [Shot 5] 3s: @Granny Lau at the roof door, arms folded. @Granny Lau: Again. Slower. |

## Training is a state

The student changes: bandaged hands, a training uniform, a sharper look by the final episode. Add each as a state and pick it in the right segments. A master's old fighting outfit is a state too, saved for the moment she fights again. States.

## Story shapes

- The hidden master: someone ordinary was once the best.
- The tournament: one opponent per episode.
- The school under threat: a rival school wants the building.
- Revenge with a lesson: the master teaches restraint.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Retro Hong Kong, Ink Jianghu; Costume Style Streetwear, Jianghu Travelwear; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: a training state for the student and a fighting state for the master; the noodle shop and the rooftop as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post. Choose Chinese dialogue for a Hong Kong story.

| Episode | Lesson | The ending beat |
| 1 | Balance | The noodle-shop owner stops a thief with one finger |
| 2 | Patience | A rival school's challenge |
| 3 | Defence | Ming loses in front of everyone |
| 4 | The master's past | Her old opponent arrives |
| 5 | Restraint | The rooftop duel |

## FAQ

### How do I make a kung fu drama with AI?

Pick an era in SceneMixer's Step 1, write each technique as its own short shot, and show training through states of the student.

### Can AI do martial arts fights?

Yes, one move per shot; long choreography in a single take is where models fail.

### Which style suits kung fu?

Shot on Film for a seventies feel, Digital HD for modern, Ink Wash Animation or Chinese 3D Animation for animated.

### Can I show injuries?

No. Show the hit stopping, the fall, and the aftermath; graphic violence is blocked.

### Can the dialogue be in Cantonese?

No. Dialogue on SceneMixer is performed in Mandarin Chinese or any of the other 14 languages; Cantonese is not a dialogue option. Put the Hong Kong flavour into names, places and the voice descriptions.

### Is this different from wuxia?

Yes: grounded martial arts rather than flying sects; there is a separate wuxia guide.

### Can I copy a famous kung fu star?

No. Real people's likenesses are refused; invent your fighters.

### How long should fight shots be?

Two to three seconds each.

Updated November 3, 2026


---

# Make an AI Language-Learning Drama: Short Scenes in 15 Languages

URL: https://scenemixer.com/guides/make-an-ai-language-learning-drama/
Make language-learning video dramas with AI: everyday scenes, short repeatable lines, dialogue in any of 15 languages, subtitles with SRT, and a second-language version.

# How to make an AI language-learning drama

Quick answer: Learners remember language in situations: ordering coffee, asking the way, apologising for being late. In SceneMixer, write each episode as one everyday scene with short lines and a phrase that comes back, choose the dialogue language in Step 1 (any of the 15 interface languages), and add subtitles in the script's words. A version in another language is a second project from the same script, with the same faces.

Brief for AI Create: A language-learning drama in Spanish for beginners: a shy exchange student in Seville gets one small task a day from her landlady; each episode is one situation (the market, the bus, the pharmacy), short lines, and one phrase repeated three times.

## Write for learners

- One situation per episode. The market, the bus, the doctor.
- Short lines, said slowly by the situation: one sentence per shot gives each line its own time. Shot timing.
- Repeat the key phrase three times in the episode, in different shots.
- Write exactly what should be said, including accents and punctuation. Dialogue.

## Languages, subtitles and versions

- Dialogue language is set per project in SceneMixer's Step 1: English, Chinese, Portuguese, Korean, Spanish, Japanese, Indonesian, Russian, French, Italian, Dutch, Arabic, Turkish, German or Ukrainian. Dialogue language.
- Subtitles come from your script's words, burned in or as SRT, on accounts with a purchase. Subtitles.
- A second language is a second project with the same cast. Second-language version.
- Accents can be described per character. Give a character an accent.

## The market, shot by shot

| Beat | Shot to adapt |
| The stall | [Shot 1] 4s: Wide shot, a sunny market street, fruit stalls. @Emma walks up to a stall with a basket. |
| The phrase | [Shot 2] 4s: Close-up on @Emma. @Emma (nervous): Hola. Quiero un kilo de naranjas, por favor. (voice: Emma) |
| The reply | [Shot 3] 4s: Medium shot. @Don Pepe smiles. @Don Pepe (warm): ¿Un kilo? ¡Claro! (voice: Pepe) |
| The repeat | [Shot 4] 4s: Close-up on @Emma, more confident. @Emma: Y un kilo de limones, por favor. (voice: Emma) |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Mediterranean Coast, Latin Colonial Town; Costume Style Soft Everyday; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the learner and the locals with one mark and a voice each; the market, the bus and the café as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Situation | Key phrase |
| 1 | The market. | “Quiero un kilo de…, por favor.” (I would like a kilo of…, please.) |
| 2 | The bus. | “¿Este autobús va a…?” (Does this bus go to…?) |
| 3 | The pharmacy. | “Me duele la cabeza.” (My head hurts.) |
| 4 | The café. | “Para llevar, por favor.” (To take away, please.) |
| 5 | The landlady. | “Perdón, llego tarde.” (Sorry, I am late.) |

## FAQ

### How do I make a language-learning video with AI?

Write one everyday situation per episode with short lines and a repeated key phrase, choose the dialogue language in Step 1, and add subtitles.

### Which languages can the characters speak?

Any of 15 languages.

### Can I add subtitles?

Yes, from your script's words, burned in or as SRT, on SceneMixer accounts that have made a purchase.

### Can I make the same episode in two languages?

Yes. Make a second project from the same script with the other dialogue language and bring the cast across.

### Can I make characters speak slowly?

Give each line its own shot and keep lines short; each line needs time to be spoken.

### Can a character have an accent?

Yes. Describe it on the character.

### Vertical or landscape?

9:16 Portrait for short lessons on feeds, 16:9 Landscape for a course.

### How long should an episode be?

One to three minutes, one situation each.

Updated November 3, 2026


---

# AI Love Story Video: Two Leads, Their Places, One Choice

URL: https://scenemixer.com/guides/make-an-ai-love-story-video/
Make a romance or love story video with AI: two leads drawn once, the places that matter to them, staging that keeps them facing each other, and a shape to follow.

# How to make an AI love story video

Quick answer: A love story is two people and the distance between them. In SceneMixer, draw both leads once so they look the same in every scene, give the romance three or four places that matter (where they meet, where it goes wrong, where they choose), stage each scene by saying who faces whom, and pick a shape: slow burn, second chance, forbidden, or fake turning real. Romance on screen is glances, hands and closeness; sexual content is not allowed.

Brief for AI Create: A love story: a night-shift baker and the early-morning cyclist who buys the first loaf every day have never seen each other in daylight; ten short episodes, one bakery, one bridge.

## Pick the shape first

| Shape | What keeps them apart | Guide |
| Slow burn | Nothing but time and nerve | This page |
| Second chance | What happened last time | Second chance romance |
| Enemies to lovers | Pride and a rivalry | Enemies to lovers |
| Fake turning real | A deal both signed | Contract marriage |
| Forbidden | Family, rank or sides | Royal drama, mafia romance |
| Dark | One of them is dangerous | Dark romance |
| Workplace | Everyone is watching | Office romance |

The shape decides the ending of every episode: in a slow burn each one ends a little closer; in a second-chance story each one opens a little more of the past. Hooks and cliffhangers.

## Two people in one frame

Most romance scenes are two people talking. Say who faces whom and where they stand (@Ana at the counter, @Tom by the door, facing each other) and the staging keeps them on the same sides from shot to shot. Cut between close-ups for the confession; go back to the two-shot for the silence after it. When two shots of the same pair follow each other, one of them changes the framing or the angle, so a long conversation does not look like one shot that keeps jumping. Who is in frame · Why characters swap sides.

## The almost, shot by shot

| Beat | Shot to adapt |
| The shelter | [Shot 1] 4s: Wide shot, a narrow shop doorway in heavy rain. @Ana and @Tom squeeze in under the awning, shoulders almost touching. |
| The look | [Shot 2] 3s: Close-up on @Ana, rain on her lashes, looking up at him. |
| The line | [Shot 3] 3s: Close-up on @Tom. @Tom (quietly): I come at six because you are there at six. |
| The step back | [Shot 4] 3s: Two-shot. The rain stops. @Ana steps out into the street and does not look back. |

The interrupted moment is the oldest trick in romance, and it suits short episodes: the episode ends on the step back, the next one opens on what she does next. Rain and time of day come from the script for the scene. Time and weather.

## The look of a love story

Warm, soft and close. Soft Glow gives hazy highlights; Shot on Film gives grain and warmth; Watercolor Anime turns it into a drawn love story. For places, Soft Romance is cafes, gardens and lamplight; Parisian Streets and Mediterranean Coast are the classic backdrops.

Visual Style: Shot on Film

Visual Style: Watercolor Anime

Scene Style: Soft Romance

Scene Style: Parisian Streets

Scene Style: Mediterranean Coast
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## What stays off screen

Sexual content cannot be generated, and anything that sexualises a minor is zero tolerance. A love story does not need it: a hand held a second too long, a coat put around shoulders, a door left open. Content rules.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Soft Romance, Parisian Streets; Costume Style Soft Everyday, Smart Casual; Visual Style Soft Glow, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: both leads with an everyday state and one dressed-up state; the place they meet and the place they choose as locations; one object that passes between them as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Part | What happens | Where |
| Meet | A small accident that makes them notice each other | The place they meet |
| Want | A reason to keep coming back | The same place, again |
| Lose | A secret, a misunderstanding, a rival | Somewhere public |
| Choose | One of them crosses the distance | The place they met, changed |

## FAQ

### How do I make an AI love story video?

Draw both leads once, give the story a few places that matter, say who faces whom in each scene, and follow one romance shape from start to end.

### How do I keep the couple looking the same?

Each lead has one sheet used in every shot; add states for outfits that change, such as a wedding.

### Can the characters kiss?

Romance is shown through closeness, hands and glances; sexual content is blocked by SceneMixer. Keep intimate moments brief and simple.

### Which visual style suits romance?

Soft Glow, Shot on Film, or Watercolor Anime for a drawn love story.

### Vertical or landscape?

9:16 Portrait for a short-drama feed, 16:9 Landscape for a love story film.

### Can I write the romance in another language?

Yes. Choose the dialogue language and target region in SceneMixer's Step 1.

### How long should each episode be?

One to two minutes for short drama; the shape decides how many episodes you need.

### Can I base it on a real couple?

Not with real people's faces; invent the characters. Using someone's likeness needs their consent.

Updated November 3, 2026


---

# AI Mafia Romance Drama: Power, Loyalty and Threats Off Screen

URL: https://scenemixer.com/guides/make-an-ai-mafia-romance-drama/
A forced marriage, a family ledger and a man everyone fears: how to write a mafia romance short drama and generate it without the violence getting it blocked.

# How to Make an AI Mafia Romance Drama

Quick answer: A mafia romance is a love story inside a system of loyalty and fear: an arranged marriage or a debt brings a woman into a family where every room has rules, and the most dangerous man in it decides she is his to protect. Power on screen is shown through rooms, silence and obedience, not gunfire, which suits AI video: extreme violence is blocked, threat and aftermath are not. In SceneMixer, choose the city and the tailoring in step 1 (Neon Nights or New York Metropolis with Power Tailoring, or Jazz Age for a period story), let a gun be a prop that is seen rather than fired, and cut away at the moment a threat becomes an act.

Brief for AI Create: A mafia romance: a florist is married to the heir of a crime family to settle her father's debt; he promises never to touch her; his uncle wants the marriage to fail; she learns the family's ledger by heart; dark, slow-burn, protective.

## The genre’s engine

- The bargain. A debt, a marriage, a witness under protection. It forces two people from different worlds into the same house.
- The rules of the family. Who sits where, who speaks first, what is never said at dinner. Breaking a rule is a plot point.
- The feared man who is gentle with one person. The audience watches others flinch around him, then sees him careful with her. That contrast is the romance.
- The rival inside the family. An uncle, a brother, a consigliere with his own plans. The threat comes from the dinner table, not the street.
- Loyalty tested. Each arc ends with someone choosing a side.

## Settings to pick in step 1

- Visual Style: Shot on Film for deep shadows and a cinema grade; Digital HD for the bright app look; Black & White Film for a period crime story.
- Scene Style: Neon Nights for clubs and wet streets, New York Metropolis for the city, Urban Opulence for the family’s penthouse, Mediterranean Coast for the old-country villa.
- Costume Style: Power Tailoring for a modern family, Jazz Age or Vintage Hollywood for a 1920s or mid-century one.
- Type and Aspect Ratio: Short Drama and 9:16 Portrait for a feed series.
- Target Region: Western or Mediterranean; invent the family name and its city. Real crime organisations and real people are not material.

Scene Style: Neon Nights

Scene Style: New York Metropolis

Scene Style: Urban Opulence

Costume Style: Power Tailoring

Costume Style: Jazz Age
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Power without violence on screen

Every generation passes the provider’s content check, and extreme violence is blocked; a scene built on someone being shot will fight the check. The genre’s best moments have always been the ones where nothing happens and everyone is afraid:

| Instead of | Write |
| A shooting | He sets a pistol on the table between them, grip toward the other man, and waits. Cut. |
| A beating | Two men walk a third into the back room; the door closes; she hears nothing, which is worse. |
| A murder | The next morning, an empty chair at the family table. Nobody mentions it. |
| A threat | He straightens the other man’s tie, slowly. “Your daughter starts school on Monday, doesn’t she?” |
A gun can be an object in the scene — holstered, on a table, in a drawer she finds. Firing at a person and showing the wound is the kind of shot that may be blocked as extreme violence. If a segment is blocked, change its storyboard text before retrying; repeated blocks pause video generation on SceneMixer for 1 hour. Blocked by moderation: what to change.

## Rooms that tell you who is in charge

Mafia stories are told through a handful of rooms that recur: the long dinner table, the study with the ledger, the club, the car. Make each one a location with its own image so it is the same room every episode, and use where people stand in it to show rank: who sits at the head, who stays by the door. Write the light: a single lamp on the desk, the rest of the study dark; red neon from the club sign across the wet windscreen. Same room in every episode · Everyone stays put.

Character sheets from the sample Neon Debt: one reference image per character, used in every shot they appear in.

## Three shots to copy

| Beat | Shot |
| The dinner | [Shot 1] 5s: Opening: wide shot, fixed camera. The long table; @Don Matteo at the head; @Sofia at the far end, the only one not eating. @Don Matteo (mild, not looking up): Our guest isn’t hungry. Ambient: cutlery, a clock. |
| The protector | [Shot 2] 3s: Close-up. In frame only: @Luca. He sets his fork down; the table goes quiet. No dialogue. |
| The line | [Shot 3] 4s: Medium shot. @Luca stands behind @Sofia’s chair, one hand on its back. @Luca (quiet, to the table): She’s family now. Anyone who forgets that can say so. @Sofia lips closed, no lip movement. |
Nobody draws a weapon; everybody understands. How to write one shot.

## Rings, ledgers and the family crest

The objects of a crime family — a signet ring kissed in greeting, a ledger, a crest on the gate — recur and should look the same each time. Check they have cards and images in step 2 and name them with @ in the shots they appear in. Keep the crest invented: real logos and emblems trigger the IP check. Same prop in every hand.

## Make it in SceneMixer, start to finish

- Script: yours on Upload Script, or a brief in AI Create naming the bargain, the feared man, the rival in the family and the ending.
- Step 1: the settings above; read the script.
- Step 2: generate the cast and the rooms; give the ring and the ledger their images.
- Storyboard: episode 1 first (free on your first project). Rewrite any shot where a weapon is used on a person into a threat and its aftermath.
- Video: generate on the default tier at 10 credits a second, regenerate what misses, composite and post.

| Episode | What happens | The hook |
| 1 | Her father’s debt is called in; the price is a wedding. | The groom tells her he will never touch her — and means it. |
| 2 | The family dinner; the uncle tests her. | She notices the uncle’s ring is the wrong one. |
| 3 | She finds the ledger in the study. | Her father’s name is in it twice. |
| 4 | The uncle arranges an accident. | Her husband is the one who walks away from it. |
| 5 | The family chooses sides at the christening. | “Then let’s ask her who she’s loyal to.” |

## FAQ

### Can a mafia drama show guns?

A gun can appear as an object: holstered, on a table, in a drawer. Shooting someone and showing the wound may be blocked as extreme violence; cut at the threat and show the aftermath.

### Can I use a real crime family or real gangster?

No. Real people and real organisations are not allowed on SceneMixer. Invent the family, its name, its city and its crest.

### Which style gives the dark, cinematic mafia look?

Shot on Film for deep shadows, with Neon Nights or New York Metropolis as the scene style and Power Tailoring for the wardrobe.

### Can I set it in the 1920s?

Yes. Pick Jazz Age or Vintage Hollywood as the costume style and a matching set, or write the era into the script.

### How do I make the boss look intimidating?

Through other people: they go quiet, they stand, they look away. Write their reactions and keep his own delivery low and slow.

### Can the dialogue be in Italian?

Yes. Choose Italian as the dialogue language and Mediterranean as the region in SceneMixer's step 1.

### Can there be a romance scene?

Keep it to what a broadcast drama would show. Explicit content is not generated; a look, a held hand or a kiss that the camera leaves early carries the genre better.

### How many episodes should a mafia romance have?

Enough for the bargain to play out, often twenty to sixty short episodes. Make and post the first few, then continue in the same project.

Updated November 3, 2026


---

# AI Medical Drama: Hospitals, Doctors and Life-or-Death Calls

URL: https://scenemixer.com/guides/make-an-ai-medical-drama/
Make a hospital or doctor drama with AI: the Hospital Ward scene style, Professional Uniforms, emergencies told through faces and monitors, and a patient per episode.

# How to make an AI medical drama

Quick answer: A medical drama is a patient, a decision and a doctor with something to lose. In SceneMixer, the Hospital Ward scene style gives the general hospital look (white walls with a pale green dado, shiny floors, equipment in neat rows) and Professional Uniforms dresses each department in its own cut and colour. Keep surgery and injuries off the screen: an emergency reads on faces, hands and a monitor, which is both better drama and what content review allows.

Brief for AI Create: A medical drama: a surgeon who lost her licence after a death on her table comes back as a night-shift nurse in the same hospital; the patient in bed 7 is the son of the man who sued her; 12 short episodes, each ends with a call to the emergency room.

## Set the hospital

| Setting | Scene Style | Costume Style |
| A big general hospital | Hospital Ward | Professional Uniforms |
| The board and the money | Corporate Cold | Power Tailoring |
| A country doctor | American Small Town or Pastoral Healing | Smart Casual |
| A field hospital | Military Base or WWII Front | Modern Military, WWII Era |
| A rich family's private clinic | Urban Opulence | Minimal Luxury |

## Emergencies without gore

Surgery, blood and open wounds are blocked by content review. Medical drama never needed them: the tension is in the hands, the eyes above the mask and the line on the monitor.

| Beat | Shot to adapt |
| The doors | [Shot 1] 3s: Wide shot, emergency corridor, double doors swinging open, a gurney rushed in by two nurses. |
| The monitor | [Shot 2] 3s: Close-up on a heart monitor, the line flattening. |
| The decision | [Shot 3] 3s: Close-up on @Nora above a surgical mask, eyes moving from the monitor to the door. |
| The line | [Shot 4] 4s: Medium shot. @Dr Hale (sharp): You're a nurse now. Step back. (voice: Hale) |
| The choice | [Shot 5] 3s: Close-up on @Nora's gloved hands, not moving back. |
How to write one shot · Content rules.

## Uniforms, patients and states

- Who wears what. Professional Uniforms keeps one cut and colour per department: say in the script who is a surgeon, a nurse, a resident.
- The patient gown is a state. A character who is admitted in episode 4 needs a second look. States.
- The ward is a location. The same bed by the same window in every episode. The same room in every episode.
- Medical words in dialogue are fine; write them exactly as they should be said. Dialogue in your script.

## Story shapes that work

- The fallen surgeon: a second chance in the wrong uniform.
- The genius resident: brilliant, rude and right.
- Doctor romance: two rivals on the same shift.
- The hospital cover-up: one death that the board wants forgotten.
- The patient of the week: one case per episode, one secret per season.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Hospital Ward; Costume Style Professional Uniforms; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: doctors, nurses and patients with one mark each; a patient-gown state for anyone admitted; the ward, the emergency room and the board room as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The night shift. | Bed 7's chart has a name she knows. |
| 2 | The emergency room. | She saves a patient without the right to. |
| 3 | The board room. | Her old file is on the director's desk. |
| 4 | Bed 7. | The patient wakes up and recognises her. |
| 5 | The roof. | The surgeon who replaced her makes an offer. |

## FAQ

### How do I make a medical drama with AI?

Choose the Hospital Ward scene style and Professional Uniforms in SceneMixer's Step 1, write or generate the script, and tell each emergency through faces, hands and monitors.

### Can I show surgery?

Not the open body. Show the team, the masks, the monitor and the hands; graphic injury is blocked by content review.

### Will doctors and nurses wear different uniforms?

Yes. Professional Uniforms gives each role one cut and colour; say who is who in the script.

### How do I show a character becoming a patient?

Add the hospital-gown look as a state of that character in SceneMixer's Step 2.

### Can I use real hospital names?

Invent them. Real brand marks and names can trigger the IP check.

### Which visual style suits a medical drama?

Digital HD for a modern hospital, Soft Glow for a romance, Shot on Film for a period drama.

### Vertical or landscape?

9:16 Portrait for short-drama feeds, 16:9 Landscape for a series.

### Can the dialogue be in another language?

Yes. The dialogue language is set in Step 1; 15 languages are available on SceneMixer.

Updated November 3, 2026


---

# AI Medieval Knight Drama: Castles, Tournaments and Oaths

URL: https://scenemixer.com/guides/make-an-ai-medieval-knight-drama/
Make a medieval knight drama with AI: Medieval Castle sets, Plate Armor and Medieval Nobility costumes, horses as characters, tournaments one action per shot.

# How to make an AI medieval knight drama

Quick answer: A knight's story is an oath tested. In SceneMixer, build the world with Medieval Castle and Medieval Nobility, put the knights in Plate Armor with one heraldic colour each so they can be told apart with helmets on, make important horses characters with their own sheets, and write tournaments and battles one action per shot.

Brief for AI Create: A medieval drama: a squire takes her dead brother's armour to fight in the king's tournament; only the old blacksmith knows; eight episodes, three jousts, one unmasking.

## Build the realm

| Place | Scene Style | Costume Style |
| The castle hall | Medieval Castle | Medieval Nobility |
| The tournament field | Medieval Castle exteriors | Plate Armor |
| The dark abbey | Gothic Dark | Medieval Nobility; plain monks’ habits written into the description |
| The northern march | Viking Fjord, Frozen Wilderness | Nordic Viking |
| The village | Eastern European Old Town | Eastern European Folk |

## Knights you can tell apart

With helmets down, every knight looks the same. Give each a heraldic colour and one emblem on the shield and surcoat, written into the description: blue surcoat with a white tower. The unhelmeted face is a separate state if the armour is ever removed on screen. States.

Horses that matter are characters: colour, markings, tack. Background horses are scenery. Animals as characters.

## A joust, shot by shot

| Beat | Shot to adapt |
| The lists | [Shot 1] 4s: Wide shot. The tournament field, banners, crowd in the stands; two knights at either end of the barrier. |
| The charge | [Shot 2] 3s: Low angle. @Squire in a blue surcoat spurs her grey horse forward. |
| The impact | [Shot 3] 2s: Medium shot. Lances meet; @Sir Aldric's lance shatters into splinters. |
| The reaction | [Shot 4] 3s: Close-up on @King Osric leaning forward in the royal box. |
Fights and crowds.

## Story shapes

- The disguised knight: the secret is who is under the helmet.
- The oath: a promise to a dying lord that costs everything.
- The forbidden match: a knight and a lady of the enemy house.
- The siege: one castle, a winter, a traitor inside.
History dramas · Royal drama.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Medieval Castle, Gothic Dark; Costume Style Plate Armor, Medieval Nobility; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: one heraldic colour per knight; an unhelmeted state where it matters; the hall and the tournament field as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Contest | The ending beat |
| 1 | The brother's funeral | She takes the armour |
| 2 | The first joust | She wins and cannot lift her visor |
| 3 | The feast | A lord recognises the horse |
| 4 | The second joust | Her helmet strap is cut |
| 5 | The final | The unmasking before the king |

## FAQ

### How do I make a medieval knight drama with AI?

Use Medieval Castle and Medieval Nobility in SceneMixer, Plate Armor with one heraldic colour per knight, and write tournaments one action per shot.

### How do I tell armoured knights apart?

Give each a colour and an emblem on shield and surcoat, written into the description.

### Can horses be consistent?

Yes. Make important horses characters with their own sheets.

### Can I show battles?

In wide shots with background soldiers and close-ups on named characters; graphic violence is blocked.

### Which look suits it?

Shot on Film or Digital HD for live action; Classical Oil Painting for a painted chronicle.

### Can I use King Arthur?

Public-domain legends are fine with your own designs; avoid the look of any specific film or show.

### Vertical or landscape?

Landscape suits castles and tournaments.

### How long should a joust scene be?

Four to six short shots: the field, the charge, the impact, the reaction.

Updated November 3, 2026


---

# How to Make an AI Memorial Video: A Gentle, Practical Guide

URL: https://scenemixer.com/guides/make-an-ai-memorial-video/
Make an AI memorial or tribute video: a short stylised film of memories with a narrator, not a photo slideshow. What a photo can do (members) and what it can't.

# How to make an AI memorial video

Quick answer: A memorial video made here is a short tribute film told in scenes, in a gentle drawn or soft look, with a narrator reading what you wrote. It is not a slideshow of photos, and it does not bring back anyone's voice. It takes text, so you write the memories; a member can upload one photo as a character reference, which is redrawn in the project's style with the face, hair and clothing read from it. You can also leave faces out altogether: a garden, a coat on a hook, a radio, hands. Begin with the lightest version and add only what you want.

Brief for AI Create: A two-minute tribute to a grandfather who kept a garden: the garden at six in the morning, his coat on the hook by the door, the radio he kept tuned to one station, a narrator reading five short lines; slow, warm, no faces, ends on the garden gate.

## What it can and cannot be

- Can. A tribute film in short scenes, a narrator's voice-over, ambience, a quiet score, a gentle look (Watercolor Storybook, Soft Glow, Painterly Anime).
- Cannot. Turn a folder of photos into a slideshow: the site takes text, up to four photos of one subject for an ad, or a video. What you can feed in.
- Does not offer. Cloning anyone's voice. The site does not offer cloning of a named person; a member can upload a clip they have the right to use. How voices work.

## Three versions, from lightest

| Version | What is on screen | What you need |
| 1. No faces | Places, objects, hands, weather; a narrator speaks | Only your words |
| 2. A drawn character | A figure inspired by the person, drawn in a gentle look from your written description | Your words; no photo |
| 3. A photo as reference | The character redrawn with the face, hair and clothing read from your photo | A membership; a photo you have the right to use |
Version 3 is Upload & Replace on the character's card in step 2, then Optimize into a reference sheet: it reads the face, hair and clothing from the photo and redraws them in the project's style. The clearer and larger the face, the more of it the video can hold on to. Using your own photo.

## Rights and the face check

The Terms say: “Do not upload or generate a real person's face, voice or likeness without their documented consent.” The window before your first upload asks you to confirm that you have the right to use the material and that every real person whose face or voice appears in it has given consent. Whether you hold that right for a person who has died, and whether to agree the film with the people close to them, is your decision; this page does not judge it.

Separately, the video provider's face check can refuse an image that looks like a real person. The segment fails, the credits come back automatically, and retrying with the same image fails the same way; after 12 blocked attempts in a day, video generation pauses for 1 hour. A drawn look sits far from a real face and is rarely flagged. What to do when a reference is flagged.

## Write three memories and one line

| Beat | Seconds | Picture | Sound |
| The place | 20 | The garden at six in the morning | Ambient: birds, a gate |
| The object | 20 | His coat on the hook by the door | Quiet; one line of voice-over |
| The habit | 20 | The radio tuned to one station | Ambient: a far-off tune, low |
| The line | 15 | The garden gate, left open | One sentence, then silence |
Narration is spoken off-screen, so people in frame keep their mouths closed. In the storyboard it is one line per shot, at about three words a second:

[Shot 1] 4s: Wide shot, slow push in. A garden at first light, a spade standing in the soil. @Narrator (voice-over, quiet): He was up before the light. (voice: Narrator)
[Shot 2] 4s: Close-up on a coat on a hook, the sleeve a little worn. Ambient: a clock, far off. No dialogue.

Lines, voice-over and captions.

## Pace, look and music

- Slow. Long shots, one thing in each. No cliffhanger; end on an image.
- Look. Soft Glow for a hazy, low-contrast feel; Watercolor Storybook for something tender and drawn.
- Music. The episode is scored automatically and you cannot pick the piece. To use a song they loved, turn the Music switch off, download, and lay a track you have the right to use under it in an editor. The built-in score and your own.

## Try one shot first

The storyboard is free to read and rewrite before anything is rendered on your first drama, and each account has one free preview of up to 5 seconds. Make the first segment alone, watch it with the sound on, and carry on only if it feels right. The free preview · What you get for free.

## Sharing it

A video shared through a SceneMixer link always carries the small AI badge, so anyone who watches can see it was made with AI. Your own download follows your badge setting: a free account's carries it, a member's is off by default. Shared links open without an account. Why shared videos carry the badge · Composite and share an episode.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing; Costume Style Soft Everyday; Visual Style Soft Glow, Watercolor Storybook; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw only what you want to see: a garden and a doorway as places, a coat as a prop, and a narrator; leave the person out until you choose to add them. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| You have | You can make | Needs |
| Words and memories | A tribute film in scenes with a narrator | Nothing else |
| A description of the person | A drawn character inspired by them | Nothing else |
| A photo you have the right to use | A character redrawn from it | A membership |
| A folder of photos | Not a slideshow | — |
| A recording of their voice | Not as a clone | A member may upload a clip they have the right to use |

## FAQ

### Can I make a memorial video from photos?

Not as a slideshow. SceneMixer takes text, up to four photos of one subject for an ad, or a video. A member can upload one photo as a character reference.

### Will the face look like them?

It will be a likeness in the project's style, not a photograph: the face, hair and clothing are read from the photo and redrawn. A small or unclear face gives the video less to hold on to.

### Can it use their voice?

The site does not offer cloning of a named person. A member can upload a clip they have the right to use; a recording of a real person is that person's.

### What if the face check blocks it?

The credits come back automatically. Change the image or choose a drawn look; after 12 blocked attempts in a day, SceneMixer pauses video for 1 hour.

### Which look is gentlest?

Soft Glow and Watercolor Storybook are the softest of the looks; Painterly Anime suits larger, quieter scenery.

### How long should it be?

One to three minutes. Three memories and one closing line fit about 75 seconds.

### Does it carry an AI label?

A shared link always shows the small AI badge; your own download follows your setting.

### What does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join, so 75 seconds is 75 seconds of each, before retakes.

Updated November 3, 2026


---

# AI Mermaid Fantasy Drama: Undersea Kingdoms, Tails and Two Worlds

URL: https://scenemixer.com/guides/make-an-ai-mermaid-fantasy-drama/
A mermaid story with AI: the Undersea Kingdom scene style, Merfolk costume, one character with both a tail and legs, and shore scenes that join the two worlds.

# How to make an AI mermaid fantasy drama

Quick answer: A mermaid drama lives in two worlds and has to look right in both. In Step 1 pick the scene style Undersea Kingdom — spiral-shell palaces and mother-of-pearl arches on coral reefs, kelp forests, giant anemones, water shading from pale teal to deep blue — and the costume style Merfolk: a top of pearls and shells or scale armour, a scaled tail with translucent fins, a crown of pearl and conch. Use Follow script for the region and a fantasy-friendly visual style. A mermaid who comes ashore is the same character with a second look, a tail and legs as two states of one sheet, and the shore scenes that join the worlds can be a Mediterranean cove or a harbour.

Brief for AI Create: A mermaid fantasy: a young merfolk princess is sent ashore for one summer with legs and no memory of why; a lighthouse keeper’s daughter finds her on the rocks, and both worlds start looking for her; 8 short episodes, English dialogue, every episode ends at the tide line.

## The choices that make it a sea kingdom

| Step 1 setting | Choose | Why |
| Scene Style | Undersea Kingdom | Coral and pearl: spiral-shell palaces, mother-of-pearl arches, kelp forests and giant anemones. |
| Costume Style | Merfolk | Pearl-and-shell tops or scale armour; a scaled tail with translucent fins; a pearl and conch crown. |
| Visual Style | Family 3D Animation, Watercolor Anime, Watercolor Storybook or Digital HD | 3D or watercolour suits the colour and light of water; live action works with a soft, dreamy grade (Soft Glow). |
| Target Region | Follow script | A fantasy kingdom has no real-world region; this keeps the names you chose. |
| Dialogue language | Any supported language | Spoken as each clip renders. |
All of this is written into the descriptions when the script is read, so choose it first. Why they are fixed afterwards · Every scene and costume style.

## Two worlds, two looks, one character

The signature trouble of a mermaid story is the transformation. Do not make it two characters. Make it one character with two states:

- In Step 2, build the princess once, in her tail. Then add a second look for the shore: the same face with legs and a simple dress.
- Pick the right look for the segments where she is in the sea and for the segments where she is ashore.
- Write the change as an event in the shot where it happens (the tail falls away into legs as the wave leaves) so the next segment starts from the new look.
Costume changes and scene states · What the system keeps track of between segments.

## The shore, which is not the sea

Pick shore scenes that give the two worlds a clear join, and give each its own scene card in Step 2:

- Mediterranean Coast — white walls, blue doors, stone steps down to a cove.
- Island Resort — white sand, shallow turquoise sea, boardwalks.
- Age of Sail Harbor — a wooden harbour and tall ships, for a period mermaid tale.
- Frozen Wilderness — for a cold northern sea.
Each segment stays in one place and one stretch of time; moving from the reef to the rocks is a new segment. A segment stays in one place and one timeline.

## Keep it dressed, keep it kind

The content checks stop nudity and anything sexual. Give every character a full clothing description in Step 2: a bodice of overlapping pearl shells over a fitted scale top, a scaled armour vest with a coral clasp, a flowing tail of blue-green scales with translucent fins. A mermaid story is an all-ages story by tradition; keep it that way. If a clip is blocked.

Invent your own sea folk, sea witches and kings. Characters from a known film or adaptation can trigger the IP check.

## An undersea scene, shot by shot

| Beat | Shot to adapt |
| The palace | [Shot 1] 4s: Wide shot, fixed camera. A spiral-shell palace on a coral reef, mother-of-pearl arches catching light from above, kelp swaying. @Nerea, in a pearl-shell bodice and a blue-green tail, floats at the foot of the steps. |
| The guard | [Shot 2] 3s: Medium shot. @Dorian, a merman guard in scale armour, blocks the arch with his spear held crosswise. |
| The line | [Shot 3] 4s: Close-up on @Nerea, lifting her chin. @Nerea (firm, quiet): Tell my mother I went by my own choice. (voice: Nerea) |
| The pause | [Shot 4] 3s: Close-up on @Dorian. He lowers the spear a hand’s breadth. |
| The tide | [Shot 5] 3s: Wide shot from the arch, @Nerea swimming up and away toward the bright surface until she is a small shape in the light. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- The princess ashore: a stranger on land, learning stairs, doors and money.
- The bargain: a sea witch, a price, a clause nobody read.
- Two kingdoms: reef and shore, each sure the other started it.
- The lighthouse keeper’s secret: a human who has known about the sea people all along.
- The sunken ship: a wreck, a stolen map and a treasure that belongs to neither side.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief in AI Create.
- Step 1: Type Short Drama, TV Series or Film; scene style Undersea Kingdom; costume style Merfolk; a visual style; Target Region Follow script; dialogue language; 9:16 Portrait or 16:9. Read the script.
- Step 2: build each merfolk character with their tail; add a shore look for anyone who comes ashore. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). Pick the tail look or the shore look for each segment.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | The palace of the reef. | The princess slips out through the arch. |
| 2 | The surface, for the first time. | A lighthouse beam passes over her. |
| 3 | Legs, and a beach at dawn. | The keeper’s daughter finds her on the rocks. |
| 4 | The village market. | A coin that glows when she touches it. |
| 5 | The reef searches for her. | The guard finds her footprints, and then hers are not alone. |

## FAQ

### How do I make a mermaid story with AI?

In SceneMixer's Step 1, choose the scene style Undersea Kingdom and the costume style Merfolk, then write or generate the script and render episode by episode.

### What does the Undersea Kingdom scene style look like?

Spiral-shell palaces and mother-of-pearl arches on coral reefs, kelp forests, giant anemones, and water shading from pale teal to deep blue.

### How do I show a mermaid with legs on land?

Make it one character with two looks, a tail and legs, as states of the same sheet. Pick the right look for each segment.

### Which visual style suits a mermaid story?

Family 3D Animation, Watercolor Anime or Watercolor Storybook for a bright, painted look; Digital HD or Soft Glow for live action.

### Can I show a mermaid without clothes?

No. The content checks on SceneMixer stop nudity. Give every character a full clothing description, such as a pearl-shell bodice or scale armour.

### Can I use a character from a famous mermaid film?

No. Known characters are refused by the IP check; invent your own sea folk.

### Which Target Region should I pick?

Follow script for a fantasy kingdom; it keeps the names you chose.

### Will the princess look the same in every episode?

Yes. Each character is drawn once on SceneMixer and that sheet is used in every shot, in her tail look and in her shore look.

Updated November 3, 2026


---

# AI Military Drama: Bases, Missions and Brothers in Arms

URL: https://scenemixer.com/guides/make-an-ai-military-drama/
Make a military or army drama with AI: Military Base and Modern Military styles with invented insignia, missions in short shots, combat kept off-screen.

# How to make an AI military drama

Quick answer: A military drama is a unit, a mission and an order someone does not want to follow. In SceneMixer, the Military Base scene style gives the camp and the drill ground (right angles, white lines, olive and khaki) and Modern Military dresses an invented country's army in field, service and dress uniforms; the insignia are always invented, which keeps you clear of real armies. Tell combat through orders, faces and the moment after, not wounds.

Brief for AI Create: A military drama: a disgraced sniper is sent back to the unit that blamed her for an ambush; their new mission returns to the same valley; 12 short episodes, each ends with a radio call nobody should be able to make.

## Set the unit

| Setting | Scene Style | Costume Style |
| A modern base | Military Base | Modern Military |
| A desert deployment | Frontier Vastness | Modern Military |
| A winter operation | Frozen Wilderness | Modern Military, Outdoor Gear |
| A naval or space fleet | Starship | Starship Uniform |
| The Second World War | WWII Front | WWII Era |
World War II dramas · Space operas.

## Uniforms and ranks

Modern Military gives an invented country's army in an international western style: field dress, service dress and dress uniform with shoulder boards, medals and sashes. Rank follows each character's role in the script. The dress uniform for a ceremony or a funeral is a state of that character. States.

Weapons and vehicles that matter (her rifle, the convoy truck) are props or locations with a picture. Props.

## The mission, shot by shot

| Beat | Shot to adapt |
| The ridge | [Shot 1] 4s: Wide shot, a rocky ridge at dawn, mist in the valley below. @Sergeant Hale lies prone behind a scope. |
| The radio | [Shot 2] 3s: Close-up on the radio. @Captain Ruiz (off-screen, static): Hold fire. Hold fire. |
| The valley | [Shot 3] 3s: Wide shot through the scope, a convoy entering the valley. |
| The choice | [Shot 4] 4s: Close-up on @Sergeant Hale. @Sergeant Hale (whisper): That's the same convoy. (voice: Hale) |
Combat: one action per shot, the moment before and after, no injuries on screen. Fights and groups · Content rules.

## Story shapes that work

- Boot camp: recruits, a hard instructor and one who should not be there.
- The return: back to the unit that blamed you.
- The rescue mission.
- Military romance: two officers on the same base.
- Homecoming: a soldier and the family who waited.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Military Base, Frontier Vastness; Costume Style Modern Military; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the unit with ranks in the script and one mark each; a dress-uniform state where needed; the base, the ridge and the briefing room as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The base gate. | Her old unit lines up to watch her walk in. |
| 2 | The briefing room. | The mission map shows the same valley. |
| 3 | The ridge. | The convoy she was blamed for is back. |
| 4 | The camp at night. | A radio call in her dead spotter's voice. |
| 5 | The valley. | The captain gives the order she refused last time. |

## FAQ

### How do I make a military drama with AI?

Choose Military Base and Modern Military in SceneMixer's Step 1, write ranks into the script, make key weapons and vehicles props, and tell combat through short shots.

### Are the uniforms of a real army?

No. Modern Military dresses an invented country's army and the insignia are always invented.

### Can I show combat?

Yes, through orders, faces, movement and aftermath. Graphic injuries and extreme violence are blocked on SceneMixer.

### How do I show a dress uniform at a ceremony?

Add it as a state of the character in Step 2.

### Can I set it in a real war?

Yes for history, such as the WWII Front style, but invent the people; real people cannot be depicted.

### Which visual style suits a military drama?

Digital HD or Shot on Film; Black & White Film for a historical piece.

### Vertical or landscape?

16:9 Landscape suits the scale; 9:16 Portrait suits short-drama feeds.

### Can I film a whole platoon?

Show two or three soldiers per shot and put the rest in the background line.

Updated November 3, 2026


---

# AI Mockumentary: Talking Heads, Observed Moments and Dry Jokes

URL: https://scenemixer.com/guides/make-an-ai-mockumentary/
Make a mockumentary with AI: talking-head interviews to camera, observed scenes, the same office or town drawn once, deadpan timing in shot seconds, and a narrator.

# How to make an AI mockumentary

Quick answer: A mockumentary is a comedy shot as if it were a documentary: people talk to the camera about themselves, and then we watch them do the opposite. In SceneMixer, write two kinds of shots, interviews (a character facing the camera in a fixed spot) and observed scenes, set the town or office once, and time the joke with a beat of silence after each line.

Brief for AI Create: A mockumentary about the volunteer committee of a tiny lighthouse museum preparing for its first ever visitor; six episodes, talking heads, a very serious chair.

## Two kinds of shot

| Shot type | How to write it |
| Talking head | Medium close-up, the character facing the camera, seated in the same spot every time, speaking to an unseen interviewer |
| Observed scene | Wider framing, slightly off-centre, characters not looking at the camera |
| The glance | A character looks at the camera after something absurd; no line |
Keep each person's interview spot the same: a chair, a window behind them. It becomes their location. How to write one shot.

## The joke is the gap

| Beat | Shot to adapt |
| The claim | [Shot 1] 4s: Talking head. @Gerald in a tweed jacket, seated by the window. @Gerald: I would describe myself as calm under pressure. |
| The cut | [Shot 2] 3s: Observed. @Gerald on a ladder, frantically re-hanging a sign that says WELCOME upside down. |
| The glance | [Shot 3] 2s: @Margaret holding the ladder, looks at the camera. No dialogue. |
Give the glance its own seconds; the pause is the punchline. Shot seconds and timing.

## Setting the stage

Mockumentaries live in small, ordinary places: an office in Corporate Cold, a town in American Small Town, a village hall. Pick Digital HD for the documentary look. The sample De Vuurtoren van de Fjord has a lighthouse and a wooden barn that would suit it. Sitcoms and skits.

## A narrator, or not

A deadpan documentary narrator can open and close each episode in voice-over. Or leave it to the interviews. Either way, keep the real-people rule: invented subjects, not parodies of real public figures. Content rules.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style American Small Town, Corporate Cold; Costume Style Smart Casual, Soft Everyday; Visual Style Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: a small cast with specific faces; each person's interview spot as a location; the museum or office as the main set. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Event | The gap |
| 1 | Meet the committee | Everyone claims to be in charge |
| 2 | The sign | Calm under pressure |
| 3 | The leaflet | Three typos, one is the name |
| 4 | The rehearsal | Nobody plays the visitor well |
| 5 | The visitor | He is lost and wants the toilet |

## FAQ

### How do I make a mockumentary with AI?

Write talking-head interviews to camera and observed scenes, keep each interview spot fixed, and time the joke with silence.

### Can characters look at the camera?

Yes; write it into the shot.

### Which look suits a mockumentary?

Digital HD for a documentary feel.

### Can I parody a real person?

No. Real people's likenesses are refused by SceneMixer; invent your subjects.

### How long should episodes be?

Two to five minutes.

### Can there be a narrator?

Yes, as a voice-over character.

### How do I keep the interview set the same?

Make each interview spot a location; it is drawn once and reused.

### Is there music?

A score is added automatically on SceneMixer; switch it off for a drier tone.

Updated November 3, 2026


---

# AI Monster Movie: Creatures, Scale and the Late Reveal

URL: https://scenemixer.com/guides/make-an-ai-monster-movie/
Make a monster movie with AI: the creature as a character with its own sheet, scale written against things we know, sound before sight, and a reveal saved for late.

# How to make an AI monster movie

Quick answer: A monster movie is about the thing you have not seen yet. In SceneMixer, make the creature a character with its own reference sheet, described plainly (shape, skin, size, how it moves), so it looks the same every time it appears. Write its scale against something familiar, a bus, a lighthouse, a pine tree. Hear it before you see it, and save the full reveal for late in the film.

Brief for AI Create: A monster film: a fishing town in the far north finds its nets cut every night; the harbour master's daughter films something in the water that is bigger than the ferry; four nights, one storm.

## Design the creature once

Describe it like an animal no one has seen: body shape, size, skin or scales, eyes, how it moves. Do not name a famous monster; the IP check refuses it, and a plain description gives a better sheet anyway. If it changes (wounded, enraged, glowing), add a state. Non-human characters · States.

## Scale is a comparison

A model has no idea how big enormous is. Write it against something in frame: its head rises higher than the lighthouse, a footprint wide as a car. Keep the people small in the frame when the creature is near. Say what is there.

| Setting | Scene Style | What it gives the monster |
| Northern harbour | Frozen Wilderness, Viking Fjord | Dark water, ice, fog |
| Jungle expedition | Tropical Rainforest | Cover, sound, sudden clearings |
| Small town | American Small Town | Familiar things to crush |
| Megacity | New York Metropolis, Cyber Metropolis | Towers for scale |

## The reveal, in stages

| Stage | Shot to adapt |
| The sign | [Shot 1] 3s: Close-up. A fishing net hauled onto a deck, cut clean through. Ambient: gulls, wind. |
| The sound | [Shot 2] 3s: Wide shot, night. The harbour under one lamp; a low groan rolls across the water. |
| The shape | [Shot 3] 3s: A dark ridge breaks the surface beyond the breakwater and sinks again. |
| The face | [Shot 4] 3s: Close-up on @Ingrid holding a phone camera, frozen. |
| The reveal | [Shot 5] 4s: Extreme wide. @Leviathan rises behind the lighthouse, its head higher than the lamp. |

## Story shapes

- The town that will not believe: one witness, a mayor who says it is a seal.
- The expedition: a crew goes looking and finds it.
- The creature with a reason: it is protecting something, and the humans took it.
- The small monster: one kitchen, one night, something under the floor.
Horror shorts · Sci-fi.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Frozen Wilderness, Viking Fjord; Costume Style Outdoor Gear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the creature as a character with its own sheet; an enraged or wounded state; the harbour and the lighthouse as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | What we see of the monster | The ending beat |
| 1 | Nothing: cut nets | A sound from the water |
| 2 | A shape in fog | A boat does not come back |
| 3 | One eye, one fin | The daughter's video goes online |
| 4 | All of it | It turns toward the town |
| 5 | Its reason | Someone has to give something back |

## FAQ

### How do I make a monster movie with AI?

Make the creature a character with its own sheet, write its scale against familiar objects, and reveal it in stages across the film.

### How do I keep the monster looking the same?

One reference sheet for the creature, used in every shot; changes such as wounds are states.

### How do I make it look huge?

Compare it to something in frame: a lighthouse, a bus, a tower. Keep people small when it is near.

### Can I use a famous monster?

No. Famous creatures trigger SceneMixer's IP check; describe your own.

### Can the monster attack people?

Show the threat and the aftermath, not wounds; graphic violence is blocked on SceneMixer.

### Which style suits a monster movie?

Digital HD or Shot on Film for live action; Black & White Film for a classic creature feature.

### Can the monster make sounds?

Yes. Put roars and groans in the shot's ambient line.

### Vertical or landscape?

Landscape suits scale; vertical works for a short creature series.

Updated November 3, 2026


---

# AI Movie Trailer Maker: A Trailer for Your Film Idea, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-movie-trailer/
Make a concept trailer for a film you have written: the three-act trailer shape, a narrator, a consistent cast, a score, and how to get the title on screen.

# How to Make an AI Movie Trailer for Your Own Film Idea

Quick answer: A movie trailer is the fastest way to show what a film would feel like before it exists — for a pitch, a crowdfunding page or just to see it. In SceneMixer you write a 60-to-120-second trailer script: a voice that sets up the world, three or four key moments that rise, a final beat and the title. Set the Type to Film and the aspect to 16:9 in step 1; the cast is drawn once and kept consistent across every shot, the storyboard is generated, and the trailer is rendered and joined into one video with a score. Generated video does not draw legible text, so the title is spoken by the narrator or added in an editor afterwards.

Brief for AI Create: A 90-second trailer for a sci-fi film: the last cargo pilot on a dying space station receives a distress call from a ship that was lost 30 years ago; a narrator, rising tension, the title spoken at the end.

## The three-act trailer

| Part | Seconds | What it does |
| The world | 15–25 | Calm. Who, where, what is normal. A narrator or a single line. |
| The disruption | 20–30 | The thing that changes everything. One clear image. |
| The rise | 25–45 | Faster cuts, bigger moments, the stakes said out loud. |
| The button | 5–10 | Silence, one last line or image, then the title. |
Write it as a script with those beats, not as the whole film: a trailer shows moments, it does not explain the plot. Short drama, series or film.

## A trailer script you can adapt

| Beat | Shot to adapt |
| The world | [Shot 1] 5s: Extreme wide shot, slow push in. A rusted space station turning against a red planet. @Narrator (voice-over, low): For thirty years, nothing has come to Station Kell. Ambient: deep hum. |
| The pilot | [Shot 2] 4s: Medium shot. @Ana alone in a cockpit, eating from a tin, a photo taped to the console. No dialogue. Ambient: hum, a beep. |
| The disruption | [Shot 3] 3s: Close-up. A red light blinking on the console. @Radio Voice (off-screen, crackling): Mayday. This is the Halcyon. Ambient: static. |
| The stakes | [Shot 4] 3s: Close-up on @Ana. @Ana (whisper): The Halcyon was lost before I was born. Ambient: static. |
| The rise | [Shot 5] 2s: Wide. A ship emerging from dust. [Shot 6] 2s: @Ana running a corridor. [Shot 7] 2s: An airlock opening on darkness. Ambient: alarms rising. |
| The button and title | [Shot 8] 4s: Black, then @Ana’s face lit red. @Narrator (voice-over): Halcyon. Ambient: silence. |
Short shots in the rise are what make a trailer feel like one; one action per shot. Shot seconds and timing · How to write one shot.

## The cast, the look and the voice

- Cast: two or three characters at most; each is drawn once and every shot is matched to that sheet. Same face in every shot.
- Look: Shot on Film or Digital HD for live action, Game Engine Realism for sci-fi and action, Painterly Anime or Family 3D Animation for an animated film. Visual style presets.
- The trailer voice: an Off-screen narrator with a voice description such as deep, slow, gravelly male voice. Voice-only characters.
- Score: projects can carry their own generated score. Add background music.

Visual Style: Digital HD

Visual Style: Game Engine Realism

Visual Style: Painterly Anime

Visual Style: Family 3D Animation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
Character sheets from the sample Starfall Command: one reference image per character, used in every shot they appear in.

## Getting the title on screen

Generated video does not draw legible text, so do not ask for the title inside a shot. Two ways that work: the narrator says the title over the last image, or you download the trailer and add the title card in any video editor. Leave a few seconds of a clean final shot for it. On-screen text · Download the original.

## Original films only

Fan trailers for real films, real actors’ faces and known characters are refused by the IP check. A trailer for your own story, with your own characters, is what this is for. IP check.

## Make it in SceneMixer, start to finish

- Script: the trailer script on Upload Script, or a brief in AI Create with the premise, the length and the ending beat.
- Step 1: Type Film; 16:9; the visual style; the dialogue language. Read the script.
- Step 2: check the leads’ sheets; add the narrator. The first 15 setup images on an account are free.
- Storyboard: keep the rise in short shots; one action each.
- Video: draft on the default tier at 10 credits a second, then regenerate the keepers at a higher tier; composite and download. Draft at 480P, finish at 1080P.
For a sci-fi sample, watch Starfall Command. For a book rather than a film: make an AI book trailer.

## FAQ

### How do I make an AI movie trailer?

Write a 60-to-120-second trailer script in four parts, set the Type to Film and the aspect to 16:9 in SceneMixer's step 1, add an off-screen narrator, and render; the cast stays consistent across every shot.

### Can it put the title on screen?

Not as text inside a shot: generated video does not draw legible text. Have the narrator say it, or add a title card in an editor after downloading.

### Can I make a trailer for a real movie?

No. Real films, actors and known characters are refused by SceneMixer's IP check. It is for trailers of your own stories.

### Does the trailer have music?

Projects can carry their own generated score; you can also add your own licensed music in an editor afterwards.

### How long should a trailer be?

Sixty to ninety seconds for a pitch or social post; up to two minutes for a full trailer.

### Can I get a deep trailer voice?

Yes. Add an off-screen narrator and describe the voice, for example deep, slow and gravelly.

### Can it be animated?

Yes. Pick an animation visual style in step 1, such as Painterly Anime or Family 3D Animation.

### What does a trailer cost?

Video is priced per second of the tier you pick, so a 90-second trailer on SceneMixer's default tier is about 90 seconds at that rate; the generate button shows the price first.

Updated November 3, 2026


---

# AI Mystery Drama: Clues, Suspects and a Fair Reveal, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-murder-mystery-drama/
Plant the clue, rotate the suspects, reveal the killer: how to write a mystery short drama, keep every clue looking the same, and handle a crime without showing a body.

# How to Make an AI Mystery or Detective Drama

Quick answer: A mystery is a promise of fairness: every clue the detective uses must have been on screen before the reveal. That makes consistency the genre’s whole craft, and it is what SceneMixer is built for: give each clue a prop image so the torn photograph in episode 2 is the same photograph in episode 6, give every suspect a card and a voice, and make flashbacks a state you pick shot by shot. Keep the crime itself off screen — the body is reported, the scene is examined — because gore is blocked and because the mystery is about who, not how.

Brief for AI Create: A murder mystery: a violinist returns to the street she fled seven years ago when her teacher is found dead in his shop; four neighbours each lie about one hour; each episode closes on one piece of evidence that rewrites the last; noir, quiet, precise.

## How a fair mystery is built

- One victim, a closed circle. Four or five suspects who all had a reason. Fewer and the answer is obvious; more and the audience stops tracking faces.
- Every suspect lies about one thing. Most lies are innocent secrets. The audience has to sort them.
- Clues shown, not explained. A clue appears in frame without comment. The detective’s attention to it can be a look, not a speech.
- A false solution. Two thirds of the way through, the obvious suspect is arrested. Then the clue that does not fit.
- The reveal uses only what we saw. The detective walks back through the clues; each one we remember.
For a series, end each episode on a new piece of evidence that changes what the previous one meant.

## Settings to pick in step 1

- Visual Style: Shot on Film for noir shadows, Digital HD for a modern procedural, Black & White Film for a classic whodunit.
- Scene Style: Forensic Cold White for labs and interview rooms, Neon Nights or New York Metropolis for city noir, Gothic Dark for a country house, American Small Town for a town where everyone knows everyone.
- Costume Style: Professional Uniforms for police and lab staff, Smart Casual for the suspects, Jazz Age or Vintage Hollywood for a period mystery.
- Type: Short Drama for an evidence-a-day series, Film for a single whodunit.

Visual Style: Black & White Film

Scene Style: Forensic Cold White

Scene Style: Neon Nights

Scene Style: Gothic Dark

Costume Style: Professional Uniforms
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Clues that look the same every time

A clue that changes design between episodes is not a clue. For each object the solution depends on:

- Check it has a card in step 2 with its own image — the read creates props for objects the story leans on; on a paid plan you can add one it missed. Add a prop to this episode.
- Describe it precisely, including the detail that matters: a pocket watch with a cracked crystal, stopped at 9:40, initials E.R. engraved inside the lid.
- Name it with @ in every shot it appears in, so the same reference image is used.
- If it changes — the photograph torn in half, the vase broken — make the changed version a state of the prop.
Small printed text will not be legible in a generated frame. The number on a receipt or the words in a diary should be spoken by whoever reads them. Same prop in every hand · Props and reference images.

## Flashbacks and the night of the crime

Mysteries replay the same hour from different witnesses. If a suspect looked different that night — a different coat, wet hair, a bandaged hand — add that look as a state of their character and pick it only in the flashback segments: in the episode editor, select the segment, open the character from the episode’s assets and choose Use this state in this shot. Picking it for a whole episode would put the flashback look in the present-day scenes too. If the flashback is set in the same place, the location image keeps the room identical, which is exactly what lets the audience compare two accounts. Use a character state in one scene.

Character sheets from the sample The Violin on Mercer St: one reference image per character, used in every shot they appear in.

## The crime off screen

Gore and extreme violence are blocked. The genre does not need them:

| Instead of | Write |
| The body | The shop door ajar, the violin case open on the floor, a police light turning on the wet window. |
| The murder in flashback | A raised hand in silhouette against a lit doorway; cut to the clock. |
| The autopsy | The examiner peels off her gloves. “He was dead before the fire started.” |
Blocked by moderation: what to change.

## An interrogation, written as shots

| Beat | Shot |
| The question | [Shot 3] 5s: Medium shot, fixed camera, the interview room. @Detective Hale across the table from @Mr. Ward; a single overhead lamp. @Detective Hale (flat): Where were you at nine forty? Ambient: the hum of the lamp. |
| The tell | [Shot 4] 3s: Close-up. In frame only: @Mr. Ward. His thumb turns the ring on his finger. No dialogue. |
| The clue | [Shot 5] 4s: Close-up on the @Pocket watch on the table between them, crystal cracked. @Detective Hale (off-screen, quiet): Funny. It stopped at nine forty. |
Shot 4 is the kind of clue a fair mystery plays without comment. How to write one shot · Writing who is off-screen.

## Make it in SceneMixer, start to finish

- Script: yours on Upload Script, or a brief in AI Create with the victim, the circle of suspects, the detective and the kind of ending.
- Step 1: the settings above; read the script. Check in the summary that the clue the solution needs appears before the reveal.
- Step 2: generate every suspect and the key locations; check each clue has an image; add flashback looks as states.
- Storyboard: episode 1 first (free on your first project). Make sure clues are named with @ wherever they appear.
- Video: generate on the default tier at 10 credits a second; composite; post.
For the tone, watch The Violin on Mercer St, a noir mystery where each episode closes on a piece of evidence that rewrites the previous truth.

| Episode | The evidence | What it changes |
| 1 | The teacher found dead; the shop clock stopped at 9:40. | Whoever did it was there before closing. |
| 2 | The neighbour’s alibi: a receipt from the bar across the road. | The receipt is from the wrong night. |
| 3 | A torn photograph in the violin case. | The teacher had a son nobody mentioned. |
| 4 | The son is arrested. | His hands were bandaged that night — he could not have played the recording. |
| 5 | The violinist plays the teacher’s last piece. | One wrong note, written in by the person who finished it. |

## FAQ

### How do I keep clues consistent across episodes?

Give each clue a prop image, describe the identifying detail, and name it with @ in every shot. If the clue changes, make the changed version a state of the prop.

### Can the murder be shown?

Not in detail. Gore and extreme violence are blocked. Report the body, show the scene and its aftermath, and keep a flashback of the act to a silhouette and a cut.

### How do I show the same night from different witnesses?

Reuse the same location so the room is identical, and give any changed look its own state picked only in the flashback segments.

### Can the camera read a note or a receipt?

Not reliably. Generated text is not legible. Have a character read it aloud while the shot shows the object.

### How many suspects should there be?

Four or five. Fewer makes the answer obvious; more makes it hard to keep faces straight across short episodes.

### Is a mystery better as a series or a single film?

A series with a new piece of evidence each episode suits feeds; a single whodunit suits Film and a landscape frame.

### Can I use a famous detective?

No. Known characters trigger SceneMixer's IP check. Create your own detective with their own habits and voice.

### Can AI Create write a mystery that plays fair?

It can draft one from a brief. Check in the summary that the clue the solution needs appears before the reveal, and edit the storyboard if it does not.

Updated November 3, 2026


---

# How to Make a Music Video With AI (and What It Can't Do)

URL: https://scenemixer.com/guides/make-an-ai-music-video/
Make a music video with AI: SceneMixer takes no audio, so build the story half at your song's length, then lay your licensed track under it in an editor. Plan included.

# How to make a music video with AI

Quick answer: SceneMixer does not accept audio. It takes text, up to four photos for an ad, or a video, so you cannot upload a song and get a video cut to it. What you can make is the picture half of a story-led music video: write the song's story as a script, make an episode at about the song's length, download the version without the built-in score, and lay your own licensed track under it in any video editor. It will not make a performance video of someone singing, and it will not cut to the beat for you: characters speak generated lines, and the cuts are yours to place.

Brief for AI Create: A story video for a slow three-and-a-half-minute song about leaving a coastal town: a young woman ties her bicycle to the ferry rail, three flashbacks of the harbour, the ferry pulls away; no dialogue, only the sound of the sea.

## What it can't do, so you don't pay to find out

- Take your song. There is no audio upload. What you can feed in.
- Sing it. Characters speak the lines written in the storyboard, with their lips moving to those words; they do not lip-sync to a track you bring.
- Cut to the beat. Segments are 15 or 30 seconds and the shots inside them have their own seconds; where the picture cuts against your beat is something you set in your editor.
- Print the lyrics. Video models do not draw readable text reliably, and the subtitles feature shows the lines your characters speak, not your lyrics. Lyrics go on in your editor.
What is left is a narrative or concept video: a story, told in pictures, that runs as long as your song.

## Plan the picture to the song's length

| Song | Seconds | Segments (15 s / 30 s) | Type |
| 2:30 | 150 | 10 / 5 | Short Drama |
| 3:30 | 210 | 14 / 7 | Short Drama |
| 4:00 | 240 | 16 / 8 | Short Drama (its limit) |
| 4:30 | 270 | 18 / 9 | TV Series (up to 300 s) |
| 6:00 | 360 | 24 / 12 | Film (no length cap) |
A Short Drama episode runs up to 240 seconds and a TV Series episode up to 300; longer than that, choose Film. Video on SceneMixer is 10 credits a second on the default tier and joining is 1 credit a second, so a 3:30 song is 210 seconds of each, before retakes. Short drama, series or film · Length and cost.

## Map the song, section by section

For a 3:30 song (210 seconds, fourteen 15-second segments):

| Song part | Seconds | Segments | What the picture does |
| Intro | 15 | 1 | The harbour at dawn, no people |
| Verse 1 | 45 | 3 | She packs the bicycle; the kitchen window; the road down |
| Chorus | 30 | 2 | The quay: the ferry, the rail, the knot |
| Verse 2 | 45 | 3 | Three flashbacks of the harbour, one per season |
| Chorus | 30 | 2 | The ferry's rope is thrown; her hand on the rail |
| Bridge | 15 | 1 | The kitchen window, a lit lamp, nobody there |
| Outro | 30 | 2 | The wake widens; the town gets small |
That adds up to 210 seconds and 14 segments. Mark the sections in your script and SceneMixer’s storyboard follows them.

## Write it as a silent film

Say what is on screen and what the place sounds like; leave dialogue out. One segment of 15 seconds, four shots:

[Shot 1] 4s: Wide shot, slow push in. A small harbour at dawn, a ferry at the quay, gulls on the rail. Ambient: water against the hull, a far-off bell. No dialogue.
[Shot 2] 4s: Medium shot. @Nell, 20s, in a yellow raincoat, ties a bicycle to the ferry's deck rail. No dialogue.
[Shot 3] 3s: Close-up on a lit kitchen window, a hand lifting the curtain. No dialogue.
[Shot 4] 4s: Wide shot, the ferry's wake widening behind it. Ambient: engine, wind. No dialogue.

The shot seconds add up to the segment (4 + 4 + 3 + 4 = 15). Dialogue and sound in one generation · Shot seconds and timing.

## Put your track under it

- SceneMixer scores the episode for you by default. Turn the Music switch off for the shots and the composite; downloads then come without the score.
- Download the composite and open it in any video editor, or in the editing tools of the platform you post to.
- Lay your licensed track under the whole piece and cut or retime to the beat there.
Use a track you have the right to use; the Terms do not allow generating protected music without permission, and platforms have their own checks. Background music: the built-in score and your own · Trimming a clip in the editor · Composite and share an episode.

## Pick a look for a song

A music video lives on its look. Painterly Anime suits big scenery and quiet feeling; Watercolor Storybook suits something tender and slow. The look is fixed once you continue past step 1, so choose it before the script is read.

Visual Style: Watercolor Storybook

Visual Style: Soft Glow
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday; Visual Style Watercolor Storybook, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the young woman and the harbour and ferry; no voices are needed for a silent story, and the music switch is the last thing you turn off. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| You want | Possible? | How |
| A story video that runs as long as your song | Yes | Plan in segments, silent shots, your track laid under it |
| Cuts that land on the beat | In your editor | Cut and retime there |
| Someone singing your song | No | Characters speak generated lines |
| Your lyrics on screen | In your editor | Readable text is not drawn reliably |
| Upload the song | No | No audio input |

## FAQ

### Can AI make a music video from a song?

Not from the song itself: there is no audio upload on SceneMixer. You can make the story video and lay your track under it in an editor.

### Can the characters sing my lyrics?

No. Characters on SceneMixer speak generated lines; they do not lip-sync to a track you bring.

### Can it sync to the beat?

No. The timeline in SceneMixer is one card per segment, and fine cutting to a beat happens in your own video editor.

### Can I add lyrics?

Add them in your editor. The subtitles feature on SceneMixer shows the lines your characters speak, and readable on-screen text is not drawn reliably.

### Which look suits a music video?

Whichever fits the mood; Painterly Anime and Watercolor Storybook are common starting points for quiet songs.

### How long can it be?

On SceneMixer, up to 240 seconds as a Short Drama episode, 300 as a TV Series episode, and longer as a Film.

### Is the song's copyright my concern?

Use a track you have the right to use. SceneMixer’s Terms do not allow generating protected music without permission.

### What does it cost?

Every second of video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join, before retakes. A 3:30 song is 210 seconds.

Updated November 3, 2026


---

# AI New Year Video: A Brand Greeting or a Short Story for 2027

URL: https://scenemixer.com/guides/make-an-ai-new-year-video/
Make a New Year video with AI: a 15-second brand greeting with your name on the end card, a Lunar New Year spot set in East Asia, or a one-minute story.

# How to make an AI New Year video

Quick answer: New Year's Day 2027 is Friday, January 1; Lunar New Year is Saturday, February 6, 2027. Two kinds of New Year video come out of SceneMixer. A greeting is an ad project: one photo of your product or shop, a Warm or Epic tone, 15 seconds, and an end card that reads Happy New Year from your brand. A story is a drama project: a one-minute episode about a resolution, a reunion or a countdown, with your own characters. For Lunar New Year, set the Target Region to East Asian and the dialogue to Chinese, and write the red, the lanterns and the family table into the scene.

Greeting line for the Ad Video box: Our family bakery, open since 1987; a thank-you to the neighbourhood and Happy New Year; new hours from January 4. Story brief for AI Create: A one-minute short: at 11:58 on New Year's Eve a courier delivers a letter to the wrong flat and stays for the countdown; one stairwell, one balcony, fireworks off screen.

## Greeting or story: pick the project

| You want | Project | Length | What carries the message |
| A thank-you to customers | Ad Video, Warm tone | 15 seconds | The spoken line and the end card |
| A spectacle with your logo at the end | Ad Video, Epic tone | 15 or 30 seconds | Fireworks and scale in the shots; the end card |
| A Lunar New Year spot | Ad Video, Warm or Epic, Target Region East Asian | 15 or 30 seconds | Red, gold, lanterns, a family table written into the line |
| A short story about a new start | Short Drama, 9:16 | 60–120 seconds | The last shot |

A greeting is cheapest as an ad: one quoted price, one end card, done. The AI ad video generator · An ad from one photo.

## The greeting, from one photo

Upload a photo of the shop front, the team's product or the thing you make. The read tells you what it saw; the one line says who you are thanking and what changes in the new year (hours, a new menu, a new location). Pick Warm for people and feeling, Epic for fireworks and a skyline. Fifteen seconds is one beat: the place, the people, the wish. The end card carries Happy New Year, your brand, and whatever you want them to do next, because nothing legible is drawn into the picture itself. The end card · Which photos work.

## Lunar New Year, set in the right place

For a Lunar New Year video the setting matters more than the date. In a drama project, set Target Region to East Asian in Step 1 so faces, streets and signage follow, and the dialogue language to Chinese; write red lanterns, a round table of twelve, red envelopes, the first firecracker into the scene and the shot. In an ad, write the same into the one line. 2027 is the Year of the Goat; a goat character works in a cartoon look, and animals as characters get a sheet and a voice like anyone else. Target region · Animals as characters.

## The story: a resolution, shot by shot

| Beat | Shot to adapt |
| The last minutes | [Shot 1] 4s: Wide shot, a stairwell at night. @Mara, a courier in a wet jacket, checks a door number and knocks. Ambient: distant countdown from a TV. |
| The wrong door | [Shot 2] 3s: Medium shot. @Teo opens, in socks, a paper hat on his head. @Teo (confused): That's not for me. |
| The choice | [Shot 3] 3s: Close-up on @Mara. She looks at the letter, then at the hat. @Mara: Thirty seconds. Can I watch from your balcony? |
| The countdown | [Shot 4] 4s: Two-shot on a small balcony, city lights below. Both count the last three numbers out loud. The sky flashes off screen. |

Fireworks are best as light on faces, off screen, because a wide shot of a sky full of them is where video models invent. How to write one shot · A love story video.

## Posting on the day

New Year content is posted on three days: December 31 (the countdown), January 1 (the greeting) and the first working day (the new hours). Make the greeting in December and schedule it; make the story earlier still. Shared clips carry the SceneMixer badge; members can download a badge-free file. Posting to TikTok, Shorts and Reels · Video without a watermark.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with one mark each and a voice; the stairwell and the balcony as locations; the letter as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post. For the greeting, use the Ad Video tab instead: one photo, Warm tone, 15 seconds, Happy New Year on the end card.

| Date | Occasion | Video |
| December 31, 2026 | New Year's Eve | The countdown story |
| January 1, 2027 | New Year's Day | The brand greeting |
| January 4, 2027 | First working day | New hours, new menu: the greeting's end card edited |
| February 6, 2027 | Lunar New Year (Year of the Goat) | The East Asian greeting or story |

## FAQ

### How do I make a New Year video with AI?

For a greeting, upload one photo to the Ad Video tab in SceneMixer, pick Warm or Epic and 15 seconds, and put Happy New Year on the end card. For a story, write a one-minute Short Drama.

### Can it say Happy New Year on screen?

On the end card, yes. Text is not drawn into the picture itself; the wish is spoken and written on the card.

### Can I make a Lunar New Year video?

Yes. Set Target Region to East Asian and the dialogue to Chinese in SceneMixer; write lanterns, red envelopes and the family table into the scene.

### Can I use my logo?

The brand name read from your photo or line goes on the end card; the card is built around your product shot.

### How long should it be?

15 seconds for a greeting, 60 to 120 for a story.

### Can I make it in several languages?

Yes. Choose the dialogue language per project; the same greeting can be made again for another market.

### Can fireworks be shown?

Yes, best as light on faces and off screen; wide shots of a sky full of fireworks are where models invent.

### What does a greeting cost?

An ad on SceneMixer is quoted as one figure before it starts and that figure is a ceiling; a 15-second spot is the cheapest length.

Updated November 3, 2026


---

# AI Noir Detective Drama: Rain, Shadows and a Client Who Lies

URL: https://scenemixer.com/guides/make-an-ai-noir-detective-drama/
Make a film noir detective story with AI: Black & White Film, Jazz Age and Vintage Hollywood costumes, rain and neon as light, and a voice-over that carries the case.

# How to make an AI noir detective drama

Quick answer: Noir is a tired detective, a client who lies and a city that is always wet. In SceneMixer, Black & White Film gives the look outright; pair it with Republican Era or Retro Hong Kong scenes for a period city, or Neon Nights for neo-noir in colour. Jazz Age and Vintage Hollywood dress the 1920s to 1950s. The detective's voice-over carries the story: write it as narration in the shot.

Brief for AI Create: A noir detective story: in a rain-soaked port city in 1947, a private eye is hired to find a singer who vanished from a nightclub; the woman who hires him is the singer, using another name; 8 episodes, each opens with his voice-over and ends with a gunshot off-screen.

## The look

| Kind of noir | Visual Style | Scene Style | Costume Style |
| Classic 1940s noir | Black & White Film | Republican Era or New York Metropolis | Vintage Hollywood |
| Prohibition era | Black & White Film or Shot on Film | New York Metropolis | Jazz Age |
| Neo-noir in colour | Shot on Film | Neon Nights | Power Tailoring |
| Hong Kong noir | Shot on Film | Retro Hong Kong | Retro Hong Kong |
Visual style presets.

## Light, rain and the voice-over

- Write the light. “A streetlamp through venetian blinds” gives noir its stripes. Describe the cause.
- Rain is a setting. Set Weather to Rain on the segment; Time to Late night. Segment header.
- The voice-over is narration in the shot, in the detective's voice. Voice-over.
- The gun and the cigarette case are props. Shootings happen off-screen: a flash, a sound, a door. Content rules.

## The client, shot by shot

| Beat | Shot to adapt |
| The office | [Shot 1] 4s: Wide shot, a detective's office at night, streetlight through venetian blinds striping the desk. @Sam Cole sits with his feet up. |
| The voice-over | [Shot 2] 4s: Close-up on the door's frosted glass, a woman's shadow. @Sam Cole (voice-over): She walked in like trouble that had read the menu. (voice: Sam) |
| The client | [Shot 3] 3s: Medium shot. @Vera enters in a fitted dark dress, rain on her shoulders. |
| The lie | [Shot 4] 4s: Close-up on @Vera. @Vera (low): I'm looking for my sister. (voice: Vera) |

## Story shapes that work

- The missing person who is the client.
- The frame-up.
- The crooked city: every lead goes back to city hall.
- The femme fatale with her own reasons.
- The last case before he quits.
Murder mysteries · Police dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style New York Metropolis, Republican Era; Costume Style Vintage Hollywood; Visual Style Black & White Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the detective and the client with one mark each; the gun and the photo as props; the office, the club and the docks as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The office. | A woman hires him to find a singer. |
| 2 | The nightclub. | The singer's dressing room has been cleaned. |
| 3 | The docks. | A gunshot, off-screen. |
| 4 | His office. | A photo of the singer: it is his client. |
| 5 | City hall. | The man who owns the club owns the mayor. |

## FAQ

### How do I make a noir detective story with AI?

Choose Black & White Film as the SceneMixer visual style, a period city scene style and Vintage Hollywood or Jazz Age costumes, and carry the story with the detective's voice-over.

### Can I make noir in colour?

Yes. Neo-noir: Shot on Film with Neon Nights.

### How do I get rain?

Set Weather to Rain on the segment header in SceneMixer.

### How do I add the voice-over?

Write it as narration in the shot, in the detective's voice.

### Can I show a shooting?

Off-screen: a flash, a sound, a door. Graphic violence is blocked.

### Which costume style suits the 1940s?

Vintage Hollywood; Jazz Age for the 1920s.

### Vertical or landscape?

16:9 Landscape suits noir framing; 9:16 Portrait works for feeds.

### Can I copy a famous detective?

No. Known characters trigger the IP check. Invent your own.

Updated November 3, 2026


---

# AI Nollywood-Style Drama: Lagos Homes, Family Fights, Big Reveals

URL: https://scenemixer.com/guides/make-an-ai-nollywood-style-drama/
Make a Nollywood-style family drama with AI: Target Region Africa, African Prints and West African Regal costumes, accented English lines, and the big confrontation.

# How to make an AI Nollywood-style drama

Quick answer: Nollywood melodrama runs on family, money, faith and betrayal, with confrontations played loud and long. In SceneMixer, set Target Region to Africa so names, faces and streets match, write the dialogue in English, dress the cast in African Prints every day and West African Regal for the wedding and the burial, and let the voice description carry the accent. Yoruba, Igbo, Hausa and Pidgin are not among the dialogue languages.

Brief for AI Create: A family drama in Lagos: on the night of his fiftieth birthday party, a rich man's first wife returns from the village with the son he never acknowledged; fifteen episodes, one house, one inheritance.

## Region, language and voice

| Setting | Choose |
| Target Region | Africa: names, faces, streets and visible text follow it |
| Dialogue language | English (the lines are performed as written) |
| Accent | In each character's voice description, e.g. a deep Nigerian English accent, when no preset voice is attached |

Target region · Give a character an accent.

## Dress the occasion

Costume Style: West African Regal

Costume Style: Wedding

Scene Style: Suburban Mansion

Scene Style: Urban Opulence

Scene Style: Street Life
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

African Prints for home and market, West African Regal for weddings, chieftaincy and burials, Power Tailoring for the boardroom. Give the leads a state for each occasion so the face stays and the outfit changes. States.

## The places

| Place | Scene Style |
| The rich husband's house | Suburban Mansion |
| The company, the hotel party | Urban Opulence |
| The market, the bus stop | Street Life |
| The village compound | Described in the location: zinc roofs, a courtyard, a mango tree |

## The confrontation, shot by shot

| Beat | Shot to adapt |
| The party | [Shot 1] 4s: Wide shot, a garden party with gold chairs, guests in lace and agbada. |
| The arrival | [Shot 2] 3s: Medium shot. @Ngozi walks in through the gate holding a young man's hand. |
| The host | [Shot 3] 2s: Close-up on @Chief Okafor; his glass stops halfway. |
| The line | [Shot 4] 4s: @Ngozi (loud, to the whole garden): Tell them who this is. Tell them! |
| The guests | [Shot 5] 3s: The guests turning, phones coming out. |

Big scenes, loud lines, reaction shots of the crowd: the format's own grammar. Background people.

## Stories that travel

Inheritance and the second wife (family inheritance), the poor girl and the rich family (Cinderella), betrayal and payback (revenge). For Indian family drama see Bollywood-style.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Suburban Mansion, Street Life; Costume Style African Prints, West African Regal; Visual Style Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Target Region Africa; each lead with an everyday and an occasion state; accents in the voice descriptions; the house, the party garden and the village compound as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Beat |
| 1 | The birthday party, the arrival |
| 3 | The family meeting |
| 6 | The pastor's visit |
| 10 | The will |
| 15 | The burial and the truth |

## FAQ

### How do I make a Nollywood-style drama with AI?

Set Target Region to Africa in SceneMixer, write English dialogue with an accent in the voice description, dress the cast in African Prints and West African Regal, and play the confrontations big.

### Can the characters speak Yoruba, Igbo, Hausa or Pidgin?

Not as a dialogue language on SceneMixer. Write English and set the accent in the voice description.

### How do I give a character a Nigerian accent?

Write it in the voice description; it works when no preset voice or recording is attached.

### Which costumes suit it?

African Prints for everyday, West African Regal for occasions, Wedding for the wedding.

### Can I set it in Accra or Nairobi?

Yes; Target Region Africa, and name the city in the script.

### Which format suits it?

Short Drama in 9:16 for feeds, or TV Series in 16:9 for YouTube.

### Which look suits it?

Digital HD.

Updated November 3, 2026


---

# AI Office Romance Drama: Meetings, Elevators and Secrets

URL: https://scenemixer.com/guides/make-an-ai-office-romance-drama/
Make an office romance drama with AI: Corporate Cold sets, Power Tailoring and Smart Casual, the elevator and meeting room as stages, and a secret to keep.

# How to make an AI office romance drama

Quick answer: An office romance is a love story with an audience: colleagues everywhere, a secret that cannot get out. In SceneMixer, build the office with Corporate Cold, dress the cast in Power Tailoring and Smart Casual, use a few recurring places (the elevator, the meeting room, the rooftop) as stages, and put colleagues in the background of every shot where a secret could slip.

Brief for AI Create: An office romance: the new head of design discovers the anonymous illustrator she has been messaging for a year is the junior copywriter who sits behind her; twelve episodes, one floor.

## The floor plan is the story

| Place | What happens there |
| The open-plan floor | Everyone watching; nothing can be said |
| The elevator | Thirty seconds alone |
| The meeting room | Professional words, personal meaning |
| The rooftop or stairwell | Where the truth comes out |
| The bar after work | The almost |
Each is drawn once and reused. Glass walls in Corporate Cold make being seen part of every scene. The sample Çantasındaki Sözleşme plays in an open-plan office. Same room in every episode.

## Colleagues as background

Write the office around the leads: colleagues at their desks behind glass, one of them glancing up. Named gossips get their own sheets; the rest are scenery. Background people · Clear or add background people.

## The elevator, shot by shot

| Beat | Shot to adapt |
| The doors | [Shot 1] 3s: Medium shot. @Hana steps into the elevator; @Sam is already inside. The doors close. |
| The silence | [Shot 2] 3s: Two-shot, side by side, both looking at the floor numbers. No dialogue. Ambient: elevator hum. |
| The line | [Shot 3] 3s: @Sam (not looking at her): You draw your sevens with a line through them. |
| The doors again | [Shot 4] 3s: The doors open on the full office; @Hana walks out fast. |

## Story shapes

- The secret identity: online friends who are colleagues.
- Boss and assistant: billionaire romance is the high-gloss version.
- The rival promotion: enemies to lovers.
- The fake relationship: contract marriage.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Corporate Cold; Costume Style Power Tailoring, Smart Casual; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: two leads and one or two named colleagues; the floor, the elevator and the roof as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Place | The ending beat |
| 1 | Her first day | His sketchbook on the desk |
| 3 | The elevator | A detail only the illustrator would know |
| 5 | The pitch meeting | He presents her idea, credited to her |
| 8 | The office party | Someone sees them on the roof |
| 12 | The open floor | She tells everyone |

## FAQ

### How do I make an office romance drama with AI?

Draw the office places once, dress the cast in Power Tailoring and Smart Casual, and keep colleagues in the background where secrets could slip.

### Which scene style suits an office?

Corporate Cold for glass and steel; Nordic Minimal for a startup.

### How do I add coworkers without many sheets?

Write them as background people; give only named gossips their own sheets.

### Which format suits it?

Short Drama, 9:16 Portrait, one to two minutes per episode.

### Which look suits it?

Digital HD; Korean Webcomic for a webtoon feel.

### Can the office be the same in every episode?

Yes. Each location is drawn once on SceneMixer and used as the set.

### Can I set it in another country?

Yes; set Target Region and dialogue language in SceneMixer's Step 1.

### How many episodes?

Ten to twenty short episodes is typical.

Updated November 3, 2026


---

# AI Ottoman Palace Drama: Sultans, Viziers and a Court of Tile and Silk

URL: https://scenemixer.com/guides/make-an-ai-ottoman-palace-drama/
An Ottoman court drama with AI: the Ottoman Palace scene style, gold-brocade court dress, Turkish dialogue, and viziers, a queen mother and a succession told off screen.

# How to make an AI Ottoman palace drama

Quick answer: An Ottoman palace drama is a story about rooms: the council chamber, the women’s quarters, the courtyard, the kitchens, each with its own rules about who may speak. In Step 1 pick the scene style Ottoman Palace — Istanbul palaces and mansions, with blue-white and turquoise floral tiles on the walls, layered low seating and gilding — and the costume style Ottoman Court, robe over robe in gold brocade with big flower patterns, tall turbans and jewels. Set the dialogue language to Turkish, and Target Region to Middle East (a Turkish setting gets Turkish names) or Follow script. Then tell the cruel parts the way the genre does: by the closed door, not by the act.

Brief for AI Create: An Ottoman palace drama: when the sultan falls ill, his mother and his grand vizier each quietly back a different son; a junior scribe who copies both sets of letters is the only one who knows; 8 short episodes, Turkish dialogue, every episode ends on a sealed letter.

## The choices that make it Ottoman

| Step 1 setting | Choose | Why |
| Scene Style | Ottoman Palace | Istanbul palaces and mansions of the 16th to 19th centuries: tiled walls, low seating in layers, gilding. |
| Costume Style | Ottoman Court | The Topkapi court: robe over robe in gold brocade, tall turbans, jewels. |
| Visual Style | Digital HD, Shot on Film or Classical Oil Painting | Live action for a series feel; the oil-painting look if you want a miniature-painting richness. |
| Target Region | Middle East or Follow script | Middle East gives Turkish names and faces for a Turkish setting; Follow script keeps what your script says. |
| Dialogue language | Turkish | Spoken in Turkish as each clip renders. English lines work too. |
All of this is written into the scene and character descriptions when the script is read, so it is chosen first. Why they are fixed afterwards · Every scene and costume style.

## Court, bazaar and Gulf robes are not the same thing

Several styles sound close and are not. Use the right one for each place and person:

- Ottoman Palace and Ottoman Court — the palace and the people who live in it.
- Middle Eastern Souq — the old-city market lanes under vaults, for the street outside the palace wall.
- Middle Eastern Robes — the white and black robes of the Gulf Arab world, not Ottoman court dress.
- Persian Court and Indian Royal — other courts, for a visiting ambassador or a rival dynasty.

## Titles and names

An Ottoman court is full of titles: the sultan, the queen mother, the grand vizier, agas, a chief physician. Tie each title to one person in the script the first time it appears (“Ibrahim, the Grand Vizier”). When the script is read, the character card keeps the other names your text uses for that person on its grey Also called line (read-only; there is nothing to type), so “the Grand Vizier” in a later episode is found as the same person. A title that could fit several people is not kept. Also called.

Invent your court. A fictional sultan with a fictional name is fine, and characters from a known series can trigger the IP check. If a clip is blocked.

## Rooms and what each is for

| Room | The kind of scene |
| The council chamber | Viziers in a row, a petition, a decision announced. |
| The queen mother’s rooms | A conversation with a servant in the room and a plan under it. |
| A tiled courtyard | An ambassador kept waiting; two rivals who must greet each other. |
| The palace kitchens | A cup, a tray, a name written on a cloth. |
| The scribes’ room | A letter copied twice. |
Each place gets its own scene card in Step 2, and each segment stays in one place and one stretch of time. A segment stays in one place and one timeline.

## Power without the violence

Palace stories end people by order, not by fight: an exile, a silk cord, a sealed room. The video checks stop graphic violence, and the genre does not need it.

- A death sentence: the decree read, the man bowing his head, the corridor emptying.
- A poisoning: the cup carried in, the cup carried out.
- The women’s quarters: a court of rivals, mothers and sisters, played as politics and family. Keep it free of anything sexual.

## A council scene, shot by shot

| Beat | Shot to adapt |
| The chamber | [Shot 1] 4s: Wide shot, fixed camera. The imperial council chamber, tiled walls in blue and white, viziers seated in a row along the low divans. @Kerem, the sultan, sits raised at the far end, still and pale. |
| The vizier | [Shot 2] 3s: Medium shot on @Kemal, the grand vizier, rising with a sealed paper in both hands. |
| The line | [Shot 3] 4s: Close-up on @Kemal. @Kemal (measured, Turkish): Padişahım, divan toplandı. (voice: Kemal) |
| The scribe | [Shot 4] 3s: Close-up on @Selim, the young scribe by the door, whose hand has stopped moving on the page. |
| The silence | [Shot 5] 3s: Medium shot on the veiled lattice above the chamber. A figure behind it is standing very still. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- The succession: an ailing ruler, two heirs, two backers, one court that must pretend nothing is wrong.
- The scribe who copies both sides: the smallest person in the room holds the largest secret.
- The ambassador: a foreign envoy, a gift, and a message hidden in how it is presented.
- The queen mother: power exercised in a single sentence to a servant.
- The physician: a cure that decides who rules.
Other palaces: Joseon palace drama · History drama · Period romance.

## Make it in SceneMixer, start to finish

- Script: paste your own (Turkish or English) on Upload Script, or write the brief in AI Create.
- Step 1: Type Short Drama or TV Series; scene style Ottoman Palace; costume style Ottoman Court; a visual style; Target Region; Turkish dialogue; 9:16 Portrait or 16:9. Read the script.
- Step 2: a specific face, an age and the clothes of their office for each character. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). One place per segment.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | The council meets without the sultan’s voice. | A sealed paper is not handed to the vizier. |
| 2 | The queen mother’s quiet instruction. | The servant bows and leaves with the wrong tray. |
| 3 | The scribes’ room. | The same hand wrote both letters. |
| 4 | An ambassador kept waiting in the courtyard. | The gift is opened and is empty. |
| 5 | The physician is summoned at night. | He refuses to say who sent for him. |

## FAQ

### How do I make an Ottoman palace drama with AI?

In SceneMixer's Step 1, choose the scene style Ottoman Palace and the costume style Ottoman Court, set the dialogue language to Turkish, then write or generate the script and render episode by episode.

### What does the Ottoman Palace scene style look like?

Istanbul palaces and mansions of the 16th to 19th centuries: blue-white and turquoise floral tiles, layered low seating and gilding.

### Can the characters speak Turkish?

Yes. Turkish is one of the dialogue languages on SceneMixer; choose it in Step 1 and the lines are spoken on screen. English lines work too.

### Which Target Region should I pick?

Middle East gives Turkish names and faces for a Turkish setting. Follow script keeps whatever the script already says.

### Is Middle Eastern Robes the right costume?

No. Those are the white and black robes of the Gulf Arab world. Ottoman court dress is the Ottoman Court style.

### Can I show an execution or a poisoning?

Not the act. Show the decree, the cup and the empty corridor. The video checks on SceneMixer stop graphic violence.

### Can I use a real sultan or a famous series' character?

Use invented characters. Characters from known series can trigger the IP check, and an original sultan keeps the story yours.

### Will the vizier look the same in every episode?

Yes. Each character is drawn once and that sheet is used in every shot.

Updated November 3, 2026


---

# AI Period Romance Drama: Regency and Victorian, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-period-romance-drama/
Ballrooms, letters, a season and a scandal: how to write a Regency or Victorian romance, pick period sets and costumes, and film the ball without losing faces.

# How to Make an AI Period Romance Drama

Quick answer: A period romance is a story about rules: who may speak to whom, who may marry whom, and what one glance across a ballroom costs. In SceneMixer, choose the era in step 1 with the set and wardrobe rows — English Manor with Regency, or European Palace with Victorian, Soft Glow for the romantic haze — then build the ball the way a film would: a few named people in focus and the rest of the room in the segment’s background line. Give each ball gown its own state, let letters be read aloud rather than shown, and write the candlelight into every evening shot.

Brief for AI Create: A Regency romance: a vicar's daughter with no dowry is hired to teach a duke's ward to dance before her first season; the duke has sworn never to marry; a scandal at the final ball forces his hand; witty, restrained, swooning.

## The conventions viewers come for

- The season. A fixed run of balls and calls with a marriage at the end of it. It is the genre’s clock.
- The rule and the risk. A dance too many, a walk without a chaperone, a letter in the wrong hands. Small acts have enormous consequences, which is what makes them thrilling.
- The mismatched pair. Rank against wit, money against pride. They spar before they admit anything.
- The hand. In a world where touching is forbidden, a gloved hand is a love scene. The genre is at its best when it is restrained.
- The scandal. Gossip, a rival, a family secret. The public is always watching, and the ending is always in front of it.

## Settings to pick in step 1

- Visual Style: Soft Glow for a hazy, romantic period feel; Shot on Film for a richer, darker grade; Classical Oil Painting or Watercolor Storybook for a painted look.
- Scene Style: English Manor for Regency-era country houses and London townhouses (Regency romance guide), European Palace for continental ballrooms and courts, Medieval Castle for something older and wilder, Parisian Streets for a continental chapter, Pastoral Healing for country estates and gardens.
- Costume Style: Regency for empire waistlines and tailcoats, Victorian for later, more structured dress, European Court Gown for a royal court. The scene and costume styles are read with the script and fixed afterwards. When they take effect.
- Type: TV Series for room to linger; Short Drama for a fast feed version with a hook at every episode’s end.
- Target Region: Western, or Follow script if your script already places everything in a named county and city.

Visual Style: Shot on Film

Scene Style: European Palace

Scene Style: Parisian Streets

Costume Style: Regency

Costume Style: Victorian

Costume Style: European Court Gown
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Gowns, coats and the second outfit

A character sheet shows one outfit, and the reference image locks it. A period romance changes clothes constantly — a morning dress, a riding habit, the gown for the final ball — so decide which changes matter on screen and give those their own image:

- Generate the character’s main image in the outfit she wears most.
- On her card choose Add a state, name it (Final ball gown), and describe only the difference: an ivory silk gown with a high waist and short puffed sleeves, long white gloves, pearls woven into her hair.
- Press Generate this state; then, in the editor for the ball episode, open her from the episode’s assets and choose Use this state in this episode (or Use this state in this shot for single segments).
Minor variations — a shawl over the same dress, gloves on or off — are written into the shot instead. One outfit per character sheet · Use a state in one scene.

## Filming a ball without losing the faces

A ballroom is the genre’s big set piece and the easiest place to lose consistency: forty dancers are forty faces. Film it as a cinema would:

- Name only the people the scene is about. The couple, the rival, the mother watching from the wall. Everyone else is the segment’s background line: couples dancing in rows, guests along the walls. Background people.
- Two people per shot. A dance is a series of close two-shots, over-the-shoulder looks and a wide establishing frame — not a choreographed wide shot of everyone.
- Let the light be candles. Chandeliers of lit candles overhead, warm light on the dancers’ faces, the windows dark gives a period evening; “a beautiful ballroom” gives a hotel.
- One room, every episode. The ballroom and the drawing room are locations with their own images, so they look the same at every ball. Same room in every episode.
Crowds and group scenes.

## Letters, rings and the language of objects

Period stories run on objects: a letter, a miniature portrait, a pressed flower, a ring. Give the recurring ones their own prop image so they are recognisable, and remember that a generated letter will not hold legible handwriting. Let the reader say it, or let a voice-over read it as she reads:

| Beat | Shot to adapt |
| The letter arrives | [Shot 2] 4s: Close-up, candlelight. @Eleanor breaks the seal on the @Letter; her hand is not quite steady. No dialogue. Ambient: the fire, rain on the glass. |
| The words | [Shot 3] 6s: Medium close-up. In frame only: @Eleanor, reading. @Thomas (voice-over, low): I have tried to stay away, and I find I am a poor student of my own resolve. |
| The hand | [Shot 9] 3s: Close-up on two gloved hands that almost touch on the balustrade, then do. No dialogue. Ambient: the orchestra through the open doors. |
Props and reference images · Lines, voice-over and captions.

## Voices and the way people spoke

Formality is characterisation. Write it into each Voice description: clipped, cool, perfectly polite for the duke; bright, quick, laughs before she can stop herself for the heroine; slow, honeyed, never says what she means for the rival. Keep period vocabulary in the dialogue itself and keep lines short enough to say in the shot’s seconds — about three words a second in English. How character voices work.

## Make it in SceneMixer, start to finish

- Script: paste yours on Upload Script, or a brief in AI Create with the season, the pair, the obstacle and the ending.
- Step 1: the period settings above; read the script.
- Step 2: generate the cast, the ballroom, the drawing room and the garden; add the gowns that matter as states; give the letter and the ring their images.
- Storyboard: episode 1 first (free on your first project). Check that ball scenes name only the people they are about.
- Video: generate on the default tier at 10 credits a second, regenerate the shots that miss, composite and post.

| Episode | What happens | The hook |
| 1 | The duke’s household hires a dancing teacher with no family name worth knowing. | He watches the first lesson from the gallery and leaves before she sees him. |
| 2 | The first call of the season; a rival notices how the duke looks at the teacher. | A letter with no signature is slipped under her door. |
| 3 | The first ball. One dance, one too many. | Her name is in the morning’s gossip sheet. |
| 4 | She is dismissed to protect his ward’s prospects. | He rides after the coach. |
| 5 | The final ball, in front of everyone. | He crosses the room and asks for the dance no one expects. |

## FAQ

### Can I make a Regency romance with AI?

Yes. Choose English Manor as the scene style and Regency as the costume style in SceneMixer's step 1, then write or generate the script. The cast, rooms and costumes are kept consistent across episodes.

### How do I change a character's gown for the ball?

Make the gown a state of her character with its own image and assign it to the ball episode or to specific segments. Her other scenes keep the main outfit.

### Why does my ballroom look like a hotel?

Because the shot does not say what lights it or what period it is. Write candles, chandeliers, dark windows and the dancers' dress into the shot.

### How do I show a crowded ball without faces changing?

Name only the people the scene is about and put the rest of the room in the segment's background line. Film the dance in close two-shots.

### Can the camera read a letter?

Not reliably. Generated handwriting is not legible. Have the character or a voice-over read it aloud while the shot shows the letter and the reader.

### Can I base my story on a famous period novel or show?

Write your own story and characters. Well-known franchises and their characters trigger the IP check; public-domain classics still need care with any modern adaptation's names and designs.

### Which style gives the soft, romantic period look?

Soft Glow. Shot on Film gives a darker, richer grade; Classical Oil Painting a painted one. The style on SceneMixer is fixed after step 1.

### How long should each episode be?

One to two minutes for a feed series on Short Drama; a few minutes on TV Series if the story needs room.

Updated November 3, 2026


---

# AI Pirate or Viking Adventure: Ships, Storms and Longhouses

URL: https://scenemixer.com/guides/make-an-ai-pirate-or-viking-adventure/
A pirate or Viking adventure with AI: harbour and fjord scene styles, pirate and Norse costume, storms set per scene, and fights told by what comes before and after.

# How to make an AI pirate or Viking adventure

Quick answer: Pirates and Vikings share a stage: the ship, the water and a crew in rough clothes, and in both cases the trick is to keep the sea out of the shot list. In Step 1 pick the scene style Age of Sail Harbor (wooden decks and rigging, colourful colonial houses, a turquoise sea) with the costume style Pirate, or Viking Fjord (steep cliffs, deep blue water, turf-roofed longhouses with carved dragon ridges) with Nordic Viking. Use Follow script for the region so names stay as you wrote them. Then stage the adventure on a deck, in a harbour and in a hall, set the weather to Storm when you need one, and tell every fight by the moment before and the moment after.

Brief for AI Create: A pirate adventure: a ship’s cook who can read finds the captain’s map is a forgery, and has to decide, in a harbour full of people who want it, whether to say so; 8 short episodes, English dialogue, every episode ends on a ship leaving a harbour.

## Pick your world

| Step 1 setting | Pirates | Vikings |
| Scene Style | Age of Sail Harbor | Viking Fjord |
| Costume Style | Pirate | Nordic Viking |
| Costume extras | Fantasy Adventurer for a fantasy crew | Celtic Tribal for a rival people |
| Visual Style | Digital HD, Shot on Film; Family 3D Animation for a family pirate tale | Shot on Film, Game Engine Realism, Digital HD |
| Target Region | Follow script keeps your names; Western moves the cast to Western names |
| Dialogue language | English, or any other supported language |
What the two scene styles give you: Age of Sail Harbor is a 17th–18th-century Caribbean colonial harbour with wooden decks and rigging, colourful stucco houses and a turquoise sea. Viking Fjord is a Viking-age fjord settlement with steep cliffs around deep blue water, timber longhouses with carved dragon ridges and wooden docks. Every scene and costume style.

All of this is written into the descriptions when the script is read, so choose it first. Why they are fixed afterwards.

## Keep the ocean out of the shot list

An open sea with a fleet is the hardest picture to ask for. Build the adventure on places a camera can stand:

- The deck. Wood underfoot, rigging overhead, a rail, a wheel. Most of a ship story is people on a deck talking.
- The harbour. Where the ship arrives, the deal is made and the rumour starts.
- The hall or the tavern. Oaths, bargains and betrayals indoors, with a fire.
- The shore. A boat pulled up, a figure walking away along a beach or a jetty.
- The ship itself as a wide shot only: leaving, arriving, a silhouette on a horizon.
Each place needs its own scene card in Step 2, and each segment stays in one place and one stretch of time. A segment stays in one place and one timeline.

## Storms, fog and night

Weather is a setting, not a sentence. In the SceneMixer storyboard, each segment header has a Time and a Weather row: Time runs from Dawn to Late night, and Weather offers Clear, Cloudy, Overcast, Rain, Storm, Snow, Fog, Windy and Sandstorm. Change on the header applies your choice to the run of segments in that scene, it is free, and it takes effect the next time those segments are generated.

Wet clothes follow the story: if a character is soaked in one segment, the next segment starts with them wet until the text dries them off. What the system keeps track of between segments · The segment header, field by field.

## Fights without the gore

Swords, axes and cannon are fine to have; wounds and blood on screen are what the video checks stop. Tell the fight by its beats:

- The draw: a blade leaving its scabbard, two stances, a held breath.
- The clash: sparks, a block, a boot sliding on a wet deck.
- The result: a figure falling out of frame, a sail coming down, a shield split on the ground.
- The raid: torchlight on a hall, then the morning after.
Invent your crew and your clan. Characters from known films and series can trigger the IP check. If a clip is blocked.

## A deck scene, shot by shot

| Beat | Shot to adapt |
| The deck | [Shot 1] 4s: Wide shot, fixed camera. The deck of a three-masted wooden ship at dawn, rigging overhead, a colourful colonial harbour behind. @Tamsin, the cook, stands at the rail holding a rolled map. |
| The map | [Shot 2] 3s: Close-up on @Tamsin’s thumb running along the edge of the paper. The ink is the wrong colour. |
| The captain | [Shot 3] 3s: Medium shot. @Rooke, the captain, comes up the stairs from below, tying his sash. |
| The line | [Shot 4] 4s: Close-up on @Tamsin, steadying her voice. @Tamsin (low): Captain, the map is a copy. (voice: Tamsin) |
| The silence | [Shot 5] 3s: Close-up on @Rooke. He smiles slowly, and looks past her at the harbour. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- The forged map: a prize that is not what it seems, and one person who knows.
- The mutiny: a crew that votes with its feet before it votes with its hands.
- The exile and the oath: a clan, a broken promise, a ship that will not wait.
- The rival crew: two ships in one harbour, one dock between them.
- The saga: three generations, one hall, a feud no one remembers starting.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief in AI Create.
- Step 1: Type Short Drama, TV Series or Film; the scene and costume styles above; a visual style; Target Region; dialogue language; 9:16 Portrait for feeds or 16:9 Landscape for YouTube and film. Read the script.
- Step 2: a specific face and a mark per character (an eyepatch, a scar, a braid); the map or the sword as a prop. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). Set Weather and Time on the headers of the storm and night scenes.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | A ship’s deck at dawn. | The cook notices the map is a copy. |
| 2 | A tavern in the harbour. | Someone else is asking after the same map. |
| 3 | The storm. | The captain is not on the deck. |
| 4 | The rival crew’s ship alongside. | The cook’s name is on their list. |
| 5 | A beach at night. | A boat pulled up with no one in it. |

## FAQ

### How do I make a pirate story with AI?

In SceneMixer's Step 1, choose the scene style Age of Sail Harbor and the costume style Pirate, then write or generate the script. Stage it on decks, in harbours and in taverns rather than on open sea.

### How do I make a Viking story with AI?

Choose the scene style Viking Fjord and the costume style Nordic Viking in SceneMixer. Longhouses, docks and the shore carry most of a saga.

### Can I film a storm at sea?

Set Weather to Storm on the segment header of the scene. It applies to the run of segments in that scene and takes effect the next time they are generated.

### Can I show sword fights?

Show the draw, the clash and the result. Graphic wounds and blood are stopped by the video checks.

### Do the characters stay wet from one segment to the next?

In newer storyboards, yes: a soaked character starts the next segment wet until the text dries them off.

### Which Target Region should I pick?

Follow script keeps the names you wrote. Western moves the cast to Western names.

### Can I use a famous pirate or a Viking king from a known series?

Use invented characters. Characters from known films and series can trigger the IP check.

### Can I make it for kids?

Yes. Pick a family-friendly visual style such as Family 3D Animation and keep the fights comic.

Updated November 3, 2026


---

# AI Police and Crime Drama: Detectives, Interrogations and Cases

URL: https://scenemixer.com/guides/make-an-ai-police-or-crime-drama/
Make a police or crime series with AI: Forensic Cold White interrogation rooms, Professional Uniforms, clue props, a case per episode and violence kept off-screen.

# How to make an AI police or crime drama

Quick answer: A police drama is a case, a clock and a room where someone lies. In SceneMixer, the Forensic Cold White scene style is built for it: white tile, stainless steel, one-way glass and tube lights under grid ceilings. Dress the force in Professional Uniforms and the detectives in Power Tailoring or Smart Casual. Make every clue that comes back a prop with its own picture, and show the crime through what was left behind, not the act.

Brief for AI Create: A police drama: a rookie detective is partnered with the officer her father put in prison 20 years ago; their first case is a missing girl whose last message was sent from a police station; 12 short episodes, each ends with a new piece of evidence in a sealed bag.

## Set the precinct

| Kind of case | Scene Style | Costume Style |
| Modern procedural | Forensic Cold White | Professional Uniforms, Power Tailoring |
| Big-city crime at night | New York Metropolis or Neon Nights | Smart Casual, Streetwear |
| A small-town murder | American Small Town | Soft Everyday |
| Hong Kong cops and triads | Retro Hong Kong | Retro Hong Kong |
| The hospital case | Hospital Ward | Professional Uniforms |
Murder mysteries · Noir detectives.

## Clues are props

A phone, a ring, a photo, a key: if the case turns on it, make it a prop in Step 2 so it looks the same in the evidence bag in episode 2 and on the table in episode 9. A wall of photos and string is set dressing written in the shot. The same prop in every hand · Describing props.

Crime scenes: show the room, the tape, the chalk-white light and the detective's face, not a body. Violence and injury are kept off-screen by content review. Content rules.

## The interrogation, shot by shot

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, interrogation room, white tile, one-way glass on the left wall, a steel table. @Dana sits opposite @Victor. |
| The evidence | [Shot 2] 3s: Close-up. @Dana slides a sealed evidence bag with a silver locket across the table. |
| The tell | [Shot 3] 3s: Close-up on @Victor's hand closing into a fist. |
| The line | [Shot 4] 4s: Close-up on @Victor. @Victor (flat): I've never seen that in my life. (voice: Victor) |
| Behind the glass | [Shot 5] 3s: Medium shot behind the one-way glass. @Captain Reyes folds her arms. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- Case of the week: one crime per episode, one secret across the season.
- Partners who hate each other: the oldest engine in the genre.
- The dirty cop: the call is coming from inside the precinct.
- The cold case: twenty years, one new piece of evidence.
- Undercover: a detective too deep inside a gang.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Forensic Cold White; Costume Style Professional Uniforms, Power Tailoring; Visual Style Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: detectives and suspects with one mark each; every recurring clue as a prop; the precinct, the interrogation room and the scene as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The precinct. | Her new partner is the man from her father's case file. |
| 2 | The girl's bedroom. | A second phone hidden in a shoebox. |
| 3 | The interrogation room. | The suspect asks for her by her father's name. |
| 4 | The pier at night. | The tracker stops moving. |
| 5 | The evidence room. | The original case box is empty. |

## FAQ

### How do I make a police drama with AI?

Pick Forensic Cold White or a city scene style in SceneMixer's Step 1, Professional Uniforms or Power Tailoring for costumes, make the clues props in Step 2, and render one case at a time.

### What does the Forensic Cold White style look like?

Police and court spaces: white tile, stainless steel, one-way glass and rows of tube lights under grid ceilings.

### Can I show a crime scene?

Yes, the room and the evidence. Not the body or the violence; those are blocked by content review.

### How do I keep a clue the same across episodes?

Make it a prop with a reference picture in SceneMixer's Step 2 and tag it wherever it appears.

### Are police badges and logos real?

Keep them invented. Real logos and brand marks can trigger the IP check.

### Can I base it on a real case?

Do not depict real people. Invent the victims, suspects and officers.

### Which visual style suits a crime drama?

Digital HD for a modern procedural, Shot on Film for grit, Black & White Film for noir.

### How long should each episode be?

For a short-drama feed, one to two minutes with one reveal; for a series, one case per episode.

Updated November 3, 2026


---

# AI Post-Apocalyptic Drama: Ruins, Factions and Scarcity

URL: https://scenemixer.com/guides/make-an-ai-post-apocalyptic-drama/
Make a post-apocalyptic drama with AI: Wasteland or Frozen Wilderness worlds, salvaged costumes, wear as a state over time, and stories built on what is scarce.

# How to make an AI post-apocalyptic drama

Quick answer: A post-apocalyptic story is about what people do when something essential runs out: water, fuel, medicine, trust. In SceneMixer, build the world in Step 1 with Wasteland or Frozen Wilderness and Wasteland Scavenge costumes, make time visible as worn states of each character, and give every faction one look so viewers know who is who in a wide shot.

Brief for AI Create: Twenty years after the grid failed, a radio operator in a snowbound outpost hears a broadcast from a city everyone says is gone; she trades the outpost's last fuel to follow it; ten episodes.

## Pick the kind of ending the world had

| World | Scene Style | Costume Style |
| Desert collapse | Wasteland | Wasteland Scavenge |
| Endless winter | Frozen Wilderness | Outdoor Gear, Inuit Arctic |
| Off-world colony | Mars Base | Space Explorer |
| Drowned city | Wasteland with water written into locations | Outdoor Gear |
| Tech remnants | Cyber Metropolis ruins | Cyberpunk Augment |
For outbreak stories see zombie apocalypse; for one person against nature, survival drama.

## Factions you can tell apart

In a wide shot of a market or a convoy, viewers need to know who belongs where. Give each faction one colour and one mark (red cloth on the arm, goggles, a painted helmet) and write it into every member's description. Named members get their own sheets; the rest are background. Background people.

## Time shows on the body

Months on the road change a face: thinner, sunburnt, a beard, a scar. Add those as states and switch to them as the story moves on; the model will not age anyone from words alone. Use a character state.

| Beat | Shot to adapt |
| The broadcast | [Shot 1] 4s: Medium shot, a radio room lit by one bulb. @Vera turns a dial; static, then a voice. Ambient: wind against the walls. |
| The trade | [Shot 2] 3s: Close-up. @Vera's gloved hands push a fuel can across a table to @Brask. |
| The road | [Shot 3] 4s: Extreme wide. One sled crossing white flats toward a broken skyline. |

## Story engines

- The scarce thing: each episode, someone needs it and someone has it.
- The rumour of a safe place: the journey is the season.
- The settlement's rules: who decides, and what happens to people who break them.
- The old world's secret: someone knows why it ended.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Wasteland, Frozen Wilderness; Costume Style Wasteland Scavenge, Outdoor Gear; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: one mark per faction; worn states for later episodes; the radio room, the road and the city as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Scarce thing | The ending beat |
| 1 | Fuel | The broadcast repeats a name Vera knows |
| 2 | Shelter | The sled breaks in a whiteout |
| 3 | Trust | A stranger knows where the city is |
| 4 | Medicine | The stranger is sick |
| 5 | Truth | The city is lit |

## FAQ

### How do I make a post-apocalyptic drama with AI?

Choose a SceneMixer world such as Wasteland or Frozen Wilderness, salvaged costumes, factions with one mark each, and worn states as time passes.

### How do I show time passing?

Add worn, thinner or scarred states to each character and switch to them in later segments.

### How do I show crowds of survivors?

As background people in the shot, with named characters in front.

### Can I show violence between factions?

The threat and the aftermath, not wounds; graphic violence is blocked.

### Which look suits it?

Shot on Film for warmth and grit, Digital HD for clean modern ruins, Game Engine Realism for a game feel.

### Can I have vehicles?

Yes, as props or set dressing; a convoy reads best in wide shots.

### Is this different from a zombie drama?

Yes: the threat is scarcity and people, not the infected; there is a separate zombie guide.

### Vertical or landscape?

Landscape for the scale of the ruins; vertical for a fast short-drama cut.

Updated November 3, 2026


---

# How to Make an AI Preschool Learning Cartoon (Ages 2 to 5)

URL: https://scenemixer.com/guides/make-an-ai-preschool-learning-cartoon/
Make a preschool learning cartoon with AI: one idea in 60 to 90 seconds, one shot repeated on purpose, spoken rhymes, gentle looks, and what YouTube asks of kids' videos.

# How to make an AI preschool learning cartoon

Quick answer: A preschool cartoon is shorter, slower and more repetitive than a kids' show: one idea, 60 to 90 seconds (four to six 15-second segments), the same place and the same framing every time, a friendly guide who talks to the viewer, and a short line said again and again. In SceneMixer, draw the guide and the places once, write the same shot sentence into every segment so each one opens like the last, and keep rhymes as spoken lines: the video speaks, it does not sing a tune you bring. Check the facts and the platform's kids rules yourself.

Brief for AI Create: A 75-second counting cartoon for ages 2 to 4: a small yellow duck called Pip swims across a pond and meets one more duck each time, one to five; same pond, same camera, Pip says the number and quacks; it ends with all five ducks and Pip saying ‘Five!’

## What changes at preschool

| Point | Preschool (2 to 5) | Older kids' episode |
| Length | 60 to 90 seconds | Two to five minutes |
| Ideas | One: a number, a colour, a shape, a feeling | One question with an experiment |
| Pace | Slow; a thing happens, then it repeats | A small story |
| Words | A few, said clearly, again and again | A short script |
| The viewer | Talked to directly by the guide | Watches the characters |
These are starting points, not rules. For older children: An educational cartoon for kids · A bedtime story series.

## A counting episode, segment by segment

| Segment | On screen | Said |
| 1 | Pip alone on the pond | ‘One duck. Quack!’ |
| 2 | Pip and one more duck | ‘Two ducks. Quack, quack!’ |
| 3 | Three ducks in a row | ‘Three ducks. Quack, quack, quack!’ |
| 4 | Four ducks, one splashing | ‘Four ducks!’ |
| 5 | All five ducks; Pip waves | ‘Five! Hooray!’ |
Five segments of 15 seconds is 75 seconds. Each segment on SceneMixer is a fresh generation, so what repeats has to be written into each one. The first two shots of segment 1:

[Shot 1] 6s: Medium shot, camera still. A round blue pond with lily pads, morning light. @Pip, a small yellow duck, floats in the middle. @Pip (bright): One duck. Quack! (voice: Pip)
[Shot 2] 5s: Same medium shot, camera still. @Pip looks at the viewer and waves a wing. Ambient: gentle water, a bird far off. No dialogue.

## Repeat the shot on purpose

Keep the pond, the camera words and the order of events the same in every segment, and change one thing: the number of ducks. A segment is generated on its own, so copy the framing sentence into each one; one place per segment keeps the pond the same pond. Why the next segment opens like the last one ended · Shot-to-shot continuity.

## Rhymes: spoken, not sung

The video speaks the lines written in the storyboard; it takes no audio, so it does not sing a tune you bring. Write a rhyme as short lines with the beat in the punctuation, read steadily, at about three words a second. If you want a sung version, record or license one elsewhere and lay it under the video in an editor after turning the Music switch off. Write your own rhyme, or use one you have the right to use. The built-in score and your own · How voices work.

## Looks that suit very young children

Rounded, bright, calm. 3D Toy Cartoon and Claymation are tactile and funny; Family 3D Animation suits a guide with expression; Watercolor Storybook is the softest. The look is fixed after step 1.

Visual Style: Claymation

Visual Style: Family 3D Animation

Visual Style: Watercolor Storybook
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Before you post kids' content

YouTube's own page on setting a video's audience says creators must tell YouTube whether their videos are “made for kids”, at channel or video level in YouTube Studio, and frames it as part of complying with the Children's Online Privacy Protection Act and other laws; it also says to consult legal counsel if you are unsure. Source: Set your channel or video's audience (checked November 2, 2026). That is YouTube's wording, not our judgement.

TikTok's page on AI-generated content lists as not allowed “the likeness of young people under the age of 18” (About AI-generated content, checked November 2, 2026). Keep your cast invented: do not use a real child's photo as a reference, and the video provider's face check refuses images that look like a real person. Reference image flagged as a real person's face · Run a kids' cartoon channel with AI.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Pastoral Healing; Costume Style Soft Everyday; Visual Style Watercolor Storybook, Family 3D Animation; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw Pip and the four other ducks as non-human characters, and the pond as the one place; the pond is the same in every episode. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Age | Episode length | Segments of 15 s | One idea |
| 2 to 3 | 45 to 60 seconds | 3 to 4 | A number to three, a colour, a sound |
| 3 to 4 | 60 to 90 seconds | 4 to 6 | Counting to five, a shape, a feeling |
| 4 to 5 | 90 to 120 seconds | 6 to 8 | A letter, a routine, a kindness |

## FAQ

### Can AI make a preschool cartoon?

Yes: a guide character and a place drawn once, one idea per episode of 60 to 90 seconds, and the same shot repeated on purpose.

### Can it sing a nursery rhyme?

It speaks the lines you write; it takes no audio and does not sing a tune you bring. Add a sung or licensed track in an editor.

### How long should an episode be?

60 to 90 seconds is four to six 15-second segments; these are starting points, not rules.

### Which look suits toddlers?

3D Toy Cartoon, Claymation and Watercolor Storybook are common choices; Family 3D Animation suits a more expressive guide.

### Who checks the facts?

You do. The characters say exactly what your script says.

### Do I have to mark it made for kids on YouTube?

YouTube's page says creators must set whether their videos are made for kids, and to consult legal counsel if unsure. Read it at the link above.

### Can I use a photo of my child?

Do not. Use invented characters; the video provider's face check refuses images that look like a real person.

### What does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join, so a 75-second episode is 75 seconds of each, before retakes.

Updated November 3, 2026


---

# AI Princess and Royal Drama: Palaces, Crowns and Court Intrigue

URL: https://scenemixer.com/guides/make-an-ai-princess-or-royal-drama/
Make a princess or royal court drama with AI: European Palace and court gowns, a crown that stays the same, ball scenes with background guests, rank you can see.

# How to make an AI princess or royal drama

Quick answer: A royal drama runs on rank: who may sit, who must bow, who wears the crown. In SceneMixer, Step 1 sets the court. European Palace gives gilded Baroque halls and halls of mirrors; European Court Gown dresses everyone in corsets, wide skirts and embroidered coats, scaled by each character's rank. For other courts there are Indian Royal Palace, Thai Royal Palace, Joseon Korea and Ottoman Palace. Make the crown a prop with its own picture, and give the ball's guests a background line.

Brief for AI Create: A princess drama: the king's illegitimate daughter is brought to court to marry in her half-sister's place; the foreign prince she is promised to is a spy, and so is she; 10 short episodes, each ends at a public event where one of them nearly gives the other away.

## Choose the court

| Court | Scene Style | Costume Style |
| A European kingdom, 17th–19th century | European Palace | European Court Gown |
| A warrior princess in a medieval kingdom | Medieval Castle | Medieval Nobility, Plate Armor |
| A dark, gothic court | Gothic Dark | Gothic Court |
| A northern Indian royal house | Indian Royal Palace | Indian Royal |
| A Thai court | Thai Royal Palace | Thai Court |
| A modern royal family | English Manor or Urban Opulence | Haute Couture, Power Tailoring |
Also: Joseon palace · Ottoman palace · Regency romance.

## Rank you can see

The costume style is scaled by each character's rank when the script is read, so a queen and her maid wear the same period at different levels. Help it by saying who is who in the script: “the Queen”, “her lady-in-waiting”, “a kitchen maid”. A princess who is secretly a servant is one character with two states: the plain dress and the court gown. States · Personality the video shows.

The crown, the ring and the seal that keep coming back are props with a reference picture. The same prop in every hand.

## The ball, shot by shot

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, the hall of mirrors at night, candles in every chandelier; couples in court dress turn across the floor in the background. |
| The entrance | [Shot 2] 3s: Medium shot. @Elise stops at the top of the staircase in a pale blue gown. |
| The look | [Shot 3] 3s: Close-up on @Prince Auguste lowering his glass. |
| The line | [Shot 4] 4s: Close-up on @Queen Margit. @Queen Margit (sweetly): Smile, my dear. Everyone is watching. (voice: Margit) |
Background guests · How to write one shot.

## Story shapes that work

- The swapped princess: the wrong girl marries the prince.
- The warrior princess: a throne won on a battlefield.
- The commoner bride: the court that wants her gone.
- The lost heir: the servant with the royal birthmark.
- The regent: a young queen and the uncle who rules for her.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style European Palace; Costume Style European Court Gown; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the royal family and servants with one mark each; the crown and seal as props; the hall, the chapel and the throne room as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | A country inn. | Royal guards arrive asking for her by name. |
| 2 | The palace dressing room. | She is fitted for her sister's wedding gown. |
| 3 | The ball. | The prince recognises her from somewhere else. |
| 4 | The chapel. | A letter in his coat, in a code she knows. |
| 5 | The throne room. | The king collapses in the middle of the ceremony. |

## FAQ

### How do I make a princess drama with AI?

Pick a court scene style such as European Palace and a costume style such as European Court Gown in SceneMixer's Step 1, then write or generate the script and render episode by episode.

### Will a queen and a maid be dressed differently?

Yes. The costume style is scaled by each character's rank, so say who is who in the script.

### How do I show a servant who is secretly a princess?

Give her two states in SceneMixer's Step 2: the plain dress and the court gown, and pick the right one per segment.

### How do I keep the crown the same?

Make it a prop with a reference picture and tag it in the shots where it appears.

### Can I film a full ballroom?

Yes. Put the dancers in the background line and keep two or three named people in each shot.

### Can I set it in an Asian or Indian court?

Yes: Joseon Korea, Thai Royal Palace, Indian Royal Palace, Ottoman Palace and Imperial Splendor are all scene styles on SceneMixer.

### Can I use a real royal family?

No. Real people cannot be depicted. Invent the kingdom and the family.

### Can I copy a famous animated princess?

No. Known characters, names and iconic dresses trigger the IP check. Design your own princess.

Updated November 3, 2026


---

# AI Prison Drama: Cells, Yards, Guards and a Plan to Get Out

URL: https://scenemixer.com/guides/make-an-ai-prison-drama/
Make a prison drama with AI: cells, yards and visiting rooms as recurring locations, uniforms for inmates and guards, prison as a state, and fights kept off-screen.

# How to make an AI prison drama

Quick answer: A prison drama is a closed world with its own rules: who runs the yard, which guard can be bought, how long until parole. In SceneMixer, build the prison from locations (the cell, the yard, the canteen, the visiting room) and use them again and again. Forensic Cold White gives the hard institutional look; Professional Uniforms dresses guards and inmates in one cut per role. The prison jumpsuit is a state of each character, so the same person can walk in wearing a suit.

Brief for AI Create: A prison drama: an accountant who took the fall for her boss is sent to a women's prison where the boss's wife is already serving time; she has 18 months to find the proof before her parole hearing; 12 short episodes, each ends in the visiting room.

## Build the prison

| Space | How to make it |
| The cell, the yard, the canteen, the visiting room | Locations in Step 2, each with its own picture, reused in every episode. The same room. |
| The overall look | Scene Style Forensic Cold White (white tile, stainless steel, tube lights) or Wasteland for a dystopian prison. |
| Guards and inmates | Costume Style Professional Uniforms: one cut and colour per role. Say who is a guard and who is an inmate. |
| A historical prison | Dungeon scene with Medieval Nobility or Victorian costumes. |

## The jumpsuit is a state

The character arrives in her own clothes and is processed into the prison uniform: two states of one character. The release, in the same clothes she came in, is the first state again. States.

The yard is a crowd: background inmates in the background line, two or three named characters in front. Background people. Fights: the moment before and the moment after, no injuries on screen. Fights.

## The visiting room, shot by shot

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, a visiting room of steel tables under tube lights, a guard by the door. @Claire sits in an orange jumpsuit. |
| The visitor | [Shot 2] 3s: Medium shot. @Marcus sits down opposite in a tailored suit. |
| The line | [Shot 3] 4s: Close-up on @Marcus. @Marcus (smiling): You look well. Prison suits you. (voice: Marcus) |
| The tell | [Shot 4] 3s: Close-up on @Claire's hand sliding a folded note under her palm. |

## Story shapes that work

- The innocent inmate.
- The escape: one step per episode.
- The guard and the prisoner.
- Revenge from inside.
- The outside world: a family waiting for one phone call a week.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Forensic Cold White; Costume Style Professional Uniforms; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: inmates and guards with one mark each and a prison-uniform state; the cell, yard, canteen and visiting room as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The courtroom. | The sentence, and her boss's smile. |
| 2 | Processing. | Her cellmate is the boss's wife. |
| 3 | The yard. | Someone wants her silent. |
| 4 | The visiting room. | The boss visits with a deal. |
| 5 | The library. | A ledger page hidden in a law book. |

## FAQ

### How do I make a prison drama with AI?

Make the cell, yard and visiting room locations in SceneMixer's Step 2, choose Forensic Cold White and Professional Uniforms in Step 1, and give each inmate a prison-uniform state.

### How do I show a character entering prison?

Two states: her own clothes and the prison uniform. Switch at the processing scene.

### Can I film a prison fight?

Show the moment before and after, one action per shot; injuries are blocked.

### How do I fill the yard?

Write background inmates in the background line and keep a few named characters in front.

### Which visual style suits a prison drama?

Digital HD or Shot on Film for grit.

### Can it be a dystopian prison?

Yes. Use Wasteland or Cyber Metropolis as the scene style in SceneMixer.

### Can I use a real prison's name?

Invent it; real names and logos can trigger the IP check.

### Vertical or landscape?

9:16 Portrait for short-drama feeds, 16:9 Landscape for a series.

Updated November 3, 2026


---

# AI Psychological Thriller: Doubt, Mirrors and the Unreliable Hero

URL: https://scenemixer.com/guides/make-an-ai-psychological-thriller/
Make a psychological thriller with AI: one or two rooms seen again and again, small changes the viewer notices, an unreliable narrator, and quiet cliffhangers.

# How to make an AI psychological thriller

Quick answer: A psychological thriller lives in the viewer's doubt: is it happening, or is she imagining it? In SceneMixer, keep it to one or two locations so every change is visible, use the same framing for the same room so differences stand out, give the hero a voice-over the pictures sometimes contradict, and end episodes on a detail rather than an event.

Brief for AI Create: A psychological thriller: a woman who works from home notices that the books on her shelf are rearranged every morning; her husband says she did it; eight episodes, one apartment.

## Same room, small changes

A location is drawn once and becomes the set, so it looks the same every time. That is the thriller's tool: move one thing. A picture turned to the wall, a door ajar, a second cup. Write the change into the shot and keep the framing identical to the earlier one. Your scene image becomes the set.

| Setting | Scene Style | Why |
| Cold apartment | Nordic Minimal | Empty space makes every object loud |
| Office | Corporate Cold | Glass walls, being watched |
| Big house | Suburban Mansion | Too many rooms |
| Hospital | Hospital Ward | Who is the patient? |

## The narrator who might be wrong

Voice-over lines tell us what the hero believes; the shot shows what is there. Let them disagree once an episode. A voice-over line is written as (voice-over) before the line. Dialogue, voice-over and captions.

## A detail as a cliffhanger

| Beat | Shot to adapt |
| Routine | [Shot 1] 4s: Medium shot, morning light. @Nora sets two mugs on the counter. @Nora (voice-over): Same as every day. |
| The shelf | [Shot 2] 3s: Wide shot of the bookshelf, every book spine turned to the wall. |
| The husband | [Shot 3] 3s: @Daniel in the doorway, calm. @Daniel: You did that last night. You don't remember? |
| The detail | [Shot 4] 3s: Close-up. A third mug in the sink. |

## Looks that unsettle

Black & White Film for a classic feel; Soft Glow for a dreamy, unreliable haze; Digital HD for clinical modern dread. Weather and time are set per segment: a storm can arrive exactly when the doubt does. Day, night and rain.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Nordic Minimal, Suburban Mansion; Costume Style Soft Everyday, Smart Casual; Visual Style Black & White Film, Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: two or three characters with specific faces; one apartment as the main location; the changed objects as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | What changes | The ending detail |
| 1 | Books turned around | A third mug |
| 2 | Her own handwriting on a note | The note is dated tomorrow |
| 3 | A neighbour who knows her name | She has never met him |
| 4 | The husband's story | His phone lights up with her number |
| 5 | What is real | The shelf is back to normal |

## FAQ

### How do I make a psychological thriller with AI?

Keep to one or two locations, change small details between identical shots, and let a voice-over narrator contradict what we see.

### How do I keep the room exactly the same?

The location is drawn once on SceneMixer and used as the set; repeat the same framing and change one object in the shot text.

### Can the narrator lie?

Yes. Voice-over lines and the picture are written separately, so they can disagree.

### Which style suits a thriller?

Black & White Film, Soft Glow or Digital HD, depending on the mood.

### How long should episodes be?

One to three minutes, each ending on a detail.

### Can I show violence?

Threat and aftermath only; graphic violence is blocked.

### Can I do a twist ending?

Yes. Plan it in the script; the storyboard keeps the clues in the frame.

### Is it better as a film or a series?

Both work; a series lets each episode end on a new doubt.

Updated November 3, 2026


---

# AI Racing Drama: Invented Cars, Night Roads and the Last Lap

URL: https://scenemixer.com/guides/make-an-ai-racing-drama/
Make a racing drama or street racing movie with AI: invented cars as props, the race in short tracking shots, the cockpit as its own place, the rival off the track.

# How to make an AI racing drama

Quick answer: A racing story is won off the track: the rival, the debt, the mechanic who believes. In SceneMixer, invent the cars (real makes and logos trigger the brand check) and give each one a number and a colour as a prop, tell the race in short shots (alongside, the mirror, the gap, the line), treat the cockpit as its own location, and stage the crash as the spin and the silence after it.

Brief for AI Create: A street racing drama: a delivery driver who knows every back road in the port city is pulled into the night races to win back her late father's car; twelve short episodes, one rival, one final race at dawn.

## Cars you can recognise

| Car | Written as a prop |
| Hers | Battered 1990s coupe, faded orange, a white 27 on the door |
| The rival's | Long black grand tourer, matte paint, red brake lights |
| The father's | The same coupe, freshly painted, in a flashback state |

No real makes or badges: well-known brands set off the IP check. A car described this exactly comes out the same in every shot. Describing props · Brands and the IP check.

## The race in short shots

| Beat | Shot to adapt |
| The start | [Shot 1] 3s: Low angle, two cars side by side at a red light, engines revving. |
| Alongside | [Shot 2] 2s: Tracking shot, the camera level with the orange coupe at speed. |
| The mirror | [Shot 3] 2s: Close-up, the rear-view mirror: the black car's lights growing. |
| The gap | [Shot 4] 2s: High wide shot, both cars into a narrow tunnel. |
| The face | [Shot 5] 3s: Close-up on @Jo through the windscreen, jaw set. |

Two to three seconds each, the framing changing on every cut, is what makes speed. Camera moves.

## The cockpit is a place

Make the inside of the car its own location so the seats, the dashboard and the windows stay put; say who is driving and who is beside them. Scenes inside a car.

## Night roads and racetracks

Neon Nights for street racing, Street Life for the garage and the port, Mediterranean Coast for a coastal road, Frontier Vastness for desert rally. Racing suits are Sportswear; the crew is Streetwear.

Scene Style: Street Life

Scene Style: Mediterranean Coast

Scene Style: Frontier Vastness

Costume Style: Sportswear

Costume Style: Streetwear
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## The crash

Show the spin, the barrier and the stillness; cut to the helmet in the road and the rival's brake lights stopping. Injuries are graphic violence and get blocked; the silence is more frightening anyway. Content rules.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Street Life; Costume Style Sportswear, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the driver, the rival and the mechanic; each car as a prop with its number and colour; the garage, the road and the cockpit as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Beat |
| 1 | Her father's car, sold to the rival |
| 3 | Her first night race |
| 6 | The mechanic who knew her father |
| 9 | The crash, and who stopped |
| 12 | Dawn race for the car |

## FAQ

### How do I make a racing drama with AI?

Invent the cars as props, tell races in two- or three-second shots, make the cockpit a location and keep the story off the track.

### Can I use real car brands?

No. Real makes and logos trigger SceneMixer's brand check; describe your own car exactly.

### How do I make the car look the same every time?

Describe it as a prop with year, shape, colour and number; it gets a reference image.

### How do I show speed?

Short tracking and low shots, the mirror, the gap, the face; change the framing on every cut.

### Can there be a crash?

Show the spin and the aftermath; injuries are blocked on SceneMixer.

### Which look suits it?

Digital HD for modern street racing, Shot on Film for a seventies track story.

### Can it be motorbikes or horse racing?

Yes; the same rules apply to bikes and horses.

Updated November 3, 2026


---

# AI Rebirth Revenge Drama: Second Life, Same Enemies, New Plan

URL: https://scenemixer.com/guides/make-an-ai-rebirth-revenge-drama/
Make a rebirth drama with AI: a heroine reborn into her past, her two lives as states, the betrayal shown once, and a revenge plan revealed episode by episode.

# How to make an AI rebirth or second-life revenge drama

Quick answer: A rebirth drama starts at the end: the heroine is betrayed and dies, then wakes up years earlier with everything she knows. The two lives are the same character with two states (the broken woman of the first life and the young woman who gets a second chance), so the face stays the same and only the look changes. In SceneMixer, set her world in Step 1 (Urban Opulence and Minimal Luxury for a modern rich family, Imperial Splendor and Qing Palace Robes for a palace) and reveal the plan one move per episode.

Brief for AI Create: A rebirth revenge drama: a woman pushed from a balcony by her husband and her best friend wakes up on the morning of her engagement, five years earlier; this time she says yes, and starts collecting proof; 12 short episodes, each ends with one of them making the same mistake as before.

## Two lives, one face

- The first life is a state. Older, worn, the wedding ring she will not wear this time. States.
- Flashbacks of the first life use that state; the present uses the other. Say “in her first life” in the script so it is easy to see which is which.
- The same places, twice. The balcony, the wedding hall, the office: one location each, with a second look if they change. The same room.
- An inner voice tells the audience what she remembers. Voice-over.

## Set her world

| World | Scene Style | Costume Style |
| A modern rich family | Urban Opulence | Minimal Luxury, Haute Couture |
| A company | Corporate Cold | Power Tailoring |
| An ancient palace | Imperial Splendor | Qing Palace Robes, Imperial Court Regalia |
| A European court | European Palace | European Court Gown |

## The wake-up, shot by shot

| Beat | Shot to adapt |
| The fall | [Shot 1] 3s: Medium shot, night, a balcony railing. @Vivian's hand slips from it; the shot cuts before she falls. |
| The morning | [Shot 2] 4s: Close-up. @Vivian gasps awake in a sunlit bedroom, younger, no ring. |
| The proof | [Shot 3] 3s: Close-up on her phone: the date, five years earlier. |
| The inner voice | [Shot 4] 4s: Close-up on @Vivian's face hardening. @Vivian (inner voice): This time, I'll let you think you've won. (voice: Vivian) |
The death is never shown, only the moment before. Content rules.

## Story shapes that work

- The betrayed wife.
- The reborn heiress who takes back the company.
- The palace concubine who knows who poisons whom.
- The second-chance romance: this time she marries the man she ignored.
- The reborn villainess who refuses her fate.
Revenge dramas · ReelShort-style dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Corporate Cold; Costume Style Minimal Luxury, Power Tailoring; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the heroine with a first-life and a second-life state; the husband and the friend with one mark each; the balcony, the bedroom and the wedding hall as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The balcony, then her bedroom. | Five years earlier, the morning of the proposal. |
| 2 | The engagement dinner. | She says yes, smiling. |
| 3 | The company. | She buys the shares he will need. |
| 4 | Her best friend's flat. | The same letter she found in her first life. |
| 5 | The wedding hall. | She invites the man he ruined. |

## FAQ

### How do I make a rebirth drama with AI?

Give the heroine two states in SceneMixer's Step 2 (first life and second life), set her world in Step 1, and reveal her revenge one move per episode.

### How do I show the two lives?

The same character with two states. Pick the first-life state in flashbacks.

### Can I show her death?

Show the moment before and cut. Graphic injury and death scenes are blocked.

### How does the audience hear what she remembers?

Use her inner voice in the shot.

### Can it be a palace rebirth story?

Yes. Choose Imperial Splendor or European Palace and the matching costume style in SceneMixer.

### Which visual style suits a rebirth drama?

Digital HD or Soft Glow for modern stories, Shot on Film for period ones.

### Vertical or landscape?

9:16 Portrait; rebirth dramas live on short-drama feeds.

### How long should each episode be?

One to two minutes with one move and one ending beat.

Updated November 3, 2026


---

# AI Regency Romance Drama: Ballrooms, Letters and Country Houses

URL: https://scenemixer.com/guides/make-an-ai-regency-romance-drama/
A Regency romance with AI: the English Manor scene style, Regency gowns and tailcoats, English dialogue, and how to stage a ballroom without a hundred dancers.

# How to make an AI Regency romance drama

Quick answer: A Regency romance is made of three things: a house, a dress and a conversation that says less than it means. In Step 1 pick the scene style English Manor — country houses and London townhouses with pale stone fronts, tall windows and white plaster mouldings — and the costume style Regency: high-waisted gowns in fine cotton and silk, fitted tailcoats and high cravats. Set Target Region to Western and the dialogue language to English. Then plan the ballroom the way a small production would: two dancers in a medium shot, hands and hems in close-up, and the crowd as a murmur behind them.

Brief for AI Create: A Regency romance: a vicar’s daughter is sent to a great house for the summer to help catalogue a library, and finds that the owner’s brother has been writing to her under a false name; 8 short episodes, English dialogue, every episode ends on a letter.

## The choices that make it Regency

| Step 1 setting | Choose | Why |
| Scene Style | English Manor | Regency to Edwardian high society: stone-fronted country houses and London townhouses, tall windows, white plaster mouldings. |
| Costume Style | Regency | England from 1811 to 1820: high-waisted straight gowns in fine cotton and silk; fitted tailcoats with high cravats; light and clean. |
| Visual Style | Soft Glow, Digital HD or Shot on Film | A soft glow suits the genre; Classical Oil Painting if you want a portrait-gallery look. |
| Target Region | Western or Follow script | Western gives Western names and faces; Follow script keeps what the script says. |
| Dialogue language | English | Spoken as each clip renders; write the lines in a slightly formal register. |
These are written into the descriptions when the script is read, so choose them first. Why they are fixed afterwards · Every scene and costume style.

## Not the same as the other period looks

- Regency is 1811–1820. For later in the century use Victorian (corseted long dresses, three-piece suits and top hats), and European Palace and European Court Gown for the continental courts.
- English Manor is the house. Parisian Streets and European Palace are different places entirely.
- For a wider look at costume dramas, read Period romance with AI.

## The ballroom, without a hundred dancers

The hardest thing to ask for is a room full of couples dancing in step. The genre does not need it:

- Two dancers, medium shot. The only couple in focus is the one the story is about.
- Close-ups of hands, gloves, hems and eyes carry the dance. Give each its own shot of two or three seconds.
- The crowd is a murmur. Write “couples dancing in the background, out of focus” and let the dialogue do the work. Background people in a scene.
- Cut away to the chaperones along the wall, the fan, the clock: the room watching is a character too.

## The letter is a prop

Regency stories live on letters: written, folded, sealed, intercepted. Make the letter a prop in Step 2 with a reference image, and name it in the shots with @ so that it is the same letter, in the same hand, every time. The same prop in every hand.

Forms of address pile up (Miss Hartwell, Lady Cathcart, my lord). Use them freely in the script, but tie each to its person the first time it appears (“Lady Cathcart, whom the whole county calls the Dowager”). When the script is read, the card keeps the other names your text uses for her on its grey Also called line (read-only; there is nothing to type), so a later “the Dowager” lands on the same card. Also called.

## A ballroom scene, shot by shot

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, fixed camera. A country-house ballroom lit by candles, tall windows, white plaster ceiling, couples forming two lines in the background, out of focus. @Eliza stands near a pillar with her fan closed. |
| The approach | [Shot 2] 3s: Medium shot. @Julian crosses the floor toward @Eliza and stops a polite distance away. |
| The line | [Shot 3] 4s: Close-up on @Julian, bowing slightly. @Julian (formal, a little too quickly): Miss Hartwell, I believe the next dance is mine. (voice: Julian) |
| The answer | [Shot 4] 3s: Close-up on @Eliza, lifting her chin. @Eliza (cool): You believe a great deal, Mr. Ashworth. (voice: Eliza) |
| The hand | [Shot 5] 3s: Close-up on their gloved hands as she gives him hers, not looking at him. |
How to write one shot · Hooks and cliffhangers.

## Story shapes that work

- The season in town: a debutante, a chaperone, a list of eligible men and one who is not on it.
- The house party: a closed world for a week, a cast of rivals under one roof.
- The false name: letters from someone who is closer than they say.
- The refused proposal: the worst possible first proposal, and the long road after it.
- The inheritance: a house that must pass to someone, and the marriage that would keep it.
Keep the romance on the Regency scale: a look held a beat too long, a hand taken to cross a floor, a letter kept. The kiss, if there is one, belongs to the last episode.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief in AI Create.
- Step 1: Type Short Drama or TV Series; scene style English Manor; costume style Regency; a visual style; Target Region; English dialogue; 9:16 Portrait or 16:9. Read the script.
- Step 2: give each character a specific face and an age; add the letter as a prop. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project). One place per segment: the library, the ballroom, the garden walk.
- Video: generate on the default tier at 10 credits a second; put the episodes together and post.

| Episode | The scene | The ending beat |
| 1 | The library of the great house. | The first letter, signed with a name that is not his. |
| 2 | A walk in the garden. | The brother pretends not to know her. |
| 3 | The ball. | “I believe the next dance is mine.” |
| 4 | The rain, and a shared carriage. | He asks who she has been writing to. |
| 5 | The reading of a will. | Her name is in it. |

## FAQ

### How do I make a Regency romance with AI?

In SceneMixer's Step 1, choose the scene style English Manor and the costume style Regency, set Target Region to Western and the dialogue language to English, then write or generate the script and render episode by episode.

### What does the Regency costume style include?

England from 1811 to 1820: high-waisted straight gowns in fine cotton and silk for women, fitted tailcoats with high cravats for men.

### Can I film a ballroom full of dancing couples?

Keep the dance to two people in medium shots, add close-ups of hands and hems, and write the crowd as couples in the background. A whole room dancing in step is the hardest thing to ask for.

### How do I keep the same letter in every scene?

Make it a prop in SceneMixer's Step 2 with a reference image and name it with @ in each shot.

### Is Victorian the same style?

No. Regency is 1811 to 1820. Victorian is the later corseted look with three-piece suits and top hats.

### Can I include a kiss?

Keep it chaste and late. Anything sexual is refused by the content rules.

### How is this different from the period romance guide?

That guide covers costume romances in general. This one is specific to Regency England: the manor, the gowns, the ballroom and the letters.

### Will the leads look the same in every episode?

Yes. Each character is drawn once and that sheet is used in every shot.

Updated November 3, 2026


---

# AI Revenge Drama: Betrayal, Rebirth and the Payback, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-revenge-drama/
Betrayed, reborn, back for payback: how to structure a revenge short drama, handle a past-life flashback look, and generate each takedown in SceneMixer.

# How to Make an AI Revenge Drama

Quick answer: A revenge drama is a list: the people who wronged the lead, crossed off one at a time, each takedown bigger than the last. Episode 1 delivers the betrayal so completely that the audience wants every payoff that follows; from then on, every episode moves one name closer to the bottom of the list. In SceneMixer, write that list into your script or your AI Create brief, give the lead a second look as a state if the story has a before-and-after (the woman she was, the woman who came back), pick that look shot by shot rather than for whole episodes, and give the evidence that brings each villain down its own prop image so the same envelope, recording or ring appears every time.

Brief for AI Create: A revenge drama: a wife pushed off a yacht by her husband and her best friend wakes up one year earlier, on the morning of her engagement; she lets them think nothing has changed while she takes the company back piece by piece; she exposes both at the anniversary party; cold, precise, satisfying.

## How a revenge series is built

- The betrayal is the whole of episode 1. Do not explain it later. Show it: who did it, what they took, the moment she understood.
- The return. Either a second chance — she wakes up before it happened, with everything she knows — or a long absence and a new identity. Both give the lead an edge the villains do not know about.
- The list. Three to five targets, from the smallest to the one who planned it. Each target is an arc of a few episodes.
- The double game. The lead acts as if nothing has changed. Scenes where the audience knows more than the villains are the genre’s best tension.
- The public exposure. Each arc ends in front of witnesses: a party, a board meeting, a livestream, a wedding. Private revenge is not satisfying; humiliation needs an audience.
- No forgiveness until the end. If the lead softens, the series loses its spine. Let her soften once, late, and pay for it.
Rotate the episode endings: a villain getting closer to the truth, a new piece of evidence, an ally turning, a target falling. Hooks and cliffhangers.

## Before and after: the lead’s two looks

Rebirth and return stories usually show the lead in two versions: the trusting woman of the first life and the composed one who came back, or the plain student and the polished executive she becomes. If the difference is in her clothes, hair or face, it changes what the reference image shows, so it needs its own image:

- Generate her main image as the version who appears in most of the series.
- On her card choose Add a state, name it for the other version (First life, Before the fall), and describe only the difference: long loose hair, pale cardigan, no makeup, a hesitant posture.
- Press Generate this state, then, in the episode editor, select a segment set in the other life, open her from the episode’s assets and choose Use this state in this shot.
Pick flashback looks segment by segment, not for whole episodes. An episode that cuts between the past and the present would otherwise put her in the flashback clothes in every shot. If you never generate the state image, the shots describe the difference in words instead — weaker, but nothing breaks. Use a character state in one scene · Costume changes and states.

Character sheets from the sample The Veil at Dawn: one reference image per character, used in every shot they appear in.

## Settings to pick in step 1

- Type: Short Drama. Revenge depends on payoffs at a steady rate, which is exactly what the short-drama rhythm delivers.
- Aspect Ratio: 9:16 Portrait for feeds; faces and reactions are the payoff, and a vertical frame keeps them large.
- Visual Style: Digital HD or Shot on Film; Black & White Film is worth considering only if the whole series is a period piece.
- Scene Style: Urban Opulence or Corporate Cold for corporate revenge, Suburban Mansion for a family feud, Island Resort if the betrayal happens on a yacht or a holiday.
- Costume Style: Power Tailoring or Minimal Luxury; the lead’s transformation is better told through her state image than through a project-wide style.

Visual Style: Shot on Film

Scene Style: Urban Opulence

Scene Style: Corporate Cold

Scene Style: Island Resort

Costume Style: Minimal Luxury
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Evidence that looks the same every time

Revenge plots run on objects: the recording, the second phone, the forged contract, the necklace that was never lost. The audience has to recognise the object the moment it reappears, so if your story keeps coming back to one, check that it has a card in step 2 with its own image, and name it with @ in every shot it is in. Two rules make evidence scenes work on screen:

- Show the object, say the content. A generated phone screen or document will not hold readable text. Let the line say what it proves; let the shot show who is holding it and who sees it.
- Cut to the guilty face. The reveal is not the object, it is the reaction. Give the villain a close-up with nothing else in frame.
Same prop in every hand · Text in the picture.

## The exposure scene, written as shots

| Beat | Shot to adapt |
| The villain is comfortable | [Shot 1] 4s: Opening: medium shot. @Derek raises a glass at the head of the long table, @Claire at his side, smiling. @Derek (warm, loud): To the woman who made all this possible. Ambient: the murmur of a full room. |
| The lead moves | [Shot 2] 3s: Cut to close-up. In frame only: @Elena. She sets her glass down and stands. No dialogue. |
| The evidence | [Shot 3] 4s: Medium shot. @Elena holds up the @Second phone so the guests nearest her can see it. @Elena (calm): Then let’s play the part you left out. |
| The fall | [Shot 4] 3s: Close-up. In frame only: @Claire. The colour leaves her face; she looks at @Derek and he will not look back. No dialogue. Ambient: the room goes quiet. |
The guests are the segment’s background line, not characters. When the same two people share consecutive shots, change the framing between them so the join reads as a cut. How to write one shot · Background people.

## Keep the revenge on the right side of the content check

Revenge plots flirt with violence: a push off a yacht, a car accident, a poisoning. The video is checked on every generation, and extreme violence and gore are blocked. Show the moment before and the consequence after: the hand on her back, then the water closing over; the brake pedal going to the floor, then the hospital corridor. Keep real people, real company names and real brands out entirely. Blocked by moderation: what to change.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script or write a brief in AI Create that names the betrayal, the return, the list and the ending. Writing the one-line brief.
- Step 1: the settings above; read the script; check in the summary that each episode moves one target closer to their fall.
- Step 2: generate the cast and the places the story returns to; add the lead’s second look as a state; check the evidence props have images.
- Storyboard and video: episode 1 first. Its storyboard is free on your first project; video on the default tier is 10 credits a second.
- Fix, composite, post. Edits to storyboard text are free and every version is kept in the text history, so you can try a sharper line and go back if it reads worse. Storyboard text history.
For a second-chance story built this way, watch The Veil at Dawn: a woman wakes with memories of a past life and walks out on a toxic wedding.

| Arc | Target | Episodes | The takedown |
| 1 | The friend who helped | 2–5 | Her lies are read out at the engagement dinner |
| 2 | The family that looked away | 6–10 | They lose the share vote they took for granted |
| 3 | The husband’s partner | 11–15 | The deal he built on stolen work collapses on a live call |
| 4 | The husband | 16–20 | Exposed at the anniversary party, in front of everyone he lied to |

## FAQ

## When the revenge is a hidden identity

A name kept secret and a second look that is revealed in front of the people who wronged her: Make an AI heiress drama.

### How do I show a flashback to the lead's first life?

Give her the first-life look as a state of her character, then pick that state only in the segments set in the past. Picking it for the whole episode would put the flashback look in every shot.

### Can AI Create write a rebirth revenge story?

Yes. Name the betrayal, how she comes back, who she wants to take down, the ending and the tone in one or two sentences, then pick a length.

### How many episodes does a revenge drama need?

One arc per target is a good unit: a few episodes each. Three or four targets make a full season. Publish the first arc before writing the rest.

### Can the villain be a real company or a real person?

No. Real people are not allowed, brand names and logos trigger the IP check, and real companies do not belong in a story about crimes. Invent them.

### How do I show a crime without the video being blocked?

Show the moment before and the result after, not the act itself. A hand on her back and then the water is clearer on screen than the fall.

### Can the evidence be a readable document or text message?

Keep the reading in the dialogue. Generated screens and paper do not hold reliable text; show who holds it and how the guilty party reacts.

### Can I switch to vertical after I have started?

No. The aspect ratio on SceneMixer is fixed once you continue past step 1. Start a new project with the same script for the other format.

### Does a revenge drama cost more to make than a romance?

No. Every genre on SceneMixer is priced the same way: the read, the setup images, the storyboard and the video by the second.

Updated November 3, 2026


---

# AI Robot and Android Drama: Machines Who Feel and Cities of Glass

URL: https://scenemixer.com/guides/make-an-ai-robot-or-android-drama/
Make a robot or android series with AI: robots as characters with their own sheets, Cyber Metropolis and Starship worlds, human and machine states, a voice for each.

# How to make an AI robot or android drama

Quick answer: A robot drama asks one question: is it alive? Make each robot a character with its own card in Step 2 and describe what it is made of: “matte white shell, a seam down the left cheek, blue light behind the eyes”. An android that passes for human and is later revealed is one character with two states. Set the world in Step 1 with Cyber Metropolis or Nordic Minimal for a near future, Starship for space, and Near-Future Techwear for the people.

Brief for AI Create: A robot drama: a widowed engineer builds a household android with his late wife's voice; the android starts remembering things the wife never told him; 10 short episodes, each ends with the android doing something it was not programmed to do.

## Choose the future

| Kind of future | Scene Style | Costume Style |
| A near-future city | Cyber Metropolis | Near-Future Techwear, Cyberpunk Augment |
| A quiet home with a robot in it | Nordic Minimal | Soft Everyday, Loungewear |
| A ship or a station | Starship | Starship Uniform |
| A colony on Mars | Mars Base | Space Explorer |
| After the machines won | Wasteland | Wasteland Scavenge |
Sci-fi short dramas · Space operas.

## Machines as characters

- Describe the build. Shell, joints, face or no face, where the light comes from. Non-human characters.
- Human-passing and revealed are two states. The android with skin and the android with the panel open. States.
- Give it a voice. Match the robot to a preset voice so it sounds the same in every episode. Pick character voices.
- Background robots (delivery drones, factory arms) are set dressing in the shot, not characters.

## The awakening, shot by shot

| Beat | Shot to adapt |
| The kitchen | [Shot 1] 4s: Wide shot, a white kitchen at night, one pendant lamp. @Ada stands at the counter, still, blue light behind her eyes. |
| The detail | [Shot 2] 3s: Close-up on @Ada's hand arranging three cups the way the wife always did. |
| The look | [Shot 3] 3s: Medium shot. @Tom stops in the doorway. |
| The line | [Shot 4] 4s: Close-up on @Ada. @Ada (gentle, slightly flat): You never told me why we stopped going to the sea. (voice: Ada) |
How to write one shot.

## Story shapes that work

- The android who remembers.
- Robot romance: a human and a machine, and the company that owns it.
- The robot detective: a machine partner on a human case.
- The uprising: one robot chooses a side.
- The robot kid: a small, funny machine in a big family (works well in a 3D animation style).

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Cyber Metropolis, Nordic Minimal; Costume Style Near-Future Techwear; Visual Style Digital HD, Game Engine Realism; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: each robot as a character with a build description, a preset voice and a revealed state; the people with one mark each; the lab and the home as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The lab. | The android speaks with his late wife's voice. |
| 2 | The kitchen. | It makes her coffee exactly the way she did. |
| 3 | The company office. | A recall notice for this model. |
| 4 | The beach. | It remembers a day he never told it about. |
| 5 | The lab at night. | Someone has been sending it files. |

## FAQ

### How do I make a robot drama with AI?

Make each robot a character with a sheet in SceneMixer's Step 2, pick a future scene style such as Cyber Metropolis or Starship in Step 1, and give the robot a preset voice.

### How do I keep a robot's design the same?

Describe its build once on its sheet, with one recognisable mark, and that picture is used in every shot.

### How do I reveal that a character is an android?

Give it two states: human-looking, and with the panel or the circuitry showing.

### Can a robot talk?

Yes. It has a voice like any other character; choose a preset so it stays the same.

### Can I use a famous robot design?

No. Known characters and designs trigger SceneMixer's IP check. Design your own machine.

### Which visual style suits robots?

Digital HD or Game Engine Realism for live action, Realistic 3D or 3D Toy Cartoon for an animated robot.

### Can it be a comedy for kids?

Yes. A small robot in a family, drawn in Family 3D Animation or 3D Toy Cartoon, works well.

### Vertical or landscape?

Either. 9:16 Portrait for feeds, 16:9 Landscape for a series or a film.

Updated November 3, 2026


---

# How to Make an AI Sales Roleplay Video: Cast, Script, Objections

URL: https://scenemixer.com/guides/make-an-ai-sales-roleplay-video/
Make an AI sales roleplay video: a buyer and a rep, one objection per episode, a script that fits its seconds, other-language versions. Not an interactive trainer.

# How to make an AI sales roleplay video

Quick answer: People who search this want one of two things. One is an interactive practice partner, an AI buyer you talk to; SceneMixer does not do that, because a video plays the same every time. The other is a video of a sales conversation to show a team, and that is what it makes: a buyer and a rep drawn once, one objection per episode, the weak reply and the strong one in the same room with the same people. Write each scene so the words fit their seconds, and make the same episode again in another language for another team.

Brief for AI Create: Four two-minute training scenes for a team selling route-planning software to delivery fleets: the buyer says the price is double, the timing is wrong, a competitor is cheaper, and ‘just send me something’; each scene shows a weak reply, then a strong one; one buyer, one rep, one glass meeting room.

## Two things people mean, and the one this is

| Search intent | What it needs | SceneMixer |
| An AI buyer to practise against | Live conversation, branching, scoring | No: a video is the same every time you press play |
| A video of a sales conversation to show a team | A scene, two people, a script | Yes |
| A set of scenes for a course | The same cast across many scenes | Yes: a TV Series project with one buyer and one rep |
Use the video for what video is good at: showing, once and identically, what a good reply sounds like, so a team can discuss it. AI corporate training videos as short dramas.

## Start with the objection table

Write the table before the script; each row becomes one scene.

| Objection | The buyer says | Weak reply | Strong reply |
| Price | Your price is double what we pay now. | Offers a discount at once | Asks what a missed delivery costs them today |
| Timing | Come back next quarter. | ‘Sure, I'll email you.’ | Asks what has to be true by then, and who decides |
| Competitor | The other tool is cheaper. | Criticises the competitor | Asks which part of the other tool they use most |
| Brush-off | Just send me something. | Sends a brochure | Asks one question that earns a ten-minute call |
Keep the product and company invented. Brands and logos are not drawn reliably, and well-known brands trigger the IP check; describe the space instead. The IP check.

## One scene, shot by shot

The price objection, in the storyboard's own format. One 15-second segment, four shots:

[Shot 1] 3s: Medium shot, a glass meeting room. @Dana, a fleet manager, arms folded. @Dana (flat): Your price is double what we pay now. (voice: Dana)
[Shot 2] 4s: Over-the-shoulder on @Ravi, the rep, leaning in. @Ravi (warm, even): It is higher. What does a missed delivery cost you today? (voice: Ravi)
[Shot 3] 3s: Close-up on @Dana. @Dana (pausing): Nothing. We just apologise. (voice: Dana)
[Shot 4] 5s: Medium shot, @Ravi opens a notebook. @Ravi (calm): Then let's price the apology. (voice: Ravi)

Each line is short enough for its shot. Write the same scene twice, once with the weak reply and once with the strong, and they share the room and the people, so only the behaviour changes. Writing dialogue in the storyboard.

## A word budget per scene

English runs about three words a second, and a segment's seconds are all the speech has to fit in, with nobody doing anything else.

| Scene length | Seconds | Spoken words, at most |
| One segment | 15 | about 45 |
| A long take | 30 | about 90 |
| A one-minute scene | 60 | about 180 |
| A two-minute scene | 120 | about 360 |
Real calls ramble. Cut every line that is not the point. Shot seconds and timing.

## Same two people, many scenes

Make the four objections episodes of one TV Series project (up to 300 seconds each; two minutes is plenty). The buyer and the rep are drawn once and the glass meeting room is drawn once, so Dana looks like Dana in every episode. Why a series stays consistent · The same room in every episode.

## Versions for other teams, and delivery

- Languages. Make the episode again with another dialogue language and region. A second language version · Fifteen dialogue languages.
- Subtitles. Burned in, or an SRT file for a learning platform, on accounts that have made a purchase. Subtitles.
- Review. Colleagues can review through a share link, which opens without an account and always carries the AI badge. Using SceneMixer as a team.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Urban Opulence; Costume Style Professional Uniforms, Smart Casual; Visual Style Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw Dana, Ravi and the glass meeting room; they are the whole cast and the whole set of every episode. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Objection | The strong reply's one move |
| 1 | Price | Ask what the problem costs them now |
| 2 | Timing | Ask what has to be true by then |
| 3 | Competitor | Ask what they use most |
| 4 | Brush-off | One question that earns the call |

## FAQ

### Can AI make a sales roleplay video?

Yes, as a scripted scene: a buyer and a rep, one objection per episode, weak and strong replies in the same room.

### Is it interactive?

No. A video plays the same every time. It shows a conversation; it does not take part in one.

### Can I use our real product and logo?

Describe the space instead. Logos are not drawn reliably and well-known brands trigger the IP check.

### Can I make it in other languages?

Yes. Make the episode again with another dialogue language and region.

### How long should a scene be?

Two minutes is plenty; at about three words a second a two-minute scene holds about 360 spoken words.

### Can I add subtitles for a course?

Yes: burned in or as an SRT file, on SceneMixer accounts that have made a purchase.

### Can I use recordings of real calls?

Upload only material you have the right to use, with the consent of every real person whose voice appears. A recording of a real person is that person's. How voices work.

### What does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; a two-minute scene is 120 seconds of each, before retakes.

Updated November 3, 2026


---

# AI Samurai or Ninja Drama: Duels, Honour and Japanese Dialogue

URL: https://scenemixer.com/guides/make-an-ai-samurai-or-ninja-drama/
Make a samurai or ninja drama with AI: Japanese Machiya sets, Samurai Armor and Kimono, Japanese dialogue, duels in three beats, and ninja hidden in plain sight.

# How to make an AI samurai or ninja drama

Quick answer: A samurai drama turns on honour and a single stroke; a ninja story on being where no one looks. In SceneMixer, choose Japanese Machiya sets with Samurai Armor or Kimono, set the dialogue to Japanese, write duels in three beats (the stillness, the draw, the aftermath) and keep the cut itself off screen. The sample 浪人の誓い (The Ronin's Oath) was made this way.

Brief for AI Create: A samurai drama: a masterless samurai is hired to guard a merchant's daughter on the road to Edo; the assassin following them is her own brother; six episodes, Japanese dialogue.

## Edo-era worlds

| Place | Scene Style | Costume Style |
| Town and temple | Japanese Machiya | Kimono |
| The lord's castle | Japanese Machiya interiors | Samurai Armor |
| The mountain road | Frozen Wilderness in winter | Kimono with travelling cloaks |
| Ninja at night | Japanese Machiya rooftops | Dark close-fitting clothes written into the sheet |
Looks: Shot on Film, Black & White Film for classic samurai cinema, Ink Wash Animation or Anime for animated.

## The duel in three beats

| Beat | Shot to adapt |
| Stillness | [Shot 1] 4s: Wide shot, a snowy road between pines. @Kenji and @Masa face each other, hands on hilts. Ambient: wind. |
| The eyes | [Shot 2] 2s: Extreme close-up on @Kenji's eyes. |
| The draw | [Shot 3] 2s: Medium shot. Both men move at once; a flash of steel. |
| Aftermath | [Shot 4] 4s: Wide shot. They stand back to back. @Masa's sword falls into the snow. |
The stroke is never shown landing; the aftermath tells it. Writing fights.

## Ninja: disguise as a state

A ninja is two people: the servant, the monk, the merchant's daughter, and the figure on the roof. Draw the disguise as the main sheet and the night clothes as a state, with the face kept the same so the audience can catch it. States.

## Japanese dialogue

Set the dialogue language to Japanese and the region to East Asia in Step 1; lines are performed in Japanese, and you can add subtitles in other languages. Dialogue language · Subtitles.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Japanese Machiya, Frozen Wilderness; Costume Style Samurai Armor, Kimono; Visual Style Shot on Film, Black & White Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the ronin, the daughter and the assassin with specific faces; a night-clothes state for the ninja; the road and the inn as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post. Choose Japanese dialogue and the East Asia region.

| Episode | Road | The ending beat |
| 1 | The merchant's house | The contract is signed |
| 2 | The first pass | An arrow in the palanquin |
| 3 | The inn | The maid is not a maid |
| 4 | The river crossing | The assassin's face |
| 5 | The snow road | The duel |

## FAQ

### How do I make a samurai drama with AI?

Use Japanese Machiya and Samurai Armor or Kimono, Japanese dialogue, and duels written as stillness, draw and aftermath.

### Can the characters speak Japanese?

Yes. Set Japanese as the dialogue language in SceneMixer's Step 1.

### Can I show sword fights?

Yes, without wounds: the draw and the aftermath carry the moment.

### How do I make a ninja disguise work?

Draw the disguise as the main sheet and the ninja clothes as a state of the same face.

### Which look suits samurai stories?

Shot on Film or Black & White Film; Ink Wash Animation or Anime for animated.

### Can I use a famous samurai film or anime?

No. Known characters and designs trigger the IP check.

### Can I add English subtitles?

Yes, burned in or as SRT on accounts that have made a purchase.

### Vertical or landscape?

Landscape suits duels and roads.

Updated November 3, 2026


---

# AI School Bully Revenge Drama: Uniforms, Hallways and the Comeback

URL: https://scenemixer.com/guides/make-an-ai-school-bully-revenge-drama/
Make a school bully revenge drama with AI: high school sets and uniforms from four countries, bullying shown through the crowd, and a comeback that exposes.

# How to make an AI school bully revenge drama

Quick answer: A bully revenge drama is a story about being seen: humiliated in front of everyone, then coming back in front of everyone. In SceneMixer, choose the school (American High School, Korean High School, Japanese High School, Elite Boarding School) and its uniform, show the bullying through the crowd and the moment after rather than the blows, and make the revenge an exposure: the truth, a performance, a return no one expected. Students are minors, so violence stays before and after and nothing is sexualised.

Brief for AI Create: A school revenge drama: the scholarship student everyone laughed at transfers back a year later as the new head of the student council, and the first thing she schedules is the assembly where it happened; ten short episodes.

## Choose the school

| School | Scene Style | Costume Style |
| American high school | American High School | Prep School Uniform or Streetwear |
| Korean high school | Korean High School | Korean School Uniform |
| Japanese high school | Japanese High School | Japanese School Uniform |
| Rich kids' academy | Elite Boarding School | Prep School Uniform |

Scene Style: Korean High School

Scene Style: Japanese High School

Scene Style: Elite Boarding School

Costume Style: Korean School Uniform

Costume Style: Prep School Uniform
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

Set the Target Region and dialogue language to match the school. High school drama.

## Show the cruelty through the crowd

The hardest part to watch is not the shove, it is everyone filming it. Write the crowd: students in the hallway holding up phones, two of them laughing; then cut to the aftermath: books on the floor, a torn sketchbook, the lead picking them up alone. That is stronger drama, and it keeps shots from being blocked. Background people · Fights and crowds.

## The comeback as states

| Character | Before | After |
| The lead | Oversized uniform, hair over the face, head down | Fitted uniform, hair tied back, a council badge |
| The bully | Centre of every group, a phone always in hand | The same, until the assembly |

One face, two states: the audience recognises her at once and enjoys that the others do not. States.

## The exposure, shot by shot

| Beat | Shot to adapt |
| The hall | [Shot 1] 4s: Wide shot, the school assembly hall, every seat full. |
| The lead | [Shot 2] 3s: Medium shot. @Seo-yeon at the podium. @Seo-yeon: Last year, someone filmed this room. I kept a copy. |
| The bully | [Shot 3] 2s: Close-up on @Ji-ho in the third row; the smile goes. |
| The room turns | [Shot 4] 3s: Wide shot from the stage: rows of students turning in their seats toward @Ji-ho. |

On-screen text and video on a screen do not render reliably, so play the evidence on the reactions. Text on screen.

## The rules with teenage characters

Students are minors: keep violence to before and after, and nothing sexual, ever. Revenge works best as justice, not cruelty in return. Content rules. For revenge that starts with a second life, see rebirth revenge; for adults, revenge drama.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Korean High School, American High School; Costume Style Korean School Uniform, Prep School Uniform; Visual Style Digital HD, Korean Webcomic; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the lead with a before and an after state; the bully and two friends; the hallway, the classroom and the hall as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Beat |
| 1 | The day it happened, filmed |
| 2 | One year later: she walks back in |
| 4 | The old friends who stayed silent |
| 7 | The bully's own secret |
| 10 | The assembly |

## FAQ

### How do I make a school bully revenge drama with AI?

Pick the school and uniform in SceneMixer's Step 1, show the bullying through the crowd and the aftermath, and make the revenge an exposure.

### Which school settings are there?

The school settings in SceneMixer are American High School, Korean High School, Japanese High School, Elite Boarding School and Campus Youth.

### Can I show the bullying?

Show the crowd, the phones and the aftermath; direct blows may be blocked and are rarely needed.

### How do I show the comeback?

Give the lead a before and an after state; the face stays, everything else changes.

### Can I show the video evidence on a screen?

On-screen text and screens are unreliable; play it on the audience's reactions.

### Which format suits it?

Short Drama, 9:16 Portrait.

### Which look suits it?

Digital HD, or Korean Webcomic for a webtoon feel.

Updated November 3, 2026


---

# AI Sci-Fi Short Drama: Starships, Cyberpunk Cities and Helmets

URL: https://scenemixer.com/guides/make-an-ai-sci-fi-short-drama/
Space stations, neon cities and a face behind a visor: how to write sci-fi that AI video renders, keep helmeted characters recognisable and stage affordable battles.

# How to Make an AI Sci-Fi Short Drama

Quick answer: Sci-fi on a small budget has always been about people in rooms: a bridge, a cargo bay, a rain-soaked street under a corporate tower, with the big world implied through windows and screens. That is also what AI video renders best. In SceneMixer, pick the world in step 1 (Starship, Mars Base, Cyber Metropolis or Neon Nights sets; Starship Uniform, Space Explorer or Cyberpunk Augment wardrobe), make a helmet or visor a state so faces are only hidden when you choose, put story information in dialogue rather than on screens, and stage a space battle as reactions on the bridge rather than as a hundred ships.

Brief for AI Create: A sci-fi thriller: a disgraced fleet commander steals the command core of an orbital station under siege; the empire sends her former second-in-command to bring it back; she has until the station's orbit decays; tense, cold, one twist per episode.

## Sci-fi that works as a short drama

- A clock. Oxygen running out, an orbit decaying, a signal arriving in six hours. Science fiction gives you deadlines for free; use one per arc.
- A small cast in a closed place. A ship, a station, a corporate floor. Few locations, each with its own reference image, kept the same every episode.
- A question about people. Who is human, who is loyal, who remembers what. The technology is the setting; the episode endings are about the characters.
- A world seen through a window. The city below, a planet turning, a fleet on the horizon. Scale lives in the background of a shot about two people.

## Settings to pick in step 1

- Visual Style: Digital HD or Shot on Film for live-action sci-fi; Game Engine Realism or Realistic 3D for a CG look; Cel-shaded 3D or Anime for anime sci-fi.
- Scene Style: Starship for bridges and corridors, Mars Base for a colony, Cyber Metropolis or Neon Nights for a cyberpunk city, Wasteland for after the collapse, Alien World for an alien planet (purple sky, glowing fungal forests, crystal pillars).
- Costume Style: Starship Uniform for a fleet, Space Explorer for suits and EVA gear, Cyberpunk Augment for implants and street armour, Near-Future Techwear for the day after tomorrow, Wasteland Scavenge for survivors, Alien Tribal for non-human peoples.
- Type and Aspect Ratio: Film and 16:9 Landscape for a sci-fi short; Short Drama and 9:16 Portrait for a feed series.

Scene Style: Starship

Scene Style: Mars Base

Scene Style: Cyber Metropolis

Costume Style: Starship Uniform

Costume Style: Space Explorer

Costume Style: Cyberpunk Augment
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Helmets, visors and masks

A helmet hides the one thing that keeps a character recognisable. Make it a choice, not a default:

- Generate the character’s main image with the face visible: the uniform or suit, no helmet.
- On their card choose Add a state, name it Helmet on, and describe only the difference: a matte white helmet with a gold-tinted visor down; the blue stripe of the uniform visible on the collar.
- Pick the state only in the shots where the helmet is on. The shot where it comes off is the shot that gets the face back — often the best reveal in the episode.
Give helmeted characters something the audience can still read: a coloured stripe, a callsign on the shoulder, a damaged visor. Use a character state in one scene · My character’s face changed.

Character sheets from the sample Starfall Command: one reference image per character, used in every shot they appear in.

## Screens, readouts and holograms

Sci-fi is full of displays, and a generated display will not hold readable text or numbers. Use screens as light and shape; put the information in the dialogue:

- The console throws red warning light across her face — then she says Orbit decays in forty minutes.
- A hologram of the station turns slowly above the table, one section pulsing — then the engineer points and says which one.
The same goes for signs in a cyberpunk city: they are colour and glow, not words. Text in the picture.

## Battles you can afford

A fleet engagement with hundreds of ships, or a mech fight with complex moving parts, is the kind of shot that fails most often. Stage it as cinema has always staged it on a budget:

| Beat | Shot to adapt |
| The threat | [Shot 2] 4s: Wide shot through the bridge window. Lights of a fleet spread across the black, closing. In frame only: @Kira, back to camera. No dialogue. Ambient: a low alarm. |
| The hit | [Shot 3] 3s: Medium shot, handheld. The bridge shakes; @Kira grabs the rail; sparks from a panel behind her. No dialogue. Ambient: a deep metallic boom. |
| The decision | [Shot 4] 4s: Close-up. In frame only: @Kira, red warning light on her face. @Kira (calm, too calm): Turn every turret on their flagship. |
The battle is in the window, the shaking and the sound; the drama is in the close-up. Fights, crowds and group scenes.

## Robots, androids and AIs

- An android who passes as human is a character like any other. If there is a reveal — a panel opening, a seam of light at the jaw — make that a state.
- A robot is a non-human character: describe shape, material, how it moves, whether it has a face or a screen. Non-human characters.
- A ship’s AI that is only heard is set to Voice only (off-screen) in step 2: no image, a voice. Describe it as you would any voice: neutral, warm, slightly too patient.

## Make it in SceneMixer, start to finish

- Script: yours on Upload Script, or a brief in AI Create with the clock, the closed place, the lead and the twist.
- Step 1: the sci-fi settings above; read the script.
- Step 2: generate the cast and the few locations; add helmet states; set the ship’s AI to voice only.
- Storyboard: episode 1 first (free on your first project). Move readouts into dialogue; turn battles into bridge reactions.
- Video: generate on the default tier at 10 credits a second; regenerate what misses; composite.
Watch Starfall Command, a siege on an orbital station, and Neon Debt, a cyberpunk chase through a vertical megacity.

| Story | Scene style | Costume style | Visual style |
| Space opera on a warship | Starship | Starship Uniform | Digital HD |
| First colony on Mars | Mars Base | Space Explorer | Shot on Film |
| Cyberpunk thriller | Cyber Metropolis / Neon Nights | Cyberpunk Augment | Game Engine Realism |
| Near-future corporate | Corporate Cold | Near-Future Techwear | Digital HD |
| Post-apocalypse | Wasteland | Wasteland Scavenge | Shot on Film |

## FAQ

### Can AI make a sci-fi short film?

Yes. Pick Film as the type and a sci-fi set and wardrobe in SceneMixer's step 1, then write or generate the script. Keep the cast small and the locations few, and the world stays consistent.

### How do I keep a character recognisable in a helmet?

Generate the main image without the helmet, add the helmet as a state, and pick it only in the shots that need it. Give the helmet a mark the audience can read.

### Can the screens show readable data?

Not reliably. Generated text and numbers on displays are not legible. Use the screen as coloured light and let a character say the information.

### Can AI video do a space battle?

A battle with hundreds of ships is unreliable. Show the fleet through a window, shake the bridge, use the sound, and keep the drama in the close-ups.

### How do I make a ship's computer that talks?

Create it as a character set to Voice only (off-screen) in SceneMixer. No image is made; its lines are spoken in the voice you describe.

### Which style suits cyberpunk?

Cyber Metropolis or Neon Nights for the set, Cyberpunk Augment for the wardrobe, and Digital HD or Game Engine Realism for the visual style.

### Can I use a famous franchise's ships or uniforms?

No. Known franchises and their designs trigger the IP check. Invent your own fleet, insignia and ships.

### Does sci-fi cost more to generate?

No. SceneMixer prices video by the second at the tier you pick, whatever is in the shot.

Updated November 3, 2026


---

# AI Second Chance Romance: Exes, Reunions and Two Timelines

URL: https://scenemixer.com/guides/make-an-ai-second-chance-romance/
Make a second chance romance with AI: past and present as two states of each face, flashbacks in their own look, the place they met as a location, and the reunion beat.

# How to make an AI second chance romance

Quick answer: A second chance romance runs on two timelines: who they were then and who they are now. In SceneMixer, give each lead a younger state and a present state in Step 2, so the same face reads at two ages, keep the place they first met as one location seen in both times, and put the flashback in its own segments with different time and weather. End episodes on a look, not a kiss.

Brief for AI Create: A second chance romance: ten years after he left their seaside town without a word, a chef comes back to buy the restaurant where she still works; twelve short episodes, present and past interleaved.

## Two ages, one face

Write the main sheet as the present. Add a state for the past: younger, different hair, the clothes of that summer. Pick the past state in every flashback segment. The face stays theirs because the state comes from the same character. States · Wrong age fixes.

## The place they met

| Where they met | Scene Style | Costume Style |
| A seaside town | Mediterranean Coast | Beach Resort then, Smart Casual now |
| A small hometown | American Small Town | Soft Everyday |
| University | Elite Boarding School or Campus Youth | Prep School Uniform then, Power Tailoring now |
| A big city | Parisian Streets, New York Metropolis | Minimal Luxury |
The same location in both timelines makes the change obvious. Set summer and sun for the past, winter or rain for the present, on the segment's Time and Weather. Same room in every episode.

## The reunion, shot by shot

| Beat | Shot to adapt |
| The door | [Shot 1] 3s: Wide shot, the restaurant at closing. The bell over the door rings; @Lena looks up from the counter. |
| Him | [Shot 2] 3s: Medium shot. @Marco in the doorway, rain on his coat. No dialogue. |
| The past | [Shot 3] 3s: Flashback, same counter, summer light. Younger @Lena laughing at younger @Marco. |
| The line | [Shot 4] 4s: Close-up on @Lena. @Lena (steady): We're closed. We've been closed for ten years. |

## Story shapes

- The one who left: the reason is the season's secret.
- The divorce that was a mistake: forced to work together again.
- The wedding: one of them is marrying someone else.
- The child: a reason to see each other every week.
Hooks and cliffhangers · Billionaire romance.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Mediterranean Coast, American Small Town; Costume Style Smart Casual, Beach Resort; Visual Style Soft Glow, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: a present and a past state for each lead; the restaurant as the shared location; flashbacks set to summer. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Now | Then |
| 1 | He walks in | The summer they met |
| 2 | He makes an offer for the restaurant | Their first argument |
| 3 | She refuses | The night before he left |
| 4 | They cook together for one night | The letter she never got |
| 5 | Someone else proposes | Why he left |

## FAQ

### How do I make a second chance romance with AI?

Give each lead a past and a present state, keep the place they met as one location in both times, and interleave flashback segments.

### How do I show a younger version of a character?

Add a younger state in SceneMixer's Step 2 and pick it in every flashback segment.

### Can flashbacks look different?

Yes. Set different time and weather on those segments; the visual style stays the same for the project.

### Which style suits it?

Soft Glow for a dreamy past, Shot on Film for warmth, Digital HD for a modern feel.

### How long should episodes be?

One to three minutes, each ending on a look or a line.

### Can I make it vertical?

Yes. 9:16 Portrait for short-drama feeds.

### Can there be kissing?

Romance is fine; explicit content is not.

### Can I reuse the leads in a sequel?

Yes. Pick them from your SceneMixer library in a new project.

Updated November 3, 2026


---

# AI Secret Baby Drama: The Child With His Eyes

URL: https://scenemixer.com/guides/make-an-ai-secret-baby-drama/
Make a secret baby drama with AI: a time jump told through states, a child whose look echoes a parent, the reveal as reaction shots, and how to write child characters.

# How to make an AI secret baby drama

Quick answer: The secret baby story runs on one picture: a child who has someone's eyes. In SceneMixer, write the shared feature into both sheets (the same grey eyes, the same cowlick), tell the years between through states (the leads before and after), and play the reveal on faces, not words. The child is a character like any other, with an age written into the description and a voice that says how young they sound.

Brief for AI Create: A secret baby drama: five years after she left the city without a word, a florist delivers to the wedding of the man she loved and finds her son already talking to him by the fountain; twenty short episodes.

## The resemblance on the sheet

Put the inherited detail in both descriptions in the same words: pale grey eyes, a cowlick on the left for the father and pale grey eyes, a cowlick on the left, five years old for the son. Say the age and the build for the child, or the sheet can come out older or younger than you meant. Describe a face · Wrong age.

## Five years in two states

| Character | Before | After |
| The mother | City clothes, long hair | Work apron, hair tied back, a bike |
| The father | Young founder in a hoodie | Suit, tired eyes, a wedding coming |
| The child | (not born) | Five, a backpack, the grey eyes |

Each column is a state of one character, so the face stays the same while the life around it changes. States.

## The reveal, on faces

| Beat | Shot to adapt |
| The meeting | [Shot 1] 4s: Wide shot, a hotel garden with a fountain. @Noah, five, talks to @Daniel, who crouches to his height. |
| The detail | [Shot 2] 2s: Close-up, the boy's grey eyes looking up. |
| The second detail | [Shot 3] 2s: Close-up, @Daniel's grey eyes, the same shape. |
| She arrives | [Shot 4] 3s: Medium shot. @Mia stops on the path, holding a box of flowers. |
| Nobody speaks | [Shot 5] 3s: Close-up on @Mia. The box slips in her hands. |

Two matching close-ups say more than a line of dialogue. Shot-to-shot continuity.

## Child characters

Children are ordinary characters: a card, a sheet, a voice description (a five-year-old boy, bright and fast). Keep them in family scenes; any sexualisation of minors is zero tolerance, and a photo of a real child needs a parent's consent and is likely to be refused by the face check. Voice descriptions · Real faces.

## Where it lives

The secret baby is a billionaire-romance staple: billionaire romance, hidden heiress, single dad and nanny. Glossy places: Urban Opulence, Suburban Mansion, Corporate Cold.

Scene Style: Urban Opulence

Scene Style: Suburban Mansion

Scene Style: Corporate Cold

Costume Style: Power Tailoring

Costume Style: Soft Everyday
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Suburban Mansion; Costume Style Power Tailoring, Soft Everyday; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the parents with a before and an after state; the child with an age and a voice description; the shared feature written the same way in both sheets. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Beat |
| 1 | The night, the leaving |
| 2 | Five years later: the mother and son |
| 4 | The meeting by the fountain |
| 8 | He counts the years |
| 14 | The test, or the confession |
| 20 | The family, chosen |

## FAQ

### How do I make a secret baby drama with AI?

Write the shared feature into both sheets, show the years through states, and play the reveal in matching close-ups.

### How do I make the child look like the father?

Use the same words for the inherited detail in both descriptions: eye colour, hair, a dimple.

### How do I show a time jump?

Give each lead a before and an after state; the face stays, the clothes and life change.

### Can children be characters?

Yes. Write their age into the description and the voice; keep them in family scenes.

### Can I use a photo of a real child?

Only with consent, and realistic photos of real people are often refused by the face check; describe the child instead.

### Which format suits it?

Short Drama, 9:16 Portrait, one to two minutes an episode.

### Which look suits it?

Digital HD for the glossy billionaire feel; Soft Glow for a gentler story.

Updated October 31, 2026


---

# AI Shojo Romance Anime: Blushes, Cherry Blossoms and Confessions

URL: https://scenemixer.com/guides/make-an-ai-shojo-romance-anime/
Make a shojo romance anime with AI: Watercolor Anime and Anime styles, Japanese High School and Soft Romance settings, seasons as segment settings, and a confession.

# How to make an AI shojo romance anime

Quick answer: A shojo romance is a school year, two people who misunderstand each other and a confession under the right sky. In SceneMixer, choose Watercolor Anime or Anime in Step 1, Japanese High School (sliding classroom doors, shoe lockers, cherry blossom in spring and red leaves in autumn) for the school and Japanese School Uniform for the students. Small moments carry this genre: give a glance, a dropped umbrella or a note its own shot.

Brief for AI Create: A shojo romance anime: a transfer student who cannot say no is mistaken for the girl who left a love letter in the school's coldest boy's shoe locker; she decides to play along until she finds the real writer; 10 episodes, each ends with a near-confession.

## School, city or café

| Setting | Scene Style | Costume Style |
| A Japanese school year | Japanese High School | Japanese School Uniform |
| A Korean school and Seoul | Korean High School | Korean School Uniform |
| A café and a small flat | Soft Romance | Soft Everyday |
| A traditional town | Japanese Machiya | Kimono for festivals |
High school dramas.

## Seasons, rain and small moments

- Cherry blossom and red leaves belong to SceneMixer’s Japanese High School style; the season is written in the shot.
- Rain and time of day are segment settings: Rain for the shared umbrella, Dusk for the rooftop. Segment header.
- The letter is a prop. Props.
- Inner voice carries the heroine's thoughts. Voice-over and inner voice.
- Festival yukata is a state of each character. States.

## The umbrella, shot by shot

| Beat | Shot to adapt |
| The entrance | [Shot 1] 4s: Wide shot, a school entrance in rain, rows of wooden shoe lockers. @Hina stands without an umbrella. |
| The umbrella | [Shot 2] 3s: Close-up. A dark blue umbrella tilts over her head. |
| The boy | [Shot 3] 3s: Medium shot. @Sota looks away, ears red. |
| The line | [Shot 4] 4s: Close-up on @Hina. @Hina (inner voice): He's the one from the letter. Isn't he? (voice: Hina) |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Japanese High School; Costume Style Japanese School Uniform; Visual Style Watercolor Anime, Anime; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with distinctive faces and festival states; the letter as a prop; the lockers, the rooftop and the gate as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The shoe lockers. | He thinks the letter was hers. |
| 2 | The rooftop. | She says yes instead of explaining. |
| 3 | The rainy entrance. | One umbrella, two people. |
| 4 | The summer festival. | The real writer is her best friend. |
| 5 | The school gate in autumn. | He says he knew all along. |

## FAQ

### How do I make a shojo romance anime with AI?

Choose Watercolor Anime or Anime with Japanese High School and Japanese School Uniform in SceneMixer's Step 1, and build each episode from small moments in short shots.

### Can I have cherry blossom and autumn leaves?

Yes. The Japanese High School style in SceneMixer includes both seasons; write the season into the shot.

### How do I add rain?

Set Weather to Rain on the segment header in SceneMixer.

### How do I show what the heroine thinks?

Use her inner voice in the shot.

### Can characters wear yukata for a festival?

Yes. Add it as a state of each character in SceneMixer.

### Can the dialogue be in Japanese?

Yes. Japanese is one of the dialogue languages on SceneMixer.

### Can I copy a famous shojo manga?

No. Known characters trigger the IP check.

### Vertical or landscape?

9:16 Portrait for short feeds, 16:9 Landscape for YouTube.

Updated November 3, 2026


---

# AI Shonen Anime: Rivals, Training Arcs and Tournaments

URL: https://scenemixer.com/guides/make-an-ai-shonen-anime/
Make a shonen anime with AI: a hero with one dream, a rival drawn as his mirror, training arcs as states, power-ups as light, and tournament fights one move per shot.

# How to make an AI shonen anime

Quick answer: Shonen is a hero with one dream, a rival who pushes him, and fights that are really about will. In SceneMixer, pick Anime or Cel-shaded 3D, design the hero and the rival as mirrors (opposite colours, same height), show training as states, make power-ups light and effect, and write every fight one move per shot with the reaction of friends in between.

Brief for AI Create: A shonen anime: a boy who cannot use magic enters a school tournament for mages armed only with a broom and his late father's footwork; twelve episodes, one rival, one final.

## Hero and rival as mirrors

Shonen pairs read at a glance: spiky dark hair against neat white hair, red against blue, a grin against a stare. Write those contrasts into the two sheets. Describe a face.

| Setting | Scene Style | Costume Style |
| School tournament | Japanese High School, Magic Academy | Japanese School Uniform, Academy Robes |
| Street fighters | Neon Nights | Streetwear |
| Martial journey | Ink Jianghu | Jianghu Travelwear |
| Sports | A court or field as a location | Sportswear |

## Training arcs are states

Torn sleeves, bandaged hands, a new headband, the final-form aura: each is a state, picked in the segments after it is earned. The visual change is the audience's reward for the training episode. States.

## A tournament exchange

| Beat | Shot to adapt |
| The arena | [Shot 1] 3s: Wide shot, a stone arena, students in the stands. @Sora with a broom; @Ren with a glowing staff. |
| The attack | [Shot 2] 2s: @Ren swings; a wave of blue light rushes forward. |
| The footwork | [Shot 3] 2s: Low angle. @Sora sidesteps, sandals skidding on stone. |
| The friends | [Shot 4] 2s: Close-up on @Mika in the stands, fists clenched. |
| The counter | [Shot 5] 3s: @Sora taps the staff out of @Ren's hand with the broom handle. |
Anime fight scenes.

## Shonen beats

- The dream said out loud in episode 1.
- The loss that starts the training arc.
- The rival's respect, earned at the midpoint.
- The friends' faith in the reaction shots.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Magic Academy, Japanese High School; Costume Style Academy Robes, Japanese School Uniform; Visual Style Anime, Cel-shaded 3D; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: hero and rival as mirrors; training and final-form states; the arena and the training ground as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Arc | The ending beat |
| 1 | The dream | He enters with a broom |
| 3 | The first loss | Ren does not even look at him |
| 5 | Training | His father's notebook |
| 9 | The semi-final | The new state, first used |
| 12 | The final | Ren offers his hand |

## FAQ

### How do I make a shonen anime with AI?

Pick an anime look, design hero and rival as mirrors, show training as states, and write fights one move per shot.

### How do I show power-ups?

As light and effect around the character, and a new state for a lasting change.

### Can AI do anime fights?

Yes, one move per shot, with reaction shots in between.

### Which look suits shonen?

Anime for classic, Cel-shaded 3D for camera moves and impact.

### Can the dialogue be Japanese?

Yes; set Japanese in SceneMixer's Step 1.

### Can I use a famous shonen character?

No. Known characters and series trigger SceneMixer's IP check.

### Can I show injuries in fights?

No wounds; show exhaustion and torn clothes as states.

### Series or film?

A series suits the arc structure.

Updated November 3, 2026


---

# How to Make an AI Short Film From a Script, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-short-film/
From script to finished file: set the Film type, write for longer takes, keep one cast consistent, generate scene by scene and composite a short film to submit or post.

# How to Make an AI Short Film

Quick answer: An AI short film in SceneMixer starts from a script, not a prompt. Paste a screenplay or write a brief in AI Create, choose Film as the Type in step 1 — slower pace, room to breathe and no cap on episode length — and 16:9 Landscape, the format the site recommends for film and large screens. The script is read into a cast, locations and props with reference images, so every face and room holds from the first scene to the last; then you generate the film segment by segment, fix individual shots, and composite it into one file you can download at full quality. A three-minute film is about 180 seconds of video, billed by the second at the tier you choose.

Brief for AI Create: A short film: a lighthouse keeper on his last night before the light is automated finds a letter from the keeper before him, addressed to whoever comes last; quiet, wintry, ends on the light going out by itself.

## Film or Short Drama: the one setting that changes everything

The Type row in step 1 decides how the script is cut and how each scene is written for the camera. For a short film, pick Film:

| | Film | Short Drama |
| Pace | Slow, atmospheric, room for silence | Fast, dense dialogue, hook-driven |
| Episode length | No cap | Capped, built for feeds |
| Structure | One piece divided into acts | Episodes that each end on a hook |
| Best for | Festival shorts, portfolio pieces, YouTube | TikTok, Reels, short-drama apps |
The type is read with the script and cannot be changed in that project afterwards. Short Drama, TV Series or Film.

## Write the script for a camera that is generated

A short film script for SceneMixer is an ordinary screenplay with a few habits that make generation reliable.

- Scene headings, character cues, dialogue. Standard screenplay layout is read as it is: INT./EXT., the place, the time, then action and lines. Use a screenplay as input.
- Few locations. Every location gets its own reference image and is kept the same each time you return to it. Two to four places is the right scale for a short.
- Action you can see. “He hesitates before opening the letter” gives the camera something; “he remembers his father” does not.
- Lines you can say in the time. About three words a second in English. A long speech becomes several shots.
- Silence on purpose. Film is allowed to be quiet. Shots with no speech are marked No dialogue in the storyboard so nobody is made to talk.
Write the dialogue exactly as it should be spoken; lines written as dialogue in your script are carried into the storyboard word for word. Write dialogue so it is picked up.

## Settings to pick in step 1

- Type: Film.
- Aspect Ratio: 16:9 Landscape, marked Recommended and described as Film / Large screen. Choose 9:16 Portrait only if the film is made for phones.
- Visual Style: Shot on Film for a cinema grade, Digital HD for a clean, filmed look, Black & White Film for monochrome, or any of the animation groups — Watercolor Anime, Family 3D Animation, Claymation — for an animated short.
- Scene Style and Costume Style: pick the world the film lives in; both are read with the script and fixed afterwards. Visual style presets, one by one.

Visual Style: Digital HD

Visual Style: Black & White Film

Visual Style: Watercolor Anime

Visual Style: Family 3D Animation

Visual Style: Claymation
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## How the film is made, scene by scene

- The read. SceneMixer turns the script into a summary, a cast list, the locations and the props. Read the summary before you continue; it is the cheapest place to catch a misunderstanding.
- The setup images. Each character, location and prop gets a reference image. Faces are decided here: describe them precisely and regenerate the ones that are generic. Describe a face so it is not a generic face.
- The storyboard. Each part of the film is written out shot by shot: framing, who is in frame, the action, the line, the sound. Editing the text is free. How to write one shot.
- The segments. Video is generated a segment at a time, each segment up to 30 seconds on the default tier, with dialogue and sound produced in the same pass as the picture. 15 or 30-second segments.
- The fixes. Regenerate one shot without redoing the rest, trim a clip, or redo a line. Edit this clip.
- The composite. When every segment is finished, compositing joins them into one file, using whichever version of each segment is active. Composite and share.
If the read splits a longer script into several parts, composite each one and join the files in any video editor.

Character sheets from the sample The Last Envelope: one reference image per character, used in every shot they appear in.

## Writing for longer takes

Film pacing lets a shot hold. Three habits make a held shot work:

- One movement per shot. A slow push in, a pan across the room, a figure crossing frame. Two movements become a new shot.
- Let the light do something. A lamp switched on, a cloud moving off the sun, headlights crossing a ceiling. Change inside a held frame is what keeps it alive.
- Sound carries time. The Ambient line at the end of a shot — wind, a kettle starting to whistle, a radio in another room — makes a quiet shot feel inhabited.

| Shot | Example |
| An opening that holds | [Shot 1] 6s: Opening: wide shot, fixed camera. The lighthouse at dusk, the lamp room dark. Snow crossing the frame. No dialogue. Ambient: wind, the sea against rock. |
| A slow reveal | [Shot 2] 5s: Slow push in. In frame only: @Arthur, seated at the keeper’s desk, turning the @Envelope over in his hands. No dialogue. Ambient: the clock in the stairwell. |
| A line with silence around it | [Shot 3] 6s: Medium close-up. In frame only: @Arthur. He reads, stops, reads the line again. @Arthur (quietly, to himself): Whoever comes last. Ambient: wind. |

## Music, the AI label and submitting your film

- Music. An episode that has its own score gets a version of the composite with the music carried across the whole timeline; each composite on SceneMixer has a Music switch, on by default. Switch it off for a version without music if you want to score the film yourself. Background music.
- Full quality. Download gives you the original, full-quality file, not the lighter copy the player streams. Download: original vs play version.
- The AI label. Free accounts carry a small AI label in the corner; on a paid plan it is off by default. Festivals and platforms may have their own rules about disclosing AI-made work; turning the label off does not remove that obligation. The AI badge switch · What you get to keep.

## What a short film costs

Four kinds of charge, each shown before you confirm: the script read by length, the setup images (the first 15 on an account are free), the storyboard text, and the video by the second — 10 credits a second on SceneMixer's default tier, more on higher resolutions. A three-minute film is 180 seconds of video; compositing adds 1 credit a second. Drafting at 480P and regenerating the best takes at a higher tier is usually cheaper than shooting everything high. A worked budget · Draft at 480P, finish at 1080P.

For a film built this way, watch The Last Envelope, a wartime espionage piece in a bombed, mirror-hall London, or The Violin on Mercer St, a quiet noir mystery.

| Length | Scenes | Locations | Roughly how much video |
| 1 minute | 2–3 | 1–2 | 60 seconds |
| 3 minutes | 4–6 | 2–3 | 180 seconds |
| 5 minutes | 6–10 | 3–4 | 300 seconds |
| 10 minutes | 10–20 | 4–6 | 600 seconds, usually split into parts |

## FAQ

### Can AI make a whole short film, not just clips?

Yes. SceneMixer reads the script into a cast, locations and a storyboard, the film is generated segment by segment with the same faces and rooms throughout, and compositing joins the segments into one file.

### Which type should I pick for a short film?

Film. It gives a slower pace, room for silence and no cap on episode length. Short Drama is for feed series that end every episode on a hook.

### Is 16:9 or 9:16 better for a short film?

16:9 Landscape, which SceneMixer recommends for film and large screens. Use 9:16 only if the film is meant for phones. The choice is fixed once you continue past step 1.

### Can I use my own screenplay?

Yes. Paste it into SceneMixer or upload a TXT, DOCX or Markdown file. Scene headings and character cues are read as they are, and dialogue lines are kept word for word.

### Does the film have sound?

Yes. SceneMixer generates dialogue, ambient sound and the picture together, and an episode can carry its own score, which you can switch off on the composite if you want to add your own.

### Can I submit an AI short film to a festival?

That depends on the festival's rules; many now have categories or disclosure rules for AI-made work. Read them before you submit, and disclose where they ask.

### How long does it take to make a three-minute film?

The read takes minutes; each segment on SceneMixer usually takes ten to twenty minutes to generate, depending on the tier and the queue, and several can run at once. Most of the time goes into reading the storyboard and fixing the shots you are not happy with.

### What resolution can a short film be?

It depends on the video tier you generate with on SceneMixer, from 480P on the default tier to 1080P and above on higher tiers. Each tier's price per second is shown before you generate.

Updated November 3, 2026


---

# Make an AI Short Film for a Festival: Script, Specs and Delivery

URL: https://scenemixer.com/guides/make-an-ai-short-film-for-a-festival/
Make an AI short film for a festival: Film type, 16:9, Enhance up to 2K, a badge-free member download, subtitles with SRT, and the festival's own rules.

# How to make an AI short film for a festival or competition

Quick answer: A festival short is judged as a film: story, craft and delivery. In SceneMixer, make it a Film project in 16:9 Landscape, render, then raise the best takes with Enhance (up to 2K) and download the full-quality file. Members can switch the AI badge off for clean downloads; subtitles can be burned in or exported as SRT. Every festival sets its own rules on AI work, length and disclosure, so read them before you start.

Brief for AI Create: A short film for an AI film festival: a lighthouse keeper on the last night before the light is automated receives a call from a ship that sank in 1952; eight minutes, two characters, one location, a quiet ending.

## Before you write

- Read the festival's rules: length limits, AI categories, disclosure, resolution, subtitles. They decide; we do not. Do AI videos need a label?.
- Small story, big craft. One location and two characters let you spend on the shots that matter. Make an AI short film.
- Original world. SceneMixer refuses known characters and real people. IP check.

## Craft

- Type Film, no forced hooks. Series or film.
- A considered look: Shot on Film, Black & White Film or a drawn style, fixed after Step 1. Visual styles.
- Fix single shots rather than whole sequences: trim, redub, or regenerate one shot. Edit this clip.
- Music is scored automatically under the voices; the Music switch gives a dialogue-only version. Background music.

## Delivery

- Resolution: render on a higher tier, or raise a finished clip with Enhance, one step at a time up to 2K. Enhance.
- No badge: members can switch the AI badge off before rendering. The AI badge switch.
- Full quality: download the original, not the play version. Original vs play version.
- Subtitles: burned in or as SRT (accounts with a purchase). Subtitles.
- Poster: a promo cover can be generated from the project. Promo cover.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Frozen Wilderness, Mediterranean Coast; Costume Style Outdoor Gear; Visual Style Shot on Film, Black & White Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: two characters with specific faces; one or two locations; render the opening sequence and review it before the rest. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Step | What to do | Where |
| 1 | Read the festival rules. | The festival's site. |
| 2 | Write a small, original script. | Upload Script or AI Create. |
| 3 | Film type, 16:9, a considered look. | Step 1. |
| 4 | Render, fix single shots, enhance. | The editor. |
| 5 | Badge off, download original, SRT. | Editor and export. |

## FAQ

### How do I make an AI short film for a festival?

Write a small original script, make a SceneMixer Film project in 16:9, render and fix single shots, enhance, and deliver a full-quality file under the festival's rules.

### What resolution can I deliver?

It depends on the tier; Enhance on SceneMixer raises a clip one step at a time up to 2K.

### Can I remove the AI badge?

Members can switch it off on SceneMixer; the setting applies to videos generated afterwards.

### Do festivals accept AI films?

Many have AI categories; each sets its own rules. Read them first.

### Can I get an SRT file?

Yes, when exporting an episode on a SceneMixer account that has made a purchase.

### Can I use real actors' faces?

No. Real people cannot be depicted.

### Can I turn the music off?

Yes. The Music switch in SceneMixer gives the dialogue-only version.

### Which look suits a festival short?

Shot on Film or Black & White Film, or a drawn style used consistently.

Updated November 3, 2026


---

# AI Single Dad and Nanny Romance: A Big House and a Small Kid

URL: https://scenemixer.com/guides/make-an-ai-single-dad-and-nanny-romance/
Make a single dad and nanny romance with AI: Suburban Mansion homes, a child character who stays the same, Soft Everyday costumes, and small scenes that grow into love.

# How to make an AI single dad and nanny romance

Quick answer: A nanny romance is a cold father, a warm stranger and a child who decides first. In SceneMixer, build the house in Step 1: Suburban Mansion (white colonnades, double-height halls, a pool) or Urban Opulence for a city penthouse; dress the nanny in Soft Everyday and the father in Power Tailoring. The child is a character with a sheet like anyone else, so the same small face appears in every episode.

Brief for AI Create: A nanny romance: a nursing student takes a summer job as a nanny for a widowed CEO's six-year-old who has not spoken since her mother died; on day one the girl whispers one word to her; 10 short episodes, each ends with the father seeing something he missed.

## Set the house

| Kind of family | Scene Style | Costume Style |
| A suburban mansion | Suburban Mansion | Soft Everyday, Power Tailoring |
| A city penthouse | Urban Opulence | Minimal Luxury |
| A ranch or farm | Wild West or Pastoral Healing | Western Cowboy, Outdoor Gear |
| A mafia boss's house | Urban Opulence | Power Tailoring |
Mafia romance · Billionaire romance.

## The child as a character

- A sheet like anyone else. Describe the child's age, hair and one thing she always carries (a toy rabbit). Describe a face.
- The toy is a prop, the same rabbit in every episode. Props.
- A child's voice comes from the preset library; play them before choosing. Voices.
- Small scenes carry this genre: breakfast, a scraped knee, a bedtime story. One beat per shot. One shot.

## The first word, shot by shot

| Beat | Shot to adapt |
| The kitchen | [Shot 1] 4s: Wide shot, a white kitchen, morning sun. @Lucy sits at the island holding a grey toy rabbit; @Hannah pours cereal. |
| The doorway | [Shot 2] 3s: Medium shot. @Adam stops in the doorway in a dark suit, phone at his ear. |
| The whisper | [Shot 3] 4s: Close-up on @Lucy leaning towards Hannah. @Lucy (whisper): Pancakes. (voice: Lucy) |
| The father | [Shot 4] 3s: Close-up on @Adam lowering the phone. |

## Story shapes that work

- The silent child.
- The fake nanny: she is there for another reason.
- The ex-wife returns.
- The contract: one summer, no feelings.
- Twins: two children, two plans to get rid of her.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Suburban Mansion; Costume Style Soft Everyday, Power Tailoring; Visual Style Soft Glow, Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the father, the nanny and the child with one mark each; the child's toy as a prop; the kitchen, the pool and the office as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The mansion's gate. | The girl hides behind the staircase. |
| 2 | The kitchen. | One word: pancakes. |
| 3 | His office. | He finds the drawings the girl made of the nanny. |
| 4 | The pool. | The girl laughs, and he hears it from upstairs. |
| 5 | A gala. | His sister-in-law wants custody. |

## FAQ

### How do I make a nanny romance with AI?

Set the house with Suburban Mansion or Urban Opulence in SceneMixer's Step 1, give the child a sheet and a favourite toy as a prop, and build the story from small domestic scenes.

### Will the child look the same in every episode?

Yes. The child is a character with a sheet in Step 2.

### Can the child have a voice?

Yes. Choose a voice from the SceneMixer preset library.

### Which costume styles suit this?

Soft Everyday for the nanny, Power Tailoring for the father.

### Which visual style suits it?

Soft Glow or Digital HD; Family 3D Animation for a family cartoon version.

### Can it be a mafia nanny story?

Yes. Keep the house in Urban Opulence and the father in Power Tailoring.

### Vertical or landscape?

9:16 Portrait; this genre lives on short-drama feeds.

### How long are the episodes?

One to two minutes, one small step forward each.

Updated November 3, 2026


---

# AI Sitcom or Comedy Skit: How to Make the Timing Work

URL: https://scenemixer.com/guides/make-an-ai-sitcom-or-comedy-skit/
Comedy is timing: how to write a sitcom episode or sketch so AI video lands the joke, with reaction shots, pauses, one living room and a cast that never changes.

# How to Make an AI Sitcom or Comedy Skit

Quick answer: Comedy fails on screen for one reason more than any other: the timing is off. A joke needs a set-up, a beat, and a reaction, and each of those needs its own seconds. In SceneMixer that means writing the punchline and the reaction as separate shots, putting the pause in the delivery note, and keeping a small cast in one or two rooms that look the same every episode — which is exactly what a sitcom is. Pick TV Series for a classic sitcom episode or Short Drama for one-minute sketches, and let the voices do the work: comedy voices are written into each character’s Voice description.

Brief for AI Create: A sitcom: three flatmates discover their landlord has been living in the cupboard under the stairs for a month; nobody wants to be the one to tell him to leave; each episode one of them tries and fails; dry, awkward, warm.

## What makes a sitcom work

- A trap. People who cannot leave each other: flatmates, a family, colleagues, a shop with one customer who never goes home.
- Fixed characters. The one who panics, the one who lies, the one who is far too calm. Each episode puts them in a situation that shows exactly that trait.
- A situation that escalates. A small lie in scene one, a bigger one in scene two, the whole thing collapsing in scene three.
- Home base. One living room, one kitchen, one office. Comedy is easier when the audience knows where the door is.
- A button. The last line of the episode: one more joke after the story is over.
A sketch is the same machine at a smaller size: one premise, three escalations, a button.

## Timing, written into the shots

AI video plays a shot for the seconds you give it. A joke crammed into one shot with its reaction will usually step on its own laugh. Separate the beats:

| Beat | Shot to adapt |
| The set-up | [Shot 1] 5s: Opening: medium shot, the kitchen. @Priya and @Tom at the table; @Priya slides a mug toward him. @Priya (brightly): So. Who’s going to tell him? @Tom lips closed, no lip movement. Ambient: the fridge humming. |
| The beat | [Shot 2] 2s: Close-up. In frame only: @Tom. He looks at the mug. He looks at her. No dialogue. |
| The punchline | [Shot 3] 3s: Close-up. In frame only: @Tom. @Tom (flat): I’ve been paying him rent. In cash. Through the door. |
| The reaction | [Shot 4] 3s: Close-up. In frame only: @Priya, frozen mid-sip. No dialogue. Ambient: a cupboard door creaks in the hall. |
Shot 2 is the most important shot in the scene and it has no words in it. Shot seconds and timing · How to write one shot.

## Delivery notes are where the comedy lives

The brackets at the start of a line tell the voice how to say it. In comedy they carry half the joke: (far too casually), (whispering, though nobody else is there), (a beat too late), (very slowly, as if to a child). Write the pause where you want it: (pauses, then). Keep each line short enough for its seconds — about three words a second in English — because a rushed punchline is a dead one. Dialogue in the storyboard.

## A cast and a set that never change

- Same faces. Each regular is a character with a card and an image, kept the same in every episode. Give each one a visual signature that matches their trait: the anxious one always in a cardigan, the liar always in sunglasses.
- Same living room. The flat is a location with its own image, so the sofa is where it was last week. Positions in a scene hold from shot to shot. Same room in every episode · Everyone stays put.
- Three named people per shot, at most. A party scene is the background line plus the two people the joke is about. Background people.
- Voices with a temperature. Dry, low, never surprised; high, fast, always apologising. Contrast between voices is a joke machine. How character voices work.

## Settings to pick in step 1

- Type: TV Series for an episode with room for scenes to build; Short Drama for one-minute sketches with a button at the end.
- Aspect Ratio: 16:9 Landscape for a classic sitcom; 9:16 Portrait for sketches on TikTok and Reels.
- Visual Style: Digital HD for live-action comedy; 3D Toy Cartoon, Family 3D Animation or Anime for animated comedy; Claymation for something handmade.
- Scene Style: Nordic Minimal or Soft Romance for a flat, Corporate Cold for an office comedy, American Small Town or Street Life for a neighbourhood one.

Visual Style: 3D Toy Cartoon

Visual Style: Family 3D Animation

Visual Style: Claymation

Scene Style: Nordic Minimal

Scene Style: Corporate Cold
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## What to keep out of the joke

Parody of real people — a celebrity, a politician — is not material here: real faces and voices are not generated, and known characters from other shows trigger the IP check. Write your own cast and let the situation be the target. Keep it broadcast-friendly; anything sexual is not generated. Content rules.

## Make it in SceneMixer, start to finish

- Script: your own on Upload Script, or a brief in AI Create with the trap, the characters’ traits, the situation and the tone.
- Step 1: the settings above; read the script.
- Step 2: generate the regulars and the home base; give each a visual signature.
- Storyboard: episode 1 first (free on your first project). Split every punchline from its reaction and give the beat its own shot.
- Video: generate on the default tier at 10 credits a second; if one joke misses, regenerate that segment only.
- Finish: composite. An episode can carry its own score; switch the composite’s Music off if you would rather the jokes land in silence. Background music.

| Format | Type | Length | Shape |
| Sketch | Short Drama | About a minute | Premise, three escalations, a button |
| Sitcom episode | TV Series | A few minutes | A problem, a lie, a collapse, a button |
| Animated comedy | Short Drama or TV Series | Either | Same as above, with an animation style |

## FAQ

### Can AI make a sitcom that is actually funny?

The jokes are yours; the timing is what you control in the storyboard. Give the punchline and the reaction separate shots, write the pause into the delivery note, and keep lines short enough for their seconds.

### Why does my punchline feel rushed?

The shot has too many words for its seconds, or the reaction is in the same shot. Split it, and give the beat before the punchline its own two-second shot.

### How do I keep the same living room every episode?

Make it a location with its own image. Every episode set there uses the same reference, and positions hold within a scene.

### Can I make an animated sitcom?

Yes. Choose 3D Toy Cartoon, Family 3D Animation, Anime or Claymation in SceneMixer's step 1.

### Can I parody a celebrity or a famous show?

No. Real people's faces and voices are not generated on SceneMixer, and known characters trigger the IP check. Write your own cast.

### Should a sitcom have music?

An episode can carry its own score; you can switch it off on the composite. Many sitcoms play better with only room sound.

### How long should a sketch be?

About a minute for feeds: one premise, three escalations, a button. Pick Short Drama and 9:16.

### Can I keep making episodes with the same cast?

Yes. Add episodes to the same project in SceneMixer and the cast and rooms carry over.

Updated November 3, 2026


---

# AI Slice of Life Anime: Small Moments, Seasons and Quiet Rooms

URL: https://scenemixer.com/guides/make-an-ai-slice-of-life-anime/
Make a slice of life anime with AI: Watercolor Anime looks, a few rooms drawn once, seasons and weather set per segment, small stakes, and silence given real shot time.

# How to make an AI slice of life anime

Quick answer: Slice of life is an anime about nothing much, which is why it is hard: the small moments have to feel true. In SceneMixer, pick Watercolor Anime or Painterly Anime, keep the world to a few rooms and streets drawn once, mark time with seasons and weather on each segment, and give silences their own shots. The stakes are a bento, a train, a goodbye.

Brief for AI Create: A slice of life anime: two sisters run their late grandmother's tiny bakery by the train tracks for one year; one season per episode block, one recipe per episode.

## A small world, drawn once

| Setting | Scene Style | Costume Style |
| School days | Japanese High School | Japanese School Uniform |
| Old town shop | Japanese Machiya | Soft Everyday |
| Countryside | Pastoral Healing | Loungewear, Outdoor Gear |
| City flat | Nordic Minimal | Soft Everyday |
Three or four places, used again and again, become home to the audience. Same room in every episode.

## Time is the plot

Set Time and Weather on each segment: cherry blossom light in spring, cicada heat in summer, the first snow. The same bakery in four seasons is the series. Day, night and weather.

## Silence gets seconds

| Beat | Shot to adapt |
| The morning | [Shot 1] 4s: Wide shot, dawn, the bakery's shutter half open; a train passes behind. Ambient: train, birds. |
| The work | [Shot 2] 4s: Close-up. @Aoi's floury hands folding dough. No dialogue. |
| The sister | [Shot 3] 3s: @Mei in the doorway with two cups of tea. @Mei: Grandma used to hum here. |
| The pause | [Shot 4] 4s: Both sisters at the window, steam rising from the cups. No dialogue. |
Choose TV Series for its slower build; the episode endings can be a feeling instead of a twist. Series or film.

## Small stakes that hold

- A recipe that will not come out right.
- A regular customer who stops coming.
- A letter from a parent abroad.
- The last day before someone moves away.
Shojo romance · Cozy farm life.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Japanese Machiya, Pastoral Healing; Costume Style Soft Everyday; Visual Style Watercolor Anime, Painterly Anime; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: two sisters with specific faces; the bakery and the station as locations; seasons set on the segments. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Season | The small thing |
| 1 | Spring | Grandma's melon bread recipe has a missing step |
| 3 | Spring | A schoolboy buys one bun every day |
| 5 | Summer | The oven breaks in the heat |
| 8 | Autumn | A letter from their mother |
| 12 | Winter | The recipe, finally right |

## FAQ

### How do I make a slice of life anime with AI?

Pick Watercolor Anime or Painterly Anime in SceneMixer, keep to a few locations, mark time with seasons and weather, and give quiet moments their own shots.

### How do I show seasons changing?

Set Time and Weather on each segment in SceneMixer; the location stays the same.

### Which type suits slice of life?

TV Series for slower pacing and soft endings.

### Can the dialogue be Japanese?

Yes; set it in Step 1, with subtitles if you like.

### Is there background music?

Yes, a score is added under the voices, and it can be switched off.

### How long should episodes be?

Three to five minutes.

### Can I copy a famous slice of life anime?

No; known characters and series trigger the IP check.

### Vertical or landscape?

Landscape suits the calm framing.

Updated November 3, 2026


---

# AI Space Opera: Starships, Empires and Crews Across the Stars

URL: https://scenemixer.com/guides/make-an-ai-space-opera/
Make a space opera with AI: Starship and Mars Base worlds, Starship Uniform ranks, the bridge as a recurring location and battles told in short shots.

# How to make an AI space opera

Quick answer: A space opera is a crew, a ship and an empire too big to fight. In SceneMixer, Step 1 builds it: Starship for curved bulkheads, flush panels and stars through the portholes, Mars Base for a colony, Alien World for the planets, and Starship Uniform with department colours and rank marks for the crew. The bridge is a location you will return to every episode; space battles work best as short cuts between the bridge, the screen and the faces.

Brief for AI Create: A space opera: the youngest officer on a damaged warship takes command when the captain disappears in a jump; the crew does not trust her, and the enemy fleet is waiting for a ship it thinks is empty; 12 episodes, each ends with an alarm.

## Build the fleet

| Where | Scene Style | Costume Style |
| Bridges, corridors, engine rooms | Starship | Starship Uniform |
| A frontier colony | Mars Base | Space Explorer |
| Alien planets | Alien World | Alien Tribal |
| The imperial capital | Cyber Metropolis or Epic Scale | Divine Regalia for an emperor, Near-Future Techwear for citizens |
| A smugglers' port | Neon Nights | Cyberpunk Augment |
Every scene and costume style.

## Crew, rank and ships

- Rank shows on the uniform. Starship Uniform uses department colours and rank marks; say each character's role in the script.
- The bridge is a location, the same room in every episode. The same room in every episode.
- The ship seen from outside is a location too, with its own picture. Damage after a battle is a state of that location. States.
- Alien crew members are characters with a body description. Alien dramas.

## The battle, shot by shot

| Beat | Shot to adapt |
| The screen | [Shot 1] 3s: Close-up on the bridge viewscreen: six enemy ships dropping out of a jump. |
| The bridge | [Shot 2] 4s: Wide shot, the bridge in red alert light, crew at their consoles turning to the command chair. |
| The order | [Shot 3] 4s: Close-up on @Lieutenant Vale. @Lieutenant Vale (steady): Cut the engines. Let them think we're dead. (voice: Vale) |
| The doubt | [Shot 4] 3s: Close-up on @Chief Okafor's hand hovering over the engine switch. |
| Outside | [Shot 5] 4s: Wide shot, the dark warship drifting past the enemy fleet against a field of stars. |
How to write one shot · Battles and groups.

## Story shapes that work

- The unlikely captain.
- The rebellion: a small crew against an empire.
- The heir among the stars: a mechanic who is secretly royal.
- The derelict: a ship found empty.
- The long voyage: a generation ship and a mutiny.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Starship, Mars Base; Costume Style Starship Uniform; Visual Style Game Engine Realism, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the crew with one mark each and their rank in the script; the bridge, the engine room and the ship's exterior as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The bridge after the jump. | The captain's chair is empty. |
| 2 | The engine room. | Someone sabotaged the jump drive. |
| 3 | The bridge. | Six enemy ships, and she cuts the engines. |
| 4 | The medical bay. | A message from the missing captain. |
| 5 | A mining colony. | The colony's leader knows her real name. |

## FAQ

### How do I make a space opera with AI?

Choose Starship, Mars Base or Alien World in SceneMixer's Step 1 and Starship Uniform for the crew, make the bridge a recurring location, and cut battles into short shots.

### Will the crew's uniforms show rank?

Yes. Starship Uniform on SceneMixer uses department colours and rank marks; say each role in the script.

### How do I show the ship from outside?

Make the exterior a location with its own picture; add a damaged state for after a battle.

### Can I have alien crew members?

Yes. They are characters with their own sheet and body description.

### Can I film space battles?

Yes, as short shots: the screen, the bridge, a face, the ship outside. One action per shot.

### Can I use names from a famous space saga?

No. Known names, ships and uniforms trigger the IP check. Invent your empire.

### Which visual style suits a space opera?

Digital HD or Game Engine Realism for live action, Realistic 3D or Painterly Anime for animation.

### Vertical or landscape?

16:9 Landscape suits the scale; 9:16 Portrait works for a short-drama feed.

Updated November 3, 2026


---

# AI Sports Drama: Underdogs, Rivals and the Big Game

URL: https://scenemixer.com/guides/make-an-ai-sports-drama/
Make a sports drama with AI: Sportswear team kit, the gym, rink or ring as locations, match moments told in short shots, and a team you can tell apart at a glance.

# How to make an AI sports drama

Quick answer: A sports drama is an underdog, a rival and a final whistle. In SceneMixer, the Sportswear costume style dresses the team in fitted kit with club badges, numbers and strong team colours; the court, rink, ring or pitch is a location you reuse every episode. Games are told in short shots of one action each (the shot, the save, the face on the bench) rather than one long rally. Hockey, boxing, basketball, racing and football all work the same way.

Brief for AI Create: A sports drama: a hockey player thrown off a top team for a fight he did not start joins a small-town team with nine players and no coach; his first game is against his old team; 10 short episodes, each ends with a score.

## Kit, venues and towns

| Sport | Venue as a location | Scene Style for the rest of the world |
| Ice hockey | The rink, the locker room | American Small Town or Frozen Wilderness |
| Boxing | The ring, the old gym | Street Life or New York Metropolis |
| Basketball, volleyball | The school gym | American High School or Japanese High School |
| Football | The pitch, the tunnel | Mediterranean Coast, Latin Colonial Town |
| Racing | The pit lane, the garage | Cyber Metropolis or Gulf Metropolis |
Costume Style Sportswear for every team; invent the club names and badges. High school dramas.

## Teams you can tell apart

- Two colours. Write each team's colours into the script; the opponent in red, your team in green.
- One mark per player. A number, a bandage, a hairstyle. The same face in every shot.
- Kit and street clothes are two states of each player. States.
- The crowd is the background line. Background people.

## The last minute, shot by shot

| Beat | Shot to adapt |
| The clock | [Shot 1] 3s: Close-up on the arena scoreboard: 2–2, 0:12 left. |
| The skate | [Shot 2] 3s: Medium tracking shot. @Cole, number 17 in green, drives the puck up the ice. |
| The rival | [Shot 3] 3s: Medium shot. @Reid in red cuts across to block him. |
| The bench | [Shot 4] 3s: Close-up on @Maggie behind the bench, both hands over her mouth. |
| The shot | [Shot 5] 4s: Close-up on the goal net as the puck hits the post and drops. |
Boxing: one punch or one reaction per shot, and no injuries in close-up. Fights.

## Story shapes that work

- The underdog team.
- The comeback: an injured champion.
- Rivals to lovers on opposing teams.
- The coach with nothing left to lose.
- The scholarship: one game decides a future.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style American Small Town, Street Life; Costume Style Sportswear; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: each player with a number or mark and a street-clothes state; the rink, the locker room and the diner as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The big-city arena. | He is thrown off the team. |
| 2 | A small-town rink. | Nine players and a broken Zamboni. |
| 3 | The diner. | The team captain's sister runs the rink. |
| 4 | Practice. | The schedule: game one is his old team. |
| 5 | The arena. | 0:12 left, 2–2. |

## FAQ

### How do I make a sports drama with AI?

Choose Sportswear for costumes, make the venue a location in SceneMixer's Step 2, give each team colours and each player a mark, and tell the games in short shots.

### Can I use a real team or league?

No. Real clubs, leagues and logos trigger SceneMixer's IP check. Invent them.

### How do I keep two teams apart?

Write each team's colours into the script and give players numbers or marks.

### Can I film a boxing match?

Yes, as short shots of one punch or reaction each. No injuries in close-up.

### How do I show the crowd?

Put it in the background line of the shot.

### Which visual style suits sports?

Digital HD for live action; Anime or Painterly Anime for a sports anime look.

### Can the players change out of kit?

Yes. Kit and street clothes are two states of each player.

### Vertical or landscape?

Either; 9:16 Portrait for short-drama feeds, 16:9 Landscape for wide game shots.

Updated November 3, 2026


---

# AI Spy Thriller: Covers, Handlers and a Mole in the Agency

URL: https://scenemixer.com/guides/make-an-ai-spy-thriller/
Make a spy thriller with AI: a cover identity as a state, Eastern European Old Town and Corporate Cold settings, dead drops as props, and one mole to find.

# How to make an AI spy thriller

Quick answer: A spy thriller is a cover, a handler and a lie that has to hold for one more day. In SceneMixer, the cover identity is a state of the spy (the accountant and the agent are one face, two looks). Build the map in Step 1: Eastern European Old Town for a cold-war city, Corporate Cold for the agency, Gulf Metropolis or Mediterranean Coast for the mission abroad. Make every dead drop, file and phone a prop with its own picture.

Brief for AI Create: A spy thriller: an agency translator is sent undercover as the new assistant to an arms dealer's wife; her handler is feeding the dealer information, and the wife knows exactly who she is; 10 episodes, each ends with a message in a place only a spy would look.

## The map in Step 1

| Kind of spy story | Scene Style | Costume Style |
| A cold-war city | Eastern European Old Town | Smart Casual, with long coats and hats written into each character’s clothing |
| The modern agency | Corporate Cold | Power Tailoring |
| A mission in the Gulf | Gulf Metropolis | Minimal Luxury |
| A summer on the coast | Mediterranean Coast | Beach Resort |
| World War II espionage | WWII Front or Parisian Streets | WWII Era |
World War II dramas.

## Covers, drops and the mole

- The cover is a state. Glasses and a cardigan for the assistant, a dark coat for the agent. States.
- Dead drops are props: the hollow coin, the newspaper, the burner phone. Props.
- Two faces that matter: the handler and the suspect. Make both specific. Describe a face.
- Action is short shots; violence stays off-screen. Fights.

## The dead drop, shot by shot

| Beat | Shot to adapt |
| The square | [Shot 1] 4s: Wide shot, an old town square in snow, a clock tower, a tram passing. @Anna sits on a bench with a newspaper. |
| The drop | [Shot 2] 3s: Close-up on @Anna's gloved hand sliding a coin under the bench slat. |
| The watcher | [Shot 3] 3s: Medium shot through the café window opposite: @Mrs Varga lowers her cup. |
| The line | [Shot 4] 4s: Close-up on @Mrs Varga. @Mrs Varga (to herself): There you are, little translator. (voice: Varga) |

## Story shapes that work

- The translator undercover.
- The mole hunt.
- Two spies, one marriage.
- The burned agent on the run.
- The sleeper who has forgotten she is one.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Eastern European Old Town, Corporate Cold; Costume Style Smart Casual, Power Tailoring; Visual Style Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the spy with a cover state, the handler and the target with specific faces; the coin and the file as props; the square and the house as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The agency. | A translator gets her first cover. |
| 2 | The dealer's house. | The wife hires her on the spot. |
| 3 | The square. | Her drop is collected by the wrong man. |
| 4 | The opera. | Her handler shakes hands with the dealer. |
| 5 | The wife's study. | A file with her real name. |

## FAQ

### How do I make a spy thriller with AI?

Give the spy a cover state in SceneMixer's Step 2, choose a city scene style in Step 1, and make dead drops and files props.

### How do I show a disguise?

Add the cover look as a state of the same character.

### Can I show shootouts?

Short shots of the moment before and after; violence is kept off-screen.

### Can I use real intelligence agencies?

Invent them. Real logos and names can trigger the IP check.

### Which visual style suits a spy thriller?

Shot on Film for a cold-war feel, Digital HD for modern stories.

### Can it be set in World War II?

Yes. WWII Front or Parisian Streets with WWII Era costumes.

### Vertical or landscape?

16:9 Landscape for a series, 9:16 Portrait for feeds.

### Can I copy a famous spy?

No. Known characters trigger the IP check.

Updated November 3, 2026


---

# AI Superhero Drama: Powers, Secret Identities and a City to Save

URL: https://scenemixer.com/guides/make-an-ai-superhero-drama/
Make an original superhero series with AI: the suit as a state, powers described as light and effect, New York or Cyber Metropolis cities, and your own design.

# How to make an AI superhero drama

Quick answer: A superhero drama is a person with a secret, a power with a cost and a city that needs both. In SceneMixer, the secret identity is two states of one character: the everyday clothes and the suit. Set the city in Step 1 (New York Metropolis, Neon Nights or Cyber Metropolis), describe powers as plain light and effect, and design an original suit: famous heroes, logos and colour schemes trigger the IP check.

Brief for AI Create: A superhero drama: a night-shift paramedic discovers she can stop time for three seconds; every time she does, she loses a memory; the city's only other hero thinks she is the villain; 10 episodes, each ends with a memory she did not know she had lost.

## The city and the suit

| Kind of hero story | Scene Style | Costume Style |
| A big city hero | New York Metropolis | Smart Casual by day |
| A neon future | Cyber Metropolis or Neon Nights | Near-Future Techwear, Cyberpunk Augment |
| A hero at school | American High School | Streetwear |
| A family of heroes (animated) | American Small Town | Soft Everyday, with Family 3D Animation visuals |
The suit is a state of the hero: describe its colours, materials and one mark, and keep it original. States · IP check.

## Powers as cause and effect

- Say what happens. “Raindrops stop in mid-air around her; only she moves”, not “amazing power”. Describe the cause.
- One power beat per shot. Build big moments from short cuts. One shot.
- Fights are short shots of one action each, with no injuries in close-up. Fights.
- The villain needs a face as specific as the hero's. Describe a face.

## Three seconds, shot by shot

| Beat | Shot to adapt |
| The crash | [Shot 1] 3s: Wide shot, a rainy avenue at night, a car skidding towards a bus stop. |
| The stop | [Shot 2] 4s: Medium shot. Raindrops freeze in the air; @Nadia, in a paramedic's jacket, is the only thing moving. |
| The pull | [Shot 3] 3s: Medium shot. @Nadia pulls a boy out of the car's path. |
| The cost | [Shot 4] 4s: Close-up on @Nadia. @Nadia (inner voice): What was my brother's name? (voice: Nadia) |

## Story shapes that work

- The power with a price.
- The hero nobody believes.
- The villain's daughter.
- The retired hero called back.
- A family of heroes for kids.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style New York Metropolis, Cyber Metropolis; Costume Style Smart Casual, Near-Future Techwear; Visual Style Digital HD, Game Engine Realism; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the hero with an everyday state and an original suit state; the villain with a specific face; the avenue and the rooftop as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The avenue at night. | Time stops for three seconds. |
| 2 | The ambulance station. | She has forgotten her own birthday. |
| 3 | A rooftop. | The city's hero calls her a villain. |
| 4 | Her flat. | A notebook she started keeping, half full. |
| 5 | The bridge. | She stops time for ten seconds. |

## FAQ

### How do I make a superhero drama with AI?

Give the hero an everyday state and a suit state, choose a city scene style in SceneMixer's Step 1, and describe powers as light and effect, one beat per shot.

### Can I use a famous superhero?

No. Known heroes, logos and suits trigger SceneMixer's IP check. Design your own.

### How do I show powers?

Describe what happens: frozen rain, a glow, a door bending. Avoid words like amazing or epic.

### Can I show fights?

Yes, as short shots of one action each, with no injuries in close-up.

### How do I keep the suit the same?

Make it a state of the hero with its own picture.

### Can it be animated for kids?

Yes. Use Family 3D Animation or 3D Toy Cartoon.

### Which visual style suits live action?

Digital HD or Game Engine Realism.

### Vertical or landscape?

Either; 9:16 Portrait for feeds, 16:9 Landscape for action.

Updated November 3, 2026


---

# AI Survival Drama: Islands, Wilderness and Who Gets Left Behind

URL: https://scenemixer.com/guides/make-an-ai-survival-drama/
Make a survival drama with AI: Tropical Rainforest, Frozen Wilderness and Island Resort worlds, Outdoor Gear that wears out as states, and a group that gets smaller.

# How to make an AI survival drama

Quick answer: A survival drama is a group, a place that wants them dead and the slow question of who leads. In SceneMixer, Step 1 picks the place: Tropical Rainforest (layer on layer of green, vines, waterfall mist), Frozen Wilderness (snow, ice lakes, one cabin), Island Resort for the beach where the plane came down, Wasteland after a collapse. Outdoor Gear dresses them; as clothes tear and dirt sets in, add a worn state per character for the later episodes.

Brief for AI Create: A survival drama: a charter plane goes down on an uninhabited island with eight passengers; the pilot is missing, the radio works but only receives, and one passenger knew the plane would crash; 12 short episodes, each ends with a sound from the jungle.

## Choose the wild

| Survival setting | Scene Style | Costume Style |
| A deserted island | Island Resort or Polynesian Island | Beach Resort turning into rags |
| The jungle | Tropical Rainforest | Outdoor Gear |
| Snow and ice | Frozen Wilderness | Outdoor Gear, Inuit Arctic |
| The desert | Frontier Vastness | Outdoor Gear |
| After the end of the world | Wasteland | Wasteland Scavenge |
Zombie apocalypse dramas.

## Wear and tear as states

Day one and day forty do not look the same. Add a worn state to each survivor (torn sleeves, sunburn, a beard) and switch to it from the episode where time has passed. The camp is a location with a second look as it grows. States.

Weather and time of day are fields on each segment: Storm on the night the shelter fails, Fog on the morning someone is missing. The segment header.

## The night, shot by shot

| Beat | Shot to adapt |
| The fire | [Shot 1] 4s: Wide shot, a beach camp at night, one fire, a shelter of palm leaves. Seven survivors sit close to the flames. |
| The count | [Shot 2] 3s: Close-up on @Jonah counting heads, stopping at seven. |
| The sound | [Shot 3] 3s: Medium shot. The jungle edge in the dark; Ambient: a branch snapping. |
| The line | [Shot 4] 4s: Close-up on @Naomi. @Naomi (very quiet): Where's Mr Kessler? (voice: Naomi) |
Danger stays off-screen: injuries and deaths are not shown, only what is left behind. Content rules.

## Story shapes that work

- The crash.
- The reality show that becomes real.
- The family lost in the woods.
- The expedition with one traitor.
- The last cabin in a winter that does not end.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Tropical Rainforest, Island Resort; Costume Style Outdoor Gear, Beach Resort; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: each survivor with one mark and a worn state; the beach camp with a second look; the jungle and the waterfall as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The plane, then the beach. | Eight survivors, no pilot. |
| 2 | The jungle edge. | Fresh footprints that are not theirs. |
| 3 | The camp. | The radio receives a voice they know. |
| 4 | The waterfall. | One of them had a map of the island. |
| 5 | The camp at night. | Seven around the fire. |

## FAQ

### How do I make a survival drama with AI?

Pick a wild scene style such as Tropical Rainforest or Frozen Wilderness in SceneMixer's Step 1, Outdoor Gear for costumes, and add worn states for the later episodes.

### How do I show the survivors getting worse?

Add a worn state per character (torn clothes, sunburn) and switch to it when time has passed.

### How do I add storms and fog?

Set the weather on each segment's header in SceneMixer: Storm, Fog, Rain, Snow and more.

### Can I show injuries?

No graphic injuries. Show the moment before and what is left after.

### Can the camp change over time?

Yes. Make it a location and add a second look as it grows.

### Which visual style suits survival?

Digital HD or Shot on Film for live action, Game Engine Realism for a gritty look.

### Vertical or landscape?

16:9 Landscape for the scale, 9:16 Portrait for short-drama feeds.

### Can it be a reality-show story?

Yes. Write the cameras and the host into the script as characters and props.

Updated November 3, 2026


---

# AI Telenovela and Soap Opera: How to Make One in Spanish or English

URL: https://scenemixer.com/guides/make-an-ai-telenovela-or-soap-opera/
Secret parentage, twins, a slap and a wedding stopped at the altar: how to write a telenovela, cast it for Latin America and generate it with native Spanish dialogue.

# How to Make an AI Telenovela or Soap Opera

Quick answer: A telenovela is melodrama at full volume: two families, one secret, love across a class line, and emotions played straight to the camera. To make one in SceneMixer, set the dialogue language to Spanish or Portuguese in step 1 — the cast speaks it natively, mouths moving to the words, not dubbed afterwards — choose Latin America as the Target Region so names, faces and streets match, and pick TV Series for longer, slower-burning episodes or Short Drama for a feed version. Write the big emotions into the delivery notes, give the twins two separate character cards, and keep the slap at the moment before contact.

Brief for AI Create: A telenovela: a seamstress in a coastal town learns on her wedding day that the groom's father is her own father; his mother has known for twenty years; she stops the wedding and claims her name; passionate, tearful, triumphant.

## What a telenovela needs

- Two families. One rich, one not, bound by an old wrong. Every subplot is a thread between the two houses.
- The secret. Who someone’s real parent is, who was switched at birth, who caused the accident. It is revealed in stages, to different characters at different times.
- The villainess. Elegant, patient and always one step ahead until the last episodes. The audience loves to hate her; give her the best lines.
- Love across the line. The heir and the maid, the landowner’s son and the fisherman’s daughter.
- The set pieces. The interrupted wedding, the deathbed confession, the slap, the letter found decades late, the return under another name.
- Feelings out loud. Characters say what they feel, at length, in close-up. This is a feature, not a flaw.

## Language and region: the two settings that make it Latin

Two separate rows in step 1 do the localising, and they work independently:

- Dialogue language decides what the cast speaks. Choose Spanish or Portuguese and every line in the storyboard is in that language, performed natively in the video. You can write the script in any language you like; the read produces the dialogue in the one you chose. Dialogue language and cast localisation.
- Target Region decides the world: choose Latin America and character names, faces, streets, signs and props follow. Anything your script states explicitly — a Mexico City address, a character called Ignacio — is kept. Target Region.
The two are set before the read and fixed afterwards. For a second market — the same story in English for the US, or in Turkish as a dizi — start a second project from the same script with a different language and region. Make a second language version.

## The rest of step 1

- Type: TV Series gives longer episodes with room for a scene to build, which suits the genre’s confrontations; Short Drama gives a faster feed version with a hook every episode.
- Aspect Ratio: 16:9 Landscape for a television or YouTube feel; 9:16 Portrait for phones.
- Visual Style: Digital HD or Shot on Film; Soft Glow for a romantic period story.
- Scene Style: Latin Colonial Town for painted streets and courtyards, Suburban Mansion for the rich family’s house, Mediterranean Coast or Island Resort for a coastal story.
- Costume Style: Haute Couture for the villainess and the galas, Latin American Folk for a rural or festival setting, Wedding if the series keeps returning to the altar.

Visual Style: Soft Glow

Scene Style: Latin Colonial Town

Scene Style: Suburban Mansion

Costume Style: Haute Couture

Costume Style: Latin American Folk
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Twins, doubles and the return in disguise

Telenovelas love a face that appears twice. How you set it up decides whether the audience can tell who is who:

- Identical twins are two characters, each with a card, an image and a voice. They can appear in the same frame, so they cannot be two states of one person. Make them distinguishable: one mark, one hairstyle, one way of dressing that the other never shares.
- A character who returns transformed — years later, rich, under another name — is one character with a state for the new look: Add a state on the card, name it Return, describe only the difference (short dark bob, tailored white suit, a scar now faded), and pick it in the episodes after her return. States.
- A character known by two names keeps one card; the other name can be added under Also called so SceneMixer recognises her either way. Also called: nicknames and aliases.
Character sheets from the sample El Secreto de Costa: one reference image per character, used in every shot they appear in.

## Melodrama, written into the shots

The emotion lives in the delivery note in brackets at the start of a line and in the close-up that follows it. Write both.

| Beat | Shot to adapt |
| The confession | [Shot 3] 6s: Medium close-up, slow push in. In frame only: @Doña Elvira, seated by the window. @Doña Elvira (trembling, holding back tears): Lo supe desde el primer día. Ambient: the sea, a church bell far off. |
| The reaction | [Shot 4] 3s: Close-up. In frame only: @Lucía. Her eyes fill; she does not look away. No dialogue. |
| The slap | [Shot 7] 3s: Medium shot. @Lucía and @Isabel face to face in the courtyard; @Lucía’s hand rises. Cut before contact. No dialogue. |
| The aftermath | [Shot 8] 3s: Close-up. In frame only: @Isabel, a hand at her cheek, eyes wide. @Isabel (barely audible): ¿Cómo te atreves? Ambient: a door slams inside the house. |
Cutting just before contact reads as a hit to the viewer and keeps the shot clear of the violence check. Dialogue in the storyboard · How to write one shot.

## Voices that carry the drama

Give each Voice description a temperature. The villainess: velvet-smooth, slow, every word chosen, a smile in the voice. The heroine: warm, quick, breaks when she cries and steadies when she is angry. The patriarch: deep, tired, used to being obeyed. A regional colour — Mexican, Caribbean, Rioplatense — can be written into the description as an accent. Give a character an accent · How character voices work.

## Make it in SceneMixer, start to finish

- Script: paste yours, in any language, on Upload Script, or write a brief in AI Create with the two families, the secret, the lovers and the ending.
- Step 1: dialogue language, Target Region and the rest as above; read the script and check in the summary that the secret is revealed in stages.
- Step 2: generate the cast and the two houses; give twins two cards and a returning character a state.
- Storyboard: episode 1 first (free on your first project); strengthen the delivery notes on the big lines.
- Video: generate on the default tier at 10 credits a second, fix the shots that miss, composite and post.
To hear a cast speaking Spanish natively, watch El Secreto de Costa; for a Turkish family drama built the same way, Çantasındaki Sözleşme.

| Episode | What happens | The hook |
| 1 | The wedding day. A letter in the groom’s father’s desk. | “You cannot marry him. He is your brother.” |
| 2 | She stops the ceremony without saying why; the town turns on her. | The groom’s mother visits her at night — and offers money to leave. |
| 3 | Her mother’s old friend tells half the story. | A photograph: two baby girls, not one. |
| 4 | The villainess moves first and has her fired from the mansion. | Across the plaza, a woman with her face steps out of a car. |
| 5 | The twins meet. | “Then which of us did they bury?” |

## FAQ

### Can the telenovela be in Spanish with real Spanish speech?

Yes. Choose Spanish as the dialogue language in SceneMixer's step 1. The cast speaks it natively in the video, with mouths moving to the words, rather than being dubbed afterwards.

### Do I have to write the script in Spanish?

No. You can write it in any language. The dialogue in the storyboard and the video comes out in the dialogue language you chose.

### How do I make identical twins?

Two characters with two cards, not one character with a state. Twins can share a frame, so they need separate images. Give each one a detail the other never has.

### Can I make a Turkish dizi or a Korean family drama the same way?

Yes. Pick Turkish or Korean as the dialogue language and the matching region, then write the story the same way.

### Should a telenovela be TV Series or Short Drama?

TV Series gives longer episodes with room for confrontations to build. Short Drama gives faster episodes with a hook every time, better for phones.

### Can a slap be shown?

Write the raised hand and cut before contact, then show the aftermath. It reads as a hit and keeps the shot clear of the violence check.

### Can I make the same series in English too?

Start a second project from the same script with English dialogue and the region you want. Each version on SceneMixer is generated and paid for on its own.

### How long should each episode be?

On TV Series, a few minutes; on Short Drama, one to two minutes. The type you pick in step 1 sets the rhythm and the cap.

Updated November 3, 2026


---

# AI Thai Lakorn Drama: Thai Palaces, Revenge and Forbidden Love

URL: https://scenemixer.com/guides/make-an-ai-thai-lakorn-drama/
Make a Thai lakorn-style drama with AI: Thai Royal Palace and Southeast Asian Town scenes, Thai Court costumes, and dialogue in English or 15 other languages.

# How to make an AI Thai lakorn-style drama

Quick answer: A lakorn runs on big feelings: a slap, a forced marriage, a revenge that turns into love. In SceneMixer, Step 1 gives you Thailand: Thai Royal Palace (stacked red-and-green roofs, gilded carvings, golden stupas) with Thai Court costumes for period stories, Southeast Asian Town (shophouses, coloured shutters, palms) for modern ones, and Target Region Southeast Asia for the faces and names. Thai is not one of the dialogue languages, so the lines are spoken in English or another of the available languages.

Brief for AI Create: A lakorn-style revenge drama: a maid in a Bangkok mansion is the daughter of the woman the family ruined 20 years ago; she plans to destroy them through their eldest son and falls for him instead; 12 short episodes, each ends with a secret almost revealed.

## Thailand in Step 1

| Kind of lakorn | Scene Style | Costume Style |
| A period palace story | Thai Royal Palace | Thai Court |
| A modern Bangkok family | Urban Opulence or Neon Nights | Minimal Luxury, Smart Casual |
| A town by the river | Southeast Asian Town | Soft Everyday |
| An island resort romance | Island Resort | Beach Resort |
Target Region Southeast Asia gives Southeast Asian faces and names. Target Region.

## Dialogue language, honestly

The spoken dialogue can be any of 15 languages; Thai is not one of them. Most lakorn-style creators publishing to a global audience use English. Subtitles in your script's words can be added to an episode. Choosing the dialogue language · Subtitles.

## The lakorn moment, shot by shot

| Beat | Shot to adapt |
| The staircase | [Shot 1] 4s: Wide shot, a Bangkok mansion hall, marble stairs. @Pim carries a tray up the stairs. |
| The collision | [Shot 2] 3s: Medium shot. @Krit comes round the corner and the tray tips. |
| The look | [Shot 3] 3s: Close-up on @Krit catching her wrist. |
| The line | [Shot 4] 4s: Close-up on @Khun Ying Duang at the top of the stairs. @Khun Ying Duang (cold): Let go of the maid, son. (voice: Duang) |
Slaps and fights: show the moment before and after, one action per shot. Fights.

## Story shapes that work

- Revenge through romance.
- The forced marriage.
- Two lives: a past life in the palace, a present life in Bangkok.
- The ghost bride: a supernatural lakorn.
- Two sisters, one man.
Revenge dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Thai Royal Palace, Southeast Asian Town; Costume Style Thai Court, Minimal Luxury; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the family and the maid with one mark each; Target Region Southeast Asia; the mansion's hall, kitchen and the river as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The mansion's kitchen. | She is hired under a false name. |
| 2 | The staircase. | The eldest son catches her wrist. |
| 3 | The family dinner. | His mother recognises her mother's necklace. |
| 4 | The river at night. | He follows her to her mother's old house. |
| 5 | The wedding announcement. | He is engaged to someone else. |

## FAQ

### How do I make a Thai lakorn-style drama with AI?

Choose Thai Royal Palace or Southeast Asian Town in SceneMixer's Step 1, Thai Court or modern costumes, Target Region Southeast Asia, and write the big emotional beats as short shots.

### Can the characters speak Thai?

No. Thai is not one of the dialogue languages on SceneMixer. Use English or another available language, and add subtitles if you need them.

### What does the Thai Royal Palace style look like?

Thai royal palaces and temples: stacked red-and-green tiled roofs with upturned ends, gilded eave carvings and golden bell-shaped stupas.

### How do I get Thai faces and names?

Set Target Region to Southeast Asia in Step 1.

### Can I show slaps and fights?

Yes, as short shots of the moment before and after; graphic violence is blocked.

### Can I copy a famous lakorn?

No. Known shows and actors trigger the IP check. Write your own story.

### Vertical or landscape?

9:16 Portrait for short-drama feeds, 16:9 Landscape for a series.

### Which visual style?

Digital HD or Soft Glow for live action; Shot on Film for a period piece.

Updated November 3, 2026


---

# AI Time Travel Romance: Two Eras, One Couple, a Clock Running Out

URL: https://scenemixer.com/guides/make-an-ai-time-travel-romance/
Make a time travel romance with AI: two eras in one project, the traveller's clothes as states, a modern scene style with a period location, and one rule.

# How to make an AI time travel romance

Quick answer: A time travel romance needs two eras, one rule and a reason not to go home. A SceneMixer project has one scene style and one costume style, so choose the era where most of the story happens and build the other one through its locations and a state per character: the traveller in modern clothes and in period dress is one person with two looks. Write the rule into the script early (a full moon, a locket, seven days) and make the object that moves people through time a prop with its own picture.

Brief for AI Create: A time travel romance: a museum restorer touches a Regency-era pocket watch and wakes in 1815 in the house of the duke who owned it; she has seven days before the watch can take her back, and his family thinks she is the bride they ordered from France; 10 episodes, each ends with the watch ticking louder.

## One project, two eras

| Where most of the story happens | Scene Style | Costume Style | The other era |
| Regency England | English Manor | Regency | The museum as a location; her modern clothes as a state |
| A Chinese palace | Imperial Splendor | Qing Palace Robes | A city apartment as a location |
| Today, with a visitor from the past | New York Metropolis | Smart Casual | His period dress as a state |
| Joseon Korea | Joseon Korea | Hanbok | A Seoul cafe as a location |
The styles apply when the script is read. Why they are fixed afterwards · States.

## The rule and the object

- The object (a watch, a mirror, a hairpin) is a prop with a reference picture, so it is the same in both eras. Props.
- The rule is said out loud once, early, and every episode brings the deadline closer. Hooks.
- Manners clash is the comedy: she says “okay” to a duke. Write the line exactly as it should land. Dialogue.

## The arrival, shot by shot

| Beat | Shot to adapt |
| The watch | [Shot 1] 3s: Close-up on a gold pocket watch in @Claire's gloved hand, the second hand starting to move. |
| The room | [Shot 2] 4s: Wide shot, a Regency drawing room in candlelight. @Claire stands in jeans and a museum lanyard. |
| The duke | [Shot 3] 3s: Medium shot. @Duke Ashford turns from the fireplace. |
| The line | [Shot 4] 4s: Close-up on @Duke Ashford. @Duke Ashford (dry): You are late, and you are wearing trousers. (voice: Ashford) |

## Story shapes that work

- Into the past: a modern woman in a period house.
- Out of the past: a knight in a modern city.
- The loop: the same day until she fixes one thing.
- Letters across time: two people in two eras, one desk.
- The choice: go home, or stay.
Regency romance · Period romance.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style English Manor; Costume Style Regency; Visual Style Soft Glow, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the traveller with a modern state and a period state; the time object as a prop; the museum and the drawing room as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The museum. | The watch starts ticking. |
| 2 | The duke's drawing room. | They think she is the French bride. |
| 3 | The ball. | She dances a modern step by mistake. |
| 4 | The library. | A portrait of herself, dated 1816. |
| 5 | The garden at night. | Three days left, and he asks her to stay. |

## FAQ

### How do I make a time travel romance with AI?

Pick the SceneMixer scene and costume style of the era where most of the story happens, give the traveller a second state for the other era, and make the time object a prop.

### Can one project have two eras?

One scene style and one costume style per project on SceneMixer; build the second era through its locations and character states.

### How do I keep the time-travel object the same?

Make it a prop with a reference picture in Step 2.

### Can a knight come to the modern day?

Yes. Set the project in the modern day and give him a period-armour state.

### Which visual style suits it?

Soft Glow for romance, Shot on Film for a period feel, Digital HD for modern scenes.

### Vertical or landscape?

9:16 Portrait for short-drama feeds, 16:9 Landscape for a period series.

### Can I copy a famous time travel story?

Write your own. Known characters and titles trigger the IP check.

### How do I show the jump in time?

A short shot of the object, a cut, and the new room; let the costume tell the audience where they are.

Updated November 3, 2026


---

# AI True Crime Style Video: An Invented Case, Told Like the Real Thing

URL: https://scenemixer.com/guides/make-an-ai-true-crime-style-video/
Make a true crime style video with AI: a narrator, reconstructions, interviews to camera and evidence close-ups around an invented case, and why real cases differ.

# How to make an AI true crime style video

Quick answer: True crime videos are made of four things: a narrator, reconstructions of the night, people talking to camera, and close-ups of the evidence. SceneMixer can make all four around a case you invent: the narrator is a voice-only character, witnesses are characters framed to camera, the evidence is props, and the lab is Forensic Cold White. Real cases are about real people, whose faces cannot be imitated and who must not be defamed or exposed, so the format works best as fiction.

Brief for AI Create: A true crime style mini-series about an invented case: in 1994 a lighthouse keeper on a small island disappears the same night the ferry stops running; three witnesses, three versions; six episodes with a narrator.

## The four ingredients

| Ingredient | How to make it |
| Narrator | A voice-only character; voice-over lines over the other shots |
| Reconstruction | Ordinary scenes, often without dialogue, in the period's look |
| Interviews | A witness seated, framed medium close-up, looking just off camera |
| Evidence | Props: the torch, the logbook, the ferry ticket, each in close-up |

Voice-only characters · Interviews to camera.

## Three versions of one night

The genre's engine is contradiction. Draw the location once (the lighthouse, the quay) and stage the same night three times, once per witness, with one detail different each time: the lamp on or off, the boat at the quay or gone. Because the set does not change, the audience can spot the difference. The set.

## Look and places

Visual Style: Shot on Film

Visual Style: Black & White Film

Visual Style: Digital HD

Scene Style: American Small Town

Scene Style: Frozen Wilderness
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

Shot on Film for reconstructions of the past, Digital HD for present-day interviews. A project has one visual style, so a mixed look means two projects. Captions with dates and names are added in your editor: generated video does not spell reliably. Text on screen.

## Real cases

A real case has real victims, families and suspects. Imitating their faces or voices is not allowed, and defaming or exposing real people is against the content rules. Invent the case, the place and the names; the format is what viewers come for. Content rules. Bodies and wounds are graphic violence and get blocked; the genre lives in the empty room afterwards. For a crime story told as drama, see police and crime or murder mystery.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style American Small Town, Forensic Cold White; Costume Style Outdoor Gear, Professional Uniforms; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the narrator as a voice-only character; three witnesses; the main location drawn once; the evidence as props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Part |
| 1 | The night: what everyone agrees on |
| 2 | Witness one |
| 3 | Witness two |
| 4 | The evidence nobody looked at |
| 5 | Witness three |
| 6 | What the logbook says |

## FAQ

### How do I make a true crime video with AI?

Use a narrator as a voice-only character, reconstruction scenes, witnesses framed to camera and evidence as props, around a case you invent.

### Can I make a video about a real case?

Real people's faces and voices cannot be imitated and real people must not be defamed or exposed; invent the case.

### Can I show the crime?

Graphic violence is blocked on SceneMixer; show the place before and the empty room after.

### How do I add dates and names on screen?

In your editor after export; generated video does not spell reliably.

### Which look suits it?

Shot on Film for the past, Digital HD for interviews; one style per project.

### Vertical or landscape?

16:9 Landscape for YouTube, 9:16 Portrait for feeds.

### How long is an episode?

Two to five minutes, built from segments and joined.

Updated November 3, 2026


---

# AI Turkish Dizi Drama: Istanbul Mansions, Family Pride and Long Looks

URL: https://scenemixer.com/guides/make-an-ai-turkish-dizi-drama/
Make a Turkish dizi-style drama with AI: Istanbul mansions and the Bosphorus, Middle Eastern Souq and Mediterranean Coast scenes, and dialogue in Turkish.

# How to make an AI Turkish dizi-style drama

Quick answer: A dizi is a family, a mansion on the water and a look across a room that lasts three seconds too long. In SceneMixer, set Istanbul in Step 1: Urban Opulence for the mansion, Middle Eastern Souq for the old-city lanes, Mediterranean Coast for the summer house, Ottoman Palace for a period story. Turkish is one of the dialogue languages, so the lines can be spoken in Turkish; write the names into the script and keep Target Region on Follow script.

Brief for AI Create: A dizi-style drama: a seamstress from an old Istanbul neighbourhood is hired to make the wedding dress for the daughter of the family that ruined her father; the groom is the man she loved at sixteen; 12 short episodes, each ends with a look across a room.

## Istanbul in Step 1

| Kind of dizi | Scene Style | Costume Style |
| A rich family on the Bosphorus | Urban Opulence | Minimal Luxury, Haute Couture |
| An old neighbourhood | Middle Eastern Souq | Soft Everyday |
| A summer by the sea | Mediterranean Coast | Beach Resort |
| An Ottoman period story | Ottoman Palace | Ottoman Court |
| A family company | Corporate Cold | Power Tailoring |
Ottoman palace dramas.

## Turkish dialogue and the long look

Choose Turkish as the dialogue language in Step 1, before the script is read; it cannot change afterwards. For an English version, make a second project from the same script. Dialogue language · A second-language version.

Dizi pacing is slow on purpose: give the look its own shot of three or four seconds with no dialogue. Shot timing.

## The fitting, shot by shot

| Beat | Shot to adapt |
| The room | [Shot 1] 4s: Wide shot, a mansion dressing room with tall windows onto the Bosphorus. @Zeynep pins a white dress on @Elif. |
| The door | [Shot 2] 3s: Medium shot. @Kerem stops in the doorway. |
| The look | [Shot 3] 4s: Close-up on @Zeynep, a pin between her lips, not moving. No dialogue. |
| The line | [Shot 4] 4s: Close-up on @Elif. @Elif (bright, unaware): Kerem, meet my seamstress. (voice: Elif) |

## Story shapes that work

- The family that ruined mine.
- The forced marriage between two feuding families.
- The girl from the neighbourhood and the heir.
- The return after ten years abroad.
- The secret child.
Revenge dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Middle Eastern Souq; Costume Style Minimal Luxury, Soft Everyday; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: both families with one mark each; Turkish dialogue chosen in Step 1; the mansion, the shop and the old street as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The tailor's shop. | An order from the family she hates. |
| 2 | The mansion. | The groom is the boy she loved. |
| 3 | The fitting. | A look nobody else sees, except the groom's mother. |
| 4 | The old neighbourhood. | Her father's ledger names the groom's father. |
| 5 | The engagement party. | She wears a dress she made for herself. |

## FAQ

### How do I make a Turkish dizi-style drama with AI?

Choose Urban Opulence, Middle Eastern Souq or Mediterranean Coast in SceneMixer's Step 1, Turkish as the dialogue language, and give long looks their own shots.

### Can the characters speak Turkish?

Yes. Turkish is one of the dialogue languages on SceneMixer; choose it in Step 1.

### Can I make an English version too?

Yes, as a second project from the same script with English dialogue.

### How do I get Turkish names and faces?

Write the names into the script and set Target Region to Follow script.

### Which visual style suits a dizi?

Digital HD or Soft Glow; Shot on Film for a period story.

### Can I copy a famous dizi?

No. Known shows and actors trigger the IP check.

### Vertical or landscape?

9:16 Portrait for short-drama feeds; 16:9 Landscape for long-form episodes.

### Can it be an Ottoman period drama?

Yes. Use Ottoman Palace and Ottoman Court.

Updated November 3, 2026


---

# AI TV Pilot and Pitch Trailer: Show Your Series Before It Exists

URL: https://scenemixer.com/guides/make-an-ai-tv-pilot-or-pitch-trailer/
Make a TV pilot or a pitch trailer with AI: TV Series type, a full pilot from your script, a 60-second trailer with the same cast, and a cover poster.

# How to make an AI TV pilot or pitch trailer

Quick answer: A pilot shows a buyer, a producer or an audience that the show works; a pitch trailer shows it in a minute. In SceneMixer, set Type to TV Series and upload the pilot script: the cast is drawn once, the episode is storyboarded and rendered shot by shot, and you join it into one video. For a trailer, write a second short script of the best moments and render it with the same characters. Use it as previs, not as a finished broadcast.

Brief for AI Create: A pitch trailer for a TV series: a small-town coroner who can hear the last ten seconds of a dead person's life; sixty seconds, five moments, the title card at the end.

## Pilot or trailer

| You need | Make | How |
| To prove the story works | The pilot episode | Upload the pilot script, TV Series type, render the episode. |
| To sell the idea in a minute | A pitch trailer | A short script of five to eight moments with the same cast. Make an AI trailer. |
| To show the look | A mood reel | Three or four key scenes. Previs and animatics. |

## Make it look like a show

- One cast across pilot and trailer: reuse characters in a second project with Pick existing. Reuse characters.
- A considered look chosen once. Visual styles.
- A poster for the deck: the promo cover. Promo cover.
- Higher resolution for the best clips. Enhance.

## The trailer, shot by shot

| Beat | Shot to adapt |
| The town | [Shot 1] 3s: Wide shot, a small town at dawn under fog. |
| The lab | [Shot 2] 3s: Medium shot. @Dr Ellis pulls off her gloves in a cold morgue lit by tube lights. |
| The gift | [Shot 3] 4s: Close-up on @Dr Ellis closing her eyes. Ambient: a heartbeat, then a whisper. |
| The line | [Shot 4] 4s: Close-up on @Sheriff Dunn. @Sheriff Dunn (uneasy): How could you possibly know that? (voice: Dunn) |
| The title | [Shot 5] 3s: Wide shot, the town at night, every porch light off but one. |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style American Small Town, Hospital Ward; Costume Style Professional Uniforms, Smart Casual; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the main cast with specific faces; key locations; reuse them in the trailer project with Pick existing. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Moment | What it sells | Length |
| The world | Where the show lives. | 3 s |
| The hero | Who we follow. | 3 s |
| The hook | What makes it different. | 4 s |
| The conflict | Who stands in the way. | 4 s |
| The title | The promise. | 3 s |

## FAQ

### How do I make a TV pilot with AI?

Set Type to TV Series in SceneMixer, upload the pilot script, draw the cast once, and render and join the episode.

### How do I make a pitch trailer?

Write a short script of five to eight key moments with the same characters and render it.

### Can the trailer use the same cast as the pilot?

Yes. Bring the characters into the second project with Pick existing.

### Is this broadcast quality?

Use it as a pilot, previs or pitch; Enhance on SceneMixer raises clips up to 2K.

### Can I get a poster?

Yes, from the promo cover feature.

### Can I use famous actors?

No. Real people cannot be depicted.

### Vertical or landscape?

16:9 Landscape for a TV pitch.

### How long should a pitch trailer be?

About a minute.

Updated November 3, 2026


---

# AI Vampire Romance Drama: Gothic Looks, Fangs and No Gore

URL: https://scenemixer.com/guides/make-an-ai-vampire-romance-drama/
An immortal, a mortal and a rule they cannot break: how to write a vampire short drama, design the gothic look, cast the fangs as a state and keep blood off screen.

# How to Make an AI Vampire Romance Drama

Quick answer: A vampire romance is about restraint: someone who could take everything chooses not to, and the audience watches the effort it costs. In SceneMixer you build it with a gothic look chosen in step 1 (Gothic Dark sets, Gothic Court or Victorian wardrobe, Digital HD or Soft Glow), the vampire’s hunger as a state of the character (red eyes, fangs) so it appears exactly when you choose, and a firm rule about blood: you show the want and the aftermath, never the wound. Dark, candle-lit scenes come out best when you write the light source into the shot.

Brief for AI Create: A vampire romance: a night-shift nurse saves a stranger who has not aged since 1890; his clan will kill her for knowing; he must choose between his kind and her before the blood moon; gothic, slow-burn, aching.

## What the genre runs on

- The rule. Every vampire story has one: no feeding on the living, no turning without the elders’ consent, no being seen. The romance is a slow breaking of the rule.
- Old world, new city. An immortal who remembers a century of balls and duels, living in a city of neon and night buses. The contrast is half the look.
- The hunger. The single scene the audience waits for is the vampire almost losing control. Delay it, then deliver it as a look, not a bite.
- The clan. Elders, a jealous sire, a younger vampire who wants the lead’s place. They make the rule enforceable.
- The cost. Immortality as loneliness, turning as a loss. The best episode endings are choices that cannot be undone.

## Settings to pick in step 1

- Visual Style: Digital HD for a modern-city vampire, Soft Glow for a romantic, candle-lit period feel, Black & White Film for a deliberately classic horror look, Painterly Anime or Korean Webcomic for an anime audience.
- Scene Style: Gothic Dark for crypts, old churches and velvet clubs; European Palace or Medieval Castle for the clan’s seat; Neon Nights if the story is mostly modern streets after dark.
- Costume Style: Gothic Court or Victorian for the old vampires, Rock & Punk for a younger, modern clan. A vampire who dresses in two centuries at once can carry that in their own character description.
- Type and Aspect Ratio: Short Drama and 9:16 Portrait for a feed series; Film and 16:9 Landscape for a single gothic short.

Visual Style: Soft Glow

Visual Style: Black & White Film

Scene Style: Gothic Dark

Scene Style: European Palace

Costume Style: Gothic Court

Costume Style: Victorian
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## The hunger as a state

Red eyes and fangs change the face, and a reference image locks the face, so text alone will not reliably add them. Make the vampire’s hungry look a state:

- Generate the vampire’s main image as their composed, human-passing self.
- On their card choose Add a state, name it Hunger, and write only the difference: irises deep red, pupils narrowed, upper canines lengthened into fangs, skin a shade paler.
- Press Generate this state. Use it only in the shots where control slips: in the episode editor, select that segment, open the vampire from the episode’s assets and choose Use this state in this shot; every other shot keeps the main image.
Because the state is chosen per segment, the transformation becomes something you place, not something that leaks into every scene. Use a character state in one scene · State image messages.

Character sheets from the sample Crimson Vow: one reference image per character, used in every shot they appear in.

## Blood stays off screen

Every generation passes the provider’s content check, and gore is blocked. A vampire series that depends on visible blood will fight the check every segment. The genre does not need it; the best of it never showed much. Work with implication:

| Instead of | Write |
| A bite on the neck | He leans toward her throat; cut to her hand closing on his sleeve. Next shot: she is sitting alone on the stairs, a scarf at her neck. |
| Blood on his mouth | He turns away from her, jaw tight, and wipes his mouth with the back of his hand. |
| A feeding scene | A dark wine glass on the table, still full. He has not touched it in three days. |
| A vampire killed | A burst of dawn light through the stained glass; the next shot, an empty coat on the floor. |
If a segment is blocked, change its storyboard text before you retry. Blocked by moderation: what to change.

## Darkness, candles and moonlight

A video model needs light to draw by. “A dark room” produces a dim, muddy frame; “a single candle on the table throws warm light up across his face, the rest of the crypt falls away to black” produces a gothic one. Name the source every time: candles, a fireplace, moonlight through a high window, a red neon sign outside, a phone screen held under a chin. Night, fog and rain are written into the shot, not made into separate images. Describe the cause, not the result.

## Three shots to copy

| Beat | Shot |
| The first look | [Shot 1] 4s: Opening: medium close-up through the rain on the clinic window. In frame only: @Lucian. He watches the door; a streetlight throws cold white light across one side of his face. No dialogue. Ambient: rain, a distant siren. |
| The slip | [Shot 6] 3s: Close-up. In frame only: @Lucian. A drop of rain runs down her wrist in the foreground; his eyes follow it, and he goes very still. No dialogue. |
| The restraint | [Shot 7] 5s: Medium shot, candlelight. @Lucian steps back from @Mara until his shoulders meet the wall. @Lucian (low, strained): Leave. Now. Before I stop asking. @Mara lips closed, no lip movement. |
Shot 6 is the one to point at the Hunger state. How to write one shot.

## Voices for something very old

An immortal should sound like one. Write it into the Voice description: low, unhurried, very precise, an old-fashioned formality, never raises his voice. Give the mortal lead the opposite — quick, warm, interrupting herself — and the gap does half the characterisation. A vampire who has lived in many countries can carry a faint accent the same way. Give a character an accent · How character voices work.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or let AI Create draft one from a brief with the rule, the two leads, the threat and the ending.
- Step 1: the gothic settings above, then read the script.
- Step 2: generate the cast, the club, the crypt and the clan seat; add the Hunger state; set any narrator to Voice only (off-screen).
- Storyboard: episode 1 first (free on your first project). Rewrite any line that puts blood or a bite in the frame.
- Video: generate on the default tier at 10 credits a second; regenerate the shots that miss; composite and post.
To see a rain-soaked city that turns gothic, watch Crimson Vow: an alley behind a velvet-lined club, steam from the grates, and a bloodline mystery under the surface.

| Episode | What happens | How it ends |
| 1 | A nurse treats a stranger with no pulse who walks out of the ward. | She finds the dressing he left behind: dry, unstained. |
| 2 | He comes back to warn her that she has been seen. | His sire is waiting in her flat. |
| 3 | The clan gives him a night to silence her. | He tells her the truth instead. |
| 4 | She hides in his world; a younger vampire wants her gone. | Her blood type is the one the elders have been searching for. |
| 5 | The hunger: he nearly loses control. | He leaves the city rather than hurt her — and she follows. |

## FAQ

### Can a vampire drama show blood?

Not reliably. Gore is blocked by SceneMixer's content check. Show the moment before and the aftermath, a dark wine glass, a scarf at the neck; the genre works better that way anyway.

### How do I give my vampire red eyes and fangs only in some scenes?

Make the look a state of the character with its own image, then pick it in the segments where you want it. Everywhere else the main image is used.

### Which visual style suits a vampire story?

Digital HD for a modern city, Soft Glow for a candle-lit romance, Black & White Film for a classic horror feel, or an anime preset for an anime audience. The style is fixed after step 1.

### Why do my night scenes come out muddy?

Because nothing in the shot is lighting them. Name a source: a candle, a fireplace, moonlight through a window, a neon sign outside. The model lights the scene from what you name.

### Can I adapt a famous vampire novel or film?

Not without the rights. Known characters and franchises trigger the IP check. Write your own clan, names and rules.

### Can a vampire be one of several supernatural characters?

Yes. A werewolf, a witch and a vampire are each characters with their own images, states and voices. Name each with @ in the shots they appear in.

### Should it be vertical or horizontal?

Vertical, 9:16, for a feed series; horizontal, 16:9, for a single short film. Pick in SceneMixer's step 1; it cannot be changed later in that project.

### How much does one episode cost?

The same as any episode on SceneMixer: the script read, setup images, the storyboard and then video by the second at the tier you pick.

Updated November 3, 2026


---

# AI War God Returns Drama: The Humiliated Hero's Big Reveal

URL: https://scenemixer.com/guides/make-an-ai-war-god-returns-drama/
Make a war god returns drama with AI: the humble state and the revealed state of one hero, proof props, a family that mocks him, and a reveal timed for the cliffhanger.

# How to make an AI war god returns drama

Quick answer: The war god returns story is a reveal story: a man everyone looks down on is secretly the most powerful person in the room. In SceneMixer, give the hero two states in Step 2, the humble one (a worn jacket, a delivery uniform) and the revealed one (a dress uniform, a tailored suit), make the proof a prop (a medal, a token, a phone call), and save the switch for the end of an episode.

Brief for AI Create: A war god returns drama: a general who disappeared for seven years comes home as a quiet driver to protect his wife's family, who treat him as a servant; twenty short episodes, a reveal every fourth one.

## Two states, one man

The whole genre is the gap between how people treat him and who he is. Draw both: the humble state used in most segments, and the revealed state saved for the payoff. Keep the face identical; only clothes, posture and light change. States.

| Setting | Scene Style | Costume Style |
| The family's mansion | Urban Opulence | Minimal Luxury for the family |
| The company | Corporate Cold | Power Tailoring |
| The banquet | Urban Opulence ballroom | Haute Couture |
| His past | Military Base | Modern Military (invented insignia) |

## Proof as a prop

The reveal needs something everyone can see: a medal case, a black card, a token, a line of soldiers saluting at the gate. Make it a prop so it looks the same each time it appears, and plant it early. Props.

## The reveal, shot by shot

| Beat | Shot to adapt |
| The insult | [Shot 1] 4s: Medium shot, a banquet hall. @Mother-in-law points at @Chen in his driver's jacket. @Mother-in-law: Wait outside with the cars. |
| The silence | [Shot 2] 2s: Close-up on @Chen, calm. |
| The arrival | [Shot 3] 3s: Wide shot. The doors open; two officers in dress uniform stop in front of @Chen and salute. |
| The reaction | [Shot 4] 3s: Close-up on @Mother-in-law, the glass slipping in her hand. |
The reaction shot is the payoff; give it its own seconds. Hooks and cliffhangers.

## Keeping it inside the rules

Invent the army, the ranks and the insignia; real armed forces and real officials are best left out. Fights are short, one move per shot, with no wounds. Content rules · ReelShort-style dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Corporate Cold; Costume Style Power Tailoring, Minimal Luxury; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: a humble and a revealed state for the hero; the medal case or token as a prop; the mansion and the banquet hall as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Humiliation | Partial reveal |
| 1 | He serves tea at the family dinner | A soldier recognises him in the street |
| 4 | He is fired as driver | A call from a number no one can trace |
| 8 | His wife's company is about to fall | The investor bows to him |
| 12 | The family banquet | The salute |
| 20 | The final enemy | His name and rank, out loud |

## FAQ

### How do I make a war god returns drama with AI?

Give the hero a humble state and a revealed state, make the proof a prop, and save the reveal for the end of an episode.

### How do I show the transformation?

Add a revealed state to the same character in SceneMixer's Step 2 and pick it in the payoff segments.

### Can I use a real army?

Invent the army, ranks and insignia; leave real forces and officials out.

### Which format suits it?

Short Drama, 9:16 Portrait, episodes of one to two minutes.

### Which style suits it?

Digital HD; the genre plays as modern and glossy.

### How often should the reveal happen?

A small reveal every few episodes and a big one at the midpoint and the end.

### Can there be fights?

Short ones, one move per shot, without wounds.

### Can the dialogue be in English?

Yes, or any of the dialogue languages; set the region to match.

Updated November 3, 2026


---

# AI Watercolor Storybook Video: Make a Picture Book That Moves

URL: https://scenemixer.com/guides/make-an-ai-watercolor-storybook-video/
Make a watercolor storybook video with AI: the Watercolor Storybook look, a story map of spreads and shots, a read-aloud narrator and a worked storyboard.

# How to make an AI watercolor storybook video

Quick answer: A watercolor storybook video is a picture book that moves: transparent washes, pencil lines and one gentle story read aloud. In SceneMixer, choose Watercolor Storybook as the Visual Style in Step 1 (transparent watercolor and pencil, like a children's book), write the story as spreads of one to three shots each, give the narrator an Off-screen voice, and draw the cast once so every spread shows the same child, animal or toy. The look is fixed once you continue, so choose it before the script is read.

Brief for AI Create: A watercolor storybook for ages 3 to 6: a girl called Pim folds a paper boat after the rain and follows it down the lane, under the bridge and over the weir to a pond where the frogs are holding a party; a warm, slow narrator, twelve spreads, one soft line at the end.

## Which watercolor look to choose

Two of the 21 visual styles are watercolor and a third is painterly. They are not the same, and the choice is fixed after Step 1:

| Visual Style | What the card says | Choose it for |
| Watercolor Storybook | Transparent watercolor and pencil, like a children's book | Bedtime stories, fables, ages 2 to 7, anything that should feel printed on paper |
| Watercolor Anime | Watercolor-on-paper look, fine lines, natural colors | Quiet slice-of-life, gentle fantasy, older children and adults |
| Painterly Anime | Lavish, saturated scenery; simple-lined characters | Big skies and scenery, a more emotional film |

Visual Style: Watercolor Anime

Visual Style: Painterly Anime
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.
All 21 visual styles · Choosing the settings.

## Think in spreads, not scenes

A printed picture book is read in spreads: two facing pages, one picture, a line or two of text, then a page turn. The same grammar works as video, and it keeps the cost down because each spread is one place.

| In the book | In the video | How to write it |
| A spread | One segment of up to 15 or 30 seconds in one place | Give the segment a single location |
| The picture | One to three shots | One subject per shot; hold the first wide shot for 4 to 5 seconds |
| The text on the page | A narrator line | One short sentence per shot, spoken by an Off-screen narrator |
| The page turn | The cut to the next segment | End a spread on an image, not on a line |
| The last page | The first picture again, changed | Return to the doorstep, now at night |

## A twelve-spread story map

Twelve spreads of 15 seconds is a three-minute film. This is the map for the brief above: one place per spread, one narrator line each.

| Spread | Place | The picture | Narrator line |
| 1 | The doorstep | The rain stops; Pim folds a yellow paper boat | When the rain stopped, Pim folded a boat. |
| 2 | The gutter | The boat is set on a thin stream of rainwater | It was the smallest boat in the world. |
| 3 | The lane | The boat drifts past a blue door and a sleeping cat | It sailed past the blue door and the cat. |
| 4 | The bakery window | Steam on the glass, a baker asleep on a flour sack | The baker was dreaming of bread. |
| 5 | The stone bridge | The boat slips into the dark under the arch | Under the bridge, everything echoed. |
| 6 | Under the bridge | Reflections of the sky glitter on the water | Something glittered. It was only the puddle, wearing the sky. |
| 7 | The weir | A small waterfall; the boat tips at the edge | At the weir, the boat took a deep breath. |
| 8 | Below the weir | The boat lands upright in white foam | And landed on its feet. |
| 9 | The meadow stream | Lantern-lit lily pads, frogs in their best | The frogs were having a party. |
| 10 | The pond | The frogs gather around the boat | They had been waiting for a boat all night. |
| 11 | The pond | The smallest frog climbs in; the others cheer | The smallest frog became the captain. |
| 12 | Pim's window | Pim, back home, sees the pond glimmer far away | And Pim, who had folded it, fell asleep smiling. |

## The first spread, shot by shot

Narrator lines are written into the shot as voice-over, and the shot is given enough seconds to say them: SceneMixer plans English speech at about three words a second plus a pause at each sentence end, so a shot with a line cannot be much shorter than the line. Voice-over lines · How to write one shot.

| Beat | Shot |
| The lane | [Shot 1] 5s: Wide shot, slow push-in. A wet cobbled lane in soft watercolor washes, a blue front door, puddles shining under a pale sky. @Narrator (voice-over, warm and slow): When the rain stopped, Pim folded a boat. Ambient: dripping eaves, a distant bird. |
| The folding | [Shot 2] 3s: Close-up. @Pim's small hands crease a yellow sheet of paper into a boat. No dialogue. Ambient: soft paper folds. |
| The boat | [Shot 3] 4s: Medium shot, low angle. The paper boat on a thin stream of rainwater in the gutter, a loose pencil outline around it, light glinting. @Narrator (voice-over): It was the smallest boat in the world. |
| The goodbye | [Shot 4] 3s: Wide shot. The boat drifts away down the lane, @Pim kneeling with her chin on her hands. Ambient: water trickling. |
That is 15 seconds, four shots and two narrator lines.

## Write the picture, not the brush

- Name the light. Low sun after rain, lamplight on wet stone, a pale morning. Light is what a painted scene is made of.
- Give each place two or three colors. Put them in the location's description: grey-blue stone, one yellow door, green moss.
- One subject per shot. A picture book frame has one thing to look at.
- Hold the shots. 4 to 6 seconds, slow push-ins, no whip pans; the look suits stillness.
- Do not repeat the style in every shot. SceneMixer sets the look once for the whole project, so spend the words on light, place and action. Why magic words do not help.

## Voice, music and subtitles

The narrator is an Off-screen character: it gets a voice and no drawing. Play the voices in the library before you choose one and keep it for every episode. How character voices work. The built-in score is mixed in under the voices and a Music switch turns it off. Background music. Accounts that have made a purchase can add read-along subtitles or an SRT file. Add subtitles.

Picture book to video · AI kids story video generator · A bedtime story series · Fairy tale videos.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing; Costume Style Soft Everyday; Visual Style Watercolor Storybook; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw Pim, the baker and the frogs as the cast and the lane, the bridge and the pond as locations; the narrator is an Off-screen character with a voice and no drawing. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Length | Spreads of 15 s | Spreads of 30 s | Usually |
| 1 minute | 4 | 2 | A single rhyme or a bedtime moment |
| 2 minutes | 8 | 4 | A short picture book |
| 3 minutes | 12 | 6 | A full picture book |
| 5 minutes | 20 | 10 | A longer story or a chapter |

## FAQ

### How do I make a watercolor storybook video with AI?

Choose Watercolor Storybook as the Visual Style in SceneMixer's Step 1, write the story as spreads of one to three shots, add an off-screen narrator, draw the cast once, then render and join the segments.

### Which style is the picture-book look?

Watercolor Storybook on SceneMixer is transparent watercolor and pencil, like a children's book. Watercolor Anime is a finer-lined, more natural-colored watercolor look for older audiences.

### Can I change the look after I start?

No. The visual style on SceneMixer is fixed once you continue past Step 1. Start a new project to try another look.

### Can it read the story aloud?

Yes. The narrator is an off-screen character with a voice you pick from SceneMixer’s library; its lines are spoken in the video.

### How long should a watercolor storybook video be?

One to five minutes. A three-minute film is twelve spreads of 15 seconds.

### Can I turn my own picture book into a video?

Paste the text as the script. Members can upload their own illustrations to SceneMixer as reference sheets so the characters match the book.

### Can I add subtitles for early readers?

Yes, on SceneMixer accounts that have made at least one purchase: burned-in captions or an SRT file.

### How much does it cost?

Video is 10 credits a second on SceneMixer's default tier plus 1 credit a second to join. The first 15 setup images, the first storyboard on your first drama and a 5-second preview are free.

Updated November 3, 2026


---

# AI Werewolf Romance Drama: How to Make One, Step by Step

URL: https://scenemixer.com/guides/make-an-ai-werewolf-romance-drama/
Alpha, fated mate, rival pack: how to write a werewolf short drama, cast the wolf form as a character state, film the shift and keep pack fights clear of moderation.

# How to Make an AI Werewolf Romance Drama

Quick answer: A werewolf romance runs on three promises: a heroine who is rejected or underestimated, an alpha bound to her by fate, and a transformation the audience is waiting for. In SceneMixer you write the story with those beats (or let AI Create draft it from one sentence), pick Short Drama and 9:16 Portrait in step 1 if it is meant for phones, and give every shifter two looks: the human main image, and the wolf added as a state of the same character, so the wolf in episode 9 is the wolf from episode 1. Film the change across two shots instead of one, keep fights to teeth, eyes and a lunge rather than wounds, and the series stays consistent and clear of the content check.

Paste this into AI Create and press Start Writing: A werewolf romance: a waitress rejected by the alpha heir at the full-moon ceremony learns she is the last white wolf; the rival pack wants her; he has one moon to win her back; slow-burn and fierce.

## What werewolf drama viewers come for

The genre is a romance first and a monster story second. Viewers stay for the bond, the hierarchy and the moment someone is put in their place; the wolves are what make the stakes physical. The beats that carry most series:

- The rejection. The heroine is refused in public by the person fate picked for her, usually in episode 1. It is the injury the rest of the series repays.
- The bond. Two characters recognise each other as mates: a scent, a pull, a look neither can break. The audience knows before the characters admit it.
- The hierarchy. Alpha, beta, the luna at the alpha’s side, the omega at the bottom. Every scene is about who outranks whom, and who is about to stop obeying.
- The hidden heritage. The underestimated heroine is more than she seemed: a rare wolf, a lost bloodline, the true heir. The reveal is the first big payoff.
- The rival. A second alpha, a jealous she-wolf or a hunter. Someone has to want her, or want her gone.
- The moon as a clock. A ceremony, a challenge or a curse that lands on the next full moon gives the whole series a deadline.
Rotate the episode endings between those beats — a reveal, a confession left hanging, a betrayal, an arrival — so the hooks never feel mechanical. Hooks and cliffhangers.

## Settings to pick in step 1

Everything in step 1 is read with the script, so decide these before you continue; the type and the visual style cannot be changed later in the same project.

- Type: Short Drama for a feed series — fast, dense dialogue, a hook every episode. TV Series if you want longer episodes with room for pack politics. Short Drama, TV Series or Film.
- Aspect Ratio: 9:16 Portrait for TikTok, Reels and short-drama apps; the default is 16:9 Landscape. Vertical or landscape.
- Visual Style: Digital HD or Shot on Film for the look of the big romance apps; Soft Glow if the story leans on moonlit longing; Painterly Anime or Korean Webcomic for an anime audience.
- Scene Style: American Small Town for a pack hidden in an ordinary town, Suburban Mansion for the alpha family’s estate, Gothic Dark if the story drifts into curses and old blood, Frozen Wilderness for a northern pack.
- Costume Style: Streetwear or Smart Casual for a modern pack, Power Tailoring if the alpha is also a billionaire, Outdoor Gear for a pack that lives in the woods.
- Target Region: Western gives you English names, American or European towns and faces; Follow script keeps whatever your text says.

Visual Style: Soft Glow

Scene Style: American Small Town

Scene Style: Suburban Mansion

Scene Style: Frozen Wilderness

Costume Style: Streetwear
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Cast the wolf as a state of the character

A reference image locks what a character looks like. A man and a wolf are two different shapes, so the wolf cannot be talked into a shot with words alone — it needs its own image. In SceneMixer that image is a state: a second look that belongs to the same character, keeps the same card, the same voice and the same name in the storyboard.

- Generate the human main image first. A state is made from the main image, and the button waits for it: Generate the main image first — the state image is built from it.
- Open the character in step 2 and choose Add a state. Name it plainly — Wolf form.
- Write only the difference. The field asks How does it differ from the main image? Answer that and nothing else: A huge grey wolf with the same amber eyes, a white scar through the fur over the left brow, thick winter coat. The carried-over detail — the eye colour, the scar — is what makes the audience read the wolf as him.
- Press Generate this state. It is priced like any other setup image.
- Point the right shots at it. The storyboard picks a state per segment when it can tell; check the dashed AI pick and override with Use this state in this shot or Use main image where it guessed wrong — in the episode editor, select the segment and open the character from the episode’s assets.
A full wolf is an animal, so describe it as one: four legs, no clothes, its size against a person (shoulder as high as a man’s chest). If your story has an in-between form — claws out, eyes gone gold, fangs — make that a second state rather than asking one image to be both. Costume changes and states · Non-human characters.

Character sheets from the sample Crimson Vow: one reference image per character, used in every shot they appear in.

## Write the transformation in two shots

A body changing shape inside one continuous shot is the hardest thing you can ask a video model for, and the most likely to come out as a melt. The genre’s own grammar has the answer: show the trigger, cut, show the result. Two shots, each with one thing happening.

| Shot | What to write |
| The trigger | [Shot 4] 3s: Cut to close-up. In frame only: @Kael. His jaw locks; his eyes catch the moonlight and turn amber. He drops to one knee. No dialogue. Ambient: wind in the pines. |
| The result | [Shot 5] 4s: Wide shot, low angle. The wolf lands on the forest path between @Elena and the hunters, head lowered, lips drawn back from its teeth. No dialogue. Ambient: a low growl, branches cracking. |
| The reaction | [Shot 6] 3s: Close-up. In frame only: @Elena. She stares, then steps forward instead of back. @Elena (barely a whisper): It’s you. |
Shot 5 is where the Wolf form state belongs; shots 4 and 6 use the main images. Writing the cause (moonlight, a drop to one knee) instead of the effect is what makes the trigger shot readable. Describe the cause, not the result · How to write one shot.

## Pack fights without getting blocked

Every generation passes the provider’s content check, and extreme violence and gore do not pass. A werewolf series does not need them. What reads as danger on a phone screen is the threat and the motion, not the injury:

- Show the lunge, the snarl, the circling, the moment before contact — then cut to the reaction.
- Let the aftermath carry the damage: a torn jacket, a limp, someone pulling another to their feet. No open wounds, no blood on teeth.
- Keep a fight to two or three figures in frame; a whole pack charging is a crowd shot and loses its faces. Put the rest of the pack in the background line of the segment instead.
- Keep romance to what a broadcast drama would show. Explicit content is not generated, and a mark or a claim scene works better as a look and a held breath than as anything physical.
If a segment is blocked, change the storyboard text before you retry: resubmitting the same text only gets blocked again, and repeated blocks pause video generation on SceneMixer for 1 hour. Blocked by moderation: what to change · Fights, crowds and group scenes.

## Moonlight, forest and the pack house

Night, a full moon, rain and firelight are written into the shot, not made into images: they ride on top of whatever the location image shows. What deserves its own location is the place the story keeps returning to — the clearing where the ceremony happens, the alpha’s house, the diner where she works — so it is the same room in every episode. Write the moon as a cause of light (a full moon low over the treeline throws cold light across the clearing), and put the watching pack in the segment’s background line rather than naming each wolf. Same room in every episode · Background people in a scene.

## Voices for an alpha and a luna

Each character’s Voice description travels with every line they speak. Werewolf leads are usually written against each other: low and unhurried, a rasp at the ends of lines, never raises his voice for the alpha; soft but steady, gets quieter when she is angry for the heroine; bright, quick, mocking for the rival. A pack elder or a narrator who is heard and never seen can be set to Voice only (off-screen) so no image is made for them. Growls and howls belong in the shot’s Ambient line, not in dialogue. How character voices work · Give a character an accent.

## Make it in SceneMixer, start to finish

- Get a script. Paste your own on Upload Script, or use AI Create: one or two sentences with the genre, the lead and what she wants, the opponent, the ending and the tone, then pick a length and Start Writing. Writing the one-line brief.
- Set up step 1 with the choices above and let the script be read. Check the episode list in the summary: every episode should end on a turn.
- Build the cast in step 2. Generate the main images (the first 15 setup images on an account are free), then add the wolf states. Give the alpha one mark the wolf shares.
- Generate the storyboard for episode 1 (on your first project, the first episode’s storyboard is free) and read it like a script. Rewrite any transformation that happens inside a single shot into the two-shot pattern.
- Generate the segments. The default tier costs 10 credits a second; draft cheaply, finish higher if the episode works. Draft at 480P, finish at 1080P.
- Fix what came out wrong one shot at a time, then composite the episode and post it. Fix one shot · Post to TikTok, Shorts and Reels.
For a feel of how a supernatural story looks when it is built this way, watch Crimson Vow, a rain-soaked city thriller that turns into a gothic bloodline mystery.

| Episode | What happens | How it ends |
| 1 | The full-moon ceremony. The alpha heir recognises her as his mate and rejects her in front of the pack. | She walks out into the forest and something in her answers the moon. |
| 2 | She is cast down to the lowest rank; the heir cannot stop thinking about her. | A stranger from the rival pack watches her from the treeline. |
| 3 | The rival alpha offers her a place at his side. | Her eyes turn white-gold for the first time, in front of the heir. |
| 4 | The elders recognise the sign of the lost white wolf. | The heir’s father orders her taken before the next moon. |
| 5 | She escapes with help from the one person she did not expect. | The heir finds her and says the word he refused to say in episode 1. She says no. |

## FAQ

### Can SceneMixer show a werewolf transformation?

Yes, the reliable way: a trigger shot on the human (eyes, breath, a drop to one knee), then a cut to the wolf. The wolf is a state of the same character with its own image, so it looks the same every time it appears.

### Will the wolf look the same in every episode?

If the wolf form is a generated state, yes: every shot that uses the state uses the same reference image. Without an image the state is only described in words, which is weaker; generate it before you shoot the wolf scenes.

### Can I make a werewolf drama in anime style?

Yes. Pick an anime preset such as Painterly Anime or Korean Webcomic in SceneMixer's step 1. The style applies to every character, state and video in the project and cannot be changed after you continue.

### What gets blocked in a werewolf story?

Gore, open wounds, extreme violence and anything sexual. Threat, chases, snarls and the moment before a fight are fine to write. If a segment is blocked, change its storyboard text before retrying.

### How long should each episode be?

For a feed series, one to two minutes with a hook at the end. Pick Short Drama in step 1 and the script is split to that rhythm.

### Can the dialogue be in Spanish, Portuguese or another language?

Yes. Choose the dialogue language in SceneMixer's step 1; the cast speaks it natively, with mouths moving to the words, rather than being dubbed afterwards.

### Can I adapt a werewolf novel I am reading?

Only with the rights to it. A published book or a well-known franchise needs the author's or publisher's permission; an original story, or your own novel, is yours to adapt.

### How much does a werewolf episode cost?

The same as any episode of that length and tier: the script read, the setup images, the storyboard and then video by the second. The worked budget guide adds it up for a whole episode.

Updated November 3, 2026


---

# AI Western Drama: Cowboys, Frontier Towns and a Showdown at Noon

URL: https://scenemixer.com/guides/make-an-ai-western-cowboy-drama/
Make a western with AI: the Wild West scene style, Western Cowboy costumes, horses as characters, frontier towns as locations, and a showdown cut before the shot.

# How to make an AI western or cowboy drama

Quick answer: A western is a town, a stranger and a reckoning. SceneMixer builds the frontier in Step 1: Wild West (red earth, sandstone mesas, plank buildings and saloon fronts under noon sun or an orange sky) with Western Cowboy costumes (wide-brim hats, boots, vests, neckerchiefs, leather fringe). A horse that matters is a character with its own sheet. Showdowns are cut before the shot lands: the hands, the eyes, the sound.

Brief for AI Create: A western: a widow running her late husband's ranch refuses to sell to the railroad baron; the stranger she hires to protect it is the man who shot her husband; 10 episodes, each ends at noon.

## The frontier in Step 1

| Kind of western | Scene Style | Costume Style |
| Classic frontier town | Wild West | Western Cowboy |
| Modern ranch romance | Wild West or American Small Town | Western Cowboy, Outdoor Gear |
| A Mexican border town | Latin Colonial Town | Latin American Folk |
| A snowbound mountain western | Frozen Wilderness | Western Cowboy |

## Horses, hats and the town

- A horse that matters is a character. Colour, markings, one mark. A real animal is drawn on four legs. Non-human characters.
- The saloon, the jail and the ranch house are locations used again and again. The same room.
- Noon and dusk are segment settings: Time and Weather on the header. Segment header.
- Guns are props; shootings are off-screen. Content rules.

## The showdown, shot by shot

| Beat | Shot to adapt |
| The street | [Shot 1] 4s: Wide shot, a dusty main street at noon, plank storefronts, two men thirty paces apart. |
| The hand | [Shot 2] 3s: Close-up on @Wade's hand above his holster, fingers flexing. |
| The window | [Shot 3] 3s: Medium shot. @Clara watches from the general store window. |
| The line | [Shot 4] 4s: Close-up on @Baron Hollis. @Baron Hollis (calm): She's not worth dying for, son. (voice: Hollis) |
| The cut | [Shot 5] 2s: Close-up on a startled horse rearing at the hitching post. Ambient: a single gunshot. |

## Story shapes that work

- The widow's ranch.
- The stranger in town.
- The bounty hunter and the outlaw.
- The cowboy romance on a modern ranch.
- The railroad coming through.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Wild West; Costume Style Western Cowboy; Visual Style Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the widow, the stranger and the baron with one mark each; the horse as a character; the saloon, the jail and the ranch as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The ranch. | The railroad's offer, and her refusal. |
| 2 | The saloon. | She hires a stranger with no name. |
| 3 | The ranch at night. | Her husband's watch in the stranger's saddlebag. |
| 4 | The jail. | The sheriff works for the baron. |
| 5 | Main street, noon. | One shot, off-screen. |

## FAQ

### How do I make a western with AI?

Choose Wild West and Western Cowboy in SceneMixer's Step 1, make important horses characters, and build showdowns from short shots cut before the shot lands.

### What does the Wild West style look like?

The American West: red earth, sandstone mesas, plank buildings and saloon fronts under noon sun or an orange evening sky.

### Can a horse be a character?

Yes. Give it a sheet in SceneMixer's Step 2 with colour and markings.

### Can I show gunfights?

Show the moment before and the sound; shootings and wounds stay off-screen.

### Can it be a modern cowboy romance?

Yes. Wild West or American Small Town with Western Cowboy and Outdoor Gear.

### Which visual style suits a western?

Shot on Film; Black & White Film for a classic look.

### Vertical or landscape?

16:9 Landscape suits the frontier; 9:16 Portrait suits feeds.

### How do I set noon or dusk?

Choose the Time on each segment's header.

Updated November 3, 2026


---

# AI Witch and Magic Academy Drama: Spells, Houses and Rivals

URL: https://scenemixer.com/guides/make-an-ai-witch-or-magic-academy-drama/
Make a magic school or witch drama with AI: the Magic Academy scene style, Academy Robes with house colours, spells shown through light and effect, and a forbidden book.

# How to make an AI witch or magic academy drama

Quick answer: A magic academy drama is a new student, a school with secrets and a rival in the wrong house. SceneMixer has it built in: Magic Academy (Gothic stone halls, spiral stairs, tall bookshelves, floating candles, stained glass) and Academy Robes (robes over shirt and tie, house badges and clear house colours). For a witch in the woods or a modern coven, use Elven Forest, Gothic Dark or a modern city instead. Spells read best as light and effect described plainly in the shot.

Brief for AI Create: A magic academy drama: a scholarship student with no magical family is sorted into the house that expelled her mother; the library's forbidden section opens only for her; 12 episodes, each ends with a spell going wrong in front of the wrong person.

## School or coven

| Kind of magic story | Scene Style | Costume Style |
| A school of magic | Magic Academy | Academy Robes |
| A witch in the woods | Elven Forest or Pastoral Healing | Fantasy Adventurer |
| A dark coven | Gothic Dark | Gothic Court, Victorian |
| Witches in a modern city | New York Metropolis | Streetwear, Rock & Punk |

## Houses, spells and books

- Houses by colour. Academy Robes shows house colours clearly; write each student's house into the script.
- Spells are cause and effect. “A thin line of blue light runs from her fingertips to the candle, which lights” works; “powerful magic” does not. Describe the cause.
- The forbidden book, the wand, the familiar are props or characters with their own picture. A cat familiar is a character. Non-human characters.
- Original world. Famous magic school names, houses and scarves trigger the IP check. IP check.

## The forbidden section, shot by shot

| Beat | Shot to adapt |
| The library | [Shot 1] 4s: Wide shot, a towering library at night, floating candles, an iron gate across one aisle. @Iris holds a lantern. |
| The gate | [Shot 2] 3s: Close-up. The iron gate's lock glows faint blue as she touches it, then swings open. |
| The rival | [Shot 3] 3s: Medium shot. @Felix steps out of the shadows in green-trimmed robes. |
| The line | [Shot 4] 4s: Close-up on @Felix. @Felix (stunned): That gate hasn't opened in thirty years. (voice: Felix) |

## Story shapes that work

- The scholarship student.
- Rival houses, one romance.
- The teacher with a past.
- The witch's apprentice in the woods.
- The hidden witch in a modern school.
High school dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Magic Academy; Costume Style Academy Robes; Visual Style Digital HD, Painterly Anime; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: students with their house in the script and one mark each; the book and the wand as props; the library, the great hall and the tower as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The great hall. | She is sorted into her mother's old house. |
| 2 | The potions room. | Her first spell sets the room alight. |
| 3 | The library. | The forbidden gate opens for her. |
| 4 | The tower. | A portrait of her mother, scratched out. |
| 5 | The duel. | Her rival lets her win. |

## FAQ

### How do I make a magic school drama with AI?

Choose Magic Academy and Academy Robes in SceneMixer's Step 1, write each student's house, and describe spells as light and effect in each shot.

### What does the Magic Academy style look like?

An old school of magic: Gothic stone halls and spiral stairs, tall bookshelves, floating candles, stained glass and astronomical instruments.

### How do I show spells?

Describe the light and what happens: a line of blue light, a candle lighting, a door opening.

### Can I copy a famous wizard school?

No. Known names, houses and uniforms trigger the IP check. Invent your own.

### Can a familiar animal be a character?

Yes. Give it a sheet in SceneMixer's Step 2.

### Which visual style suits magic?

Digital HD or Soft Glow for live action; Painterly Anime or Watercolor Anime for a drawn look.

### Can it be a modern witch story?

Yes. Use a modern city scene style and Streetwear or Rock & Punk costumes.

### Vertical or landscape?

Either; 16:9 Landscape suits the great hall, 9:16 Portrait suits feeds.

Updated November 3, 2026


---

# AI World War II Drama: Fronts, Homefronts and Invented Insignia

URL: https://scenemixer.com/guides/make-an-ai-world-war-ii-drama/
Make a World War II drama with AI: the WWII Front scene style and WWII Era costumes with invented insignia, invented characters, and war shown before and after.

# How to make an AI World War II drama

Quick answer: A war drama is a person and an impossible choice inside a history that is bigger than them. SceneMixer has the period built in: WWII Front (trenches, sandbag bunkers, barbed wire and shell-damaged buildings on the European, Pacific and Chinese fronts) and WWII Era costumes for soldiers and civilians; the insignia are always invented. Real historical figures cannot be depicted, so your characters are invented people living through real years.

Brief for AI Create: A WWII drama: a village schoolteacher hides a downed pilot in her school's attic while enemy officers are billeted in the classrooms below; her best student starts asking questions; 10 episodes, each ends with footsteps on the stairs.

## The war in Step 1

| Where | Scene Style | Costume Style |
| The front | WWII Front | WWII Era |
| An occupied town | Parisian Streets or Eastern European Old Town | WWII Era |
| Wartime China | Republican Era | Republican Modern |
| The homefront | American Small Town or English Manor | WWII Era |
Visual Style Black & White Film or Shot on Film for the period. History dramas · Military dramas.

## History without real people, war without gore

- Invented people. Real people, including historical leaders, cannot be depicted; uniforms carry invented insignia. Content rules.
- Combat before and after. The barrage heard from a cellar, the field the morning after. Fights and crowds.
- Damage is a state of a location: the town before and after the shelling. States.
- Letters and radios carry news; make them props. Props.

## The attic, shot by shot

| Beat | Shot to adapt |
| The classroom | [Shot 1] 4s: Wide shot, a village classroom turned into an officers' billet, maps pinned over the blackboard. |
| The ceiling | [Shot 2] 3s: Close-up on the ceiling boards as dust sifts down. Ambient: a cough from above. |
| The officer | [Shot 3] 3s: Medium shot. @Captain Weiss looks up. |
| The line | [Shot 4] 4s: Close-up on @Marie. @Marie (too quickly): Mice. The building is old. (voice: Marie) |

## Story shapes that work

- The hidden pilot.
- Letters home from the front.
- The resistance cell in an occupied city.
- The nurse at a field hospital.
- The family separated at a railway station.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style WWII Front, Parisian Streets; Costume Style WWII Era; Visual Style Black & White Film, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: invented characters with one mark each; the school, the attic and the square as locations, with a damaged state where needed. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | The field at dusk. | A parachute in the trees. |
| 2 | The school. | Officers are billeted in her classrooms. |
| 3 | The attic. | A cough, and a captain looking up. |
| 4 | The village square. | Posters with the pilot's description. |
| 5 | The school at night. | Her best student is on the attic stairs. |

## FAQ

### How do I make a World War II drama with AI?

Choose WWII Front or a period town scene style and WWII Era costumes in SceneMixer's Step 1, invent your characters, and show combat before and after.

### Can I show real historical leaders?

No. Real people cannot be depicted on SceneMixer. Invent the characters; history can stay the backdrop.

### Are the uniforms accurate?

They follow the period with invented insignia.

### Can I show battles?

Through sound, faces and aftermath; graphic violence is blocked.

### How do I show a town after shelling?

Add a damaged state of the location.

### Which visual style suits WWII?

Black & White Film or Shot on Film.

### Can it be set in Asia?

Yes. WWII Front covers the Pacific and Chinese fronts; Republican Era suits wartime Chinese cities.

### Vertical or landscape?

16:9 Landscape for a period series; 9:16 Portrait for feeds.

Updated November 3, 2026


---

# AI Xianxia and Wuxia Drama: Cultivation, Sects and Sword Fights

URL: https://scenemixer.com/guides/make-an-ai-xianxia-or-wuxia-drama/
Sects, cultivation, a framed disciple and a flying sword: how to write a xianxia or wuxia short drama and build its immortal look, spells and duels in SceneMixer.

# How to Make an AI Xianxia or Wuxia Drama

Quick answer: Xianxia (immortal fantasy) and wuxia (martial-arts adventure) are the most recognisable Chinese drama genres abroad: a disciple wronged by the sect, a long climb in power, and a return to settle the debt. SceneMixer has presets built for them — the Chinese 3D Animation visual style, Celestial Colossus, Ink Jianghu and Dark Wuxia sets, Immortal Sect Robes and Jianghu Travelwear — so the look comes from step 1. Make the sword’s drawn form a state of the prop, write spells as light sources, and stage duels as single exchanges rather than long choreography.

Brief for AI Create: A cultivation revenge story: a sect's kitchen boy is framed for stealing the master's jade token and cast out; he returns ten years later as a grandmaster to expose the elder who framed him; every episode ends on a reveal; ruthless.

## What the genre runs on

- The sect. Masters, elders, inner and outer disciples. The hierarchy is the world, and the lead starts at the bottom of it.
- The injustice. A framing, a crippled talent, a stolen inheritance. Episode 1 delivers it in public.
- Cultivation. Power gained in stages: a breakthrough, a hidden technique, a treasure that chooses its owner. Each stage is a payoff.
- The token. A jade seal, a sword, a scroll — an object that proves who is right. It recurs until the final reveal.
- The duel. Not a brawl but a test, watched by the whole sect. Who wins matters less than who is exposed.
- Scale. Cloud seas, endless jade stairs, halls built for gods. The lead looks small against them on purpose.

## Settings to pick in step 1

- Visual Style: Chinese 3D Animation for the polished immortal-fantasy CG look; Digital HD for a live-action costume drama; Ink Wash Animation or Chinese Painted Style for a painted one.
- Scene Style: Celestial Colossus for floating palaces, giant stairs and cloud seas; Ink Jianghu for misty mountains and inns; Dark Wuxia for a grimmer martial world; Dark Asura for demonic realms.
- Costume Style: Immortal Sect Robes for flowing sect dress, Jianghu Travelwear for wandering swordsmen, Dark Path Robes for demonic cultivators, Divine Regalia for gods and emperors of heaven.
- Target Region and dialogue: East Asia gives East Asian names, faces and architecture; to keep them specifically Chinese, use Chinese names in the script or pick Follow script. The dialogue language can be English, Chinese or any of the others.
- Type: Short Drama for a fast reversal series; TV Series for a slower epic.
Visual style presets, one by one.

Visual Style: Ink Wash Animation

Scene Style: Celestial Colossus

Scene Style: Ink Jianghu

Scene Style: Dark Wuxia

Costume Style: Immortal Sect Robes

Costume Style: Jianghu Travelwear
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Swords, tokens and treasures

- The sword is a prop with a state. A sheathed sword and a drawn, glowing one are different shapes. Give the sword its own prop image, then choose Add a state on it, name it Drawn, and describe only the difference: blade out of the scabbard, a thin line of pale blue light along the edge. Pick it in the shots where it is drawn. States.
- The token recurs. If your plot hangs on a jade seal or a scroll, check it has a card and an image in step 2, and name it with @ every time it is in frame. Same prop in every hand.
- A broken treasure — a snapped staff, a cracked mirror — is another state, because its shape has changed.

## Spells and qi: write them as light

A video model draws light well and draws vague “power” badly. Describe a technique by what it does to the light and the air around it:

| Instead of | Write |
| He uses his ultimate technique | A ring of gold light opens on the ground around him; the wind lifts his sleeves; the stones at his feet rise an inch. |
| She breaks through to the next realm | Pale light gathers in her palms and climbs her arms; her hair lifts as if underwater; the candle flames in the hall lean toward her. |
| A powerful aura | The disciples nearest him step back without meaning to; the banners behind him snap in a wind no one else feels. |
A lasting physical change from a breakthrough — white hair, a mark on the brow — changes the reference image and belongs in a state of the character. Describe the cause, not the result.

## Duels as exchanges, not choreography

Fights that need long, continuous choreography are among the hardest things to generate. A xianxia duel does not need them; it needs three or four clear exchanges and the crowd’s reaction between them.

| Beat | Shot to adapt |
| The challenge | [Shot 1] 4s: Opening: wide shot, low angle. @Lin Feng stands alone at the foot of the jade stairs; @Elder Qin looks down from the top. @Elder Qin (cold, carrying): A thief returns to be judged. Ambient: wind across the cloud sea. |
| One exchange | [Shot 2] 3s: Medium shot. @Elder Qin’s palm strike meets @Lin Feng’s raised forearm; a ring of light bursts outward from the contact. No dialogue. Ambient: a deep boom. |
| The witnesses | [Shot 3] 3s: Wide shot. The watching disciples step back as one. No dialogue. Ambient: murmurs. |
| The turn | [Shot 4] 4s: Close-up. In frame only: @Lin Feng. He opens his hand: the @Jade token, glowing. @Lin Feng (quiet): Tell them who gave this to me. |
The watching disciples are the segment’s background line. No blood is needed: the reveal is the blow. Fights, crowds and group scenes.

## Names, titles and voices

Sects run on titles — Senior Brother, Elder, Sect Master — and characters are called by several of them. Keep one name on each card and add the titles and nicknames under Also called so the system recognises the person whichever one the script uses. Also called. For voices, an elder is slow, resonant, every sentence a verdict; a young master is light, arrogant, quick to laugh at others. How character voices work.

Character sheets from the sample 浪人の誓い: one reference image per character, used in every shot they appear in.

## Make it in SceneMixer, start to finish

- Script: paste a web novel chapter or your own script, or a brief in AI Create. Turn a web novel into a series.
- Step 1: the xianxia settings above; read the script.
- Step 2: generate the cast, the sect halls and the stairs; give the sword and the token their images and the drawn sword its state.
- Storyboard: episode 1 first (free on your first project). Break any long fight into exchanges.
- Video: generate on the default tier at 10 credits a second, regenerate what misses, composite and post.
See the look in 問劍青雲, a xianxia sample with a cloud-sea sect, giant jade stairs and a blood-stained token, and the martial counterpart in 浪人の誓い, a ronin story with native Japanese dialogue.

| Episode | What happens | The hook |
| 1 | The kitchen boy is accused of stealing the master’s jade token and cast out in front of the sect. | At the foot of the mountain, the token is in his sleeve. He never took it. |
| 2 | A hermit in the ruins sees what the sect missed: a sealed talent. | The seal breaks; the ruins light up. |
| 3 | Ten years pass. A grandmaster arrives at the sect’s tournament under another name. | The elder recognises the way he holds his sword. |
| 4 | The tournament duel. | He opens his hand: the token, glowing. |
| 5 | The elder’s defence collapses; the master’s death is reopened. | “He did not die of illness.” |

## FAQ

### Is there a xianxia visual style?

Yes. Chinese 3D Animation is a visual style preset in step 1, and Celestial Colossus, Ink Jianghu and Dark Wuxia are scene styles built for the genre.

### Can the characters speak English in a xianxia drama?

Yes. Choose English, or any other dialogue language, in SceneMixer's step 1. For Chinese names, faces and architecture, use Chinese names in the script and pick East Asia or Follow script as the region.

### How do I show a sword being drawn?

Give the sword its own prop image and a Drawn state, then pick the state in the shots where the blade is out.

### How do I show spells and qi?

As light and motion: a ring of light on the ground, sleeves lifted by a wind, stones rising. Vague words like power or aura give the model nothing to draw.

### Can AI make a good sword fight?

As a few clear exchanges with reactions between them, yes. Long continuous choreography is unreliable; cut the fight into single moves.

### Can I adapt a web novel?

Yes, if you have the rights. Paste chapters into SceneMixer; long texts are read in two passes and split into episodes.

### How do I handle characters called by many titles?

Keep one name on the card and add the titles and nicknames under Also called, so the system recognises the person either way.

### Does a xianxia drama cost more because of the effects?

No. Every genre on SceneMixer is priced the same way: the read, setup images, the storyboard and video by the second.

Updated November 3, 2026


---

# AI Zombie Apocalypse Drama: Outbreak, Survivors and Safe Zones

URL: https://scenemixer.com/guides/make-an-ai-zombie-apocalypse-drama/
Make a zombie apocalypse series with AI: Wasteland and Wasteland Scavenge styles, infection as a character state, horror without gore, and cliffhangers that keep viewers.

# How to make an AI zombie apocalypse drama

Quick answer: A zombie drama is a survival story with a clock on it: who gets bitten, who lies about it, and how long the walls hold. In SceneMixer, set the world in Step 1 with the Wasteland scene style (fallen overpasses, sand-covered city skeletons, tidy survivor camps) and Wasteland Scavenge costumes (salvaged leather, canvas and riveted plates). Make the infected look of a character a separate state in Step 2, keep the bites and the violence just off-screen, and end every episode on a door, a sound or a fever.

Brief for AI Create: A zombie apocalypse drama: a nurse who hid her own bite gets the last place in a walled settlement; the settlement's leader is her ex-husband, and he is the only one who knows the outbreak started in his lab; 12 short episodes, each ends with someone at the gate.

## Build the world in Step 1

| Stage of the outbreak | Scene Style | Costume Style |
| Day one, the city falls | New York Metropolis or Neon Nights | Smart Casual, Professional Uniforms |
| Months later, the ruins | Wasteland | Wasteland Scavenge |
| The settlement or the base | Military Base | Modern Military or Outdoor Gear |
| Small-town outbreak | American Small Town | Soft Everyday |
| The quarantine ward | Hospital Ward | Professional Uniforms |
The styles are written into every description when the script is read, so pick them first. Every scene and costume style.

## Infection is a state, not a sentence

A model copies the face it is given. If a character turns, the grey skin and the clouded eyes need their own picture: open the character in Step 2 and add a state (“infected”) next to the main sheet, then pick that state in the segments after the turn. Torn clothes after a fight are a state too. Dirt, rain and who is holding the gun are words in the shot. Costume changes and states.

Hordes are background: write a crowd of the infected shuffling between burnt-out cars in the shot and keep the named characters in front. Background people in a scene.

## Fear without gore

Graphic injuries and extreme violence are blocked by content review, and a blocked attempt costs you time. Zombie fear does not need them. Show the moment before and the moment after: a hand on a door that will not hold, the sound of glass, a survivor checking her own arm in a mirror. Fights, crowds and group scenes · Content rules.

| Beat | Shot to adapt |
| The gate | [Shot 1] 4s: Wide shot, night. A chain-link gate under one floodlight; beyond it, shapes moving between wrecked cars. @Mara stands inside with a rifle lowered. |
| The sound | [Shot 2] 3s: Close-up on the padlock as the gate shudders once. |
| The secret | [Shot 3] 3s: Close-up on @Mara pulling her sleeve down over a bandage on her forearm. |
| The line | [Shot 4] 4s: Medium shot on @Elias at the watchtower. @Elias (quiet): Nobody comes in after dark. Nobody. (voice: Elias) |

## Story shapes that work

- The hidden bite: one character is infected and nobody knows; every episode the circle of people who know grows by one.
- The safe zone: the walls are not the danger, the people who run them are.
- The road trip: a family crosses a country, one town per episode.
- The cure: a scientist, a sample and three groups who want it.
- Romance in the ruins: two rival survivors forced to share a car.
Hooks and cliffhangers · Horror shorts.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Wasteland; Costume Style Wasteland Scavenge; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: one face and one mark per survivor; an infected state for anyone who turns; the gate, the ruins and the lab as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | A hospital on the first night. | The nurse sees the bite on her own arm. |
| 2 | The road out of the city. | A settlement gate and a man she knows. |
| 3 | Inside the walls. | Someone searches the newcomers' bags. |
| 4 | The ruins on a supply run. | Her fever starts while they are pinned down. |
| 5 | The lab under the settlement. | The outbreak's first case file has his name on it. |

## FAQ

### How do I make a zombie apocalypse drama with AI?

Pick a scene style such as Wasteland and a costume style such as Wasteland Scavenge in SceneMixer's Step 1, write or generate the script, make each infected character a separate state in Step 2, and render episode by episode.

### How do I show a character turning into a zombie?

Add a state to that character in SceneMixer's Step 2 with the changed face and clothes, and pick it in every segment after the turn. Words alone will not change a face the model has a picture of.

### Can I show bites and gore?

No. Graphic injury is blocked by content review. Cut away before the bite and come back to the aftermath, a bandage, a fever, a closed door.

### How do I film a horde?

Write the horde as background people in the shot and keep two or three named characters in front of it.

### Which visual style suits a zombie drama?

Digital HD or Shot on Film for live action, Black & White Film for a bleak look, or Game Engine Realism for a video-game feel.

### Should it be vertical or landscape?

9:16 Portrait for a short-drama feed, 16:9 Landscape for a YouTube series or a film.

### Can the survivors keep the same faces for twelve episodes?

Yes. Each character is drawn once in SceneMixer's Step 2 and that sheet is used in every shot, including their states.

### Can I name a famous zombie show?

Do not. Names and iconic looks from existing films and shows trigger the IP check. Invent your own outbreak.

Updated November 3, 2026


---

# An Amazon Product Video With AI: Landscape, No Price, Just the Product

URL: https://scenemixer.com/guides/make-an-amazon-product-video-with-ai/
Make an Amazon listing video with AI from your product photos: landscape, 30 to 60 seconds, Show how it works, and an end card with no price or promotion.

# How to make an Amazon product video with AI

Quick answer: An Amazon listing video has one job: show the thing working, in the gallery, to someone already looking at it. In SceneMixer, upload one to four photos of the product, pick Landscape, 30 or 60 seconds and Show how it works, and the video is written, shot and finished. The product is redrawn as a clean studio shot and used in every scene, so it matches your gallery images. Keep the end card to the brand and the product name: Amazon's content rules for listing videos exclude prices, promotions, time-limited offers, competitor names and customer reviews, and the format and length limits are published in Seller Central and change, so check them before you upload.

One line for the box next to the upload: A 40-ounce insulated steel bottle with a straw lid, dishwasher safe; for people who forget to drink water at their desk.

## What Amazon wants from a listing video

- The product, demonstrated. Size, how it opens, what it does, what comes in the box. The buyer is already on the page; the video answers the questions the photos cannot.
- Nothing time-sensitive. No prices, no sale, no limited offer, no holiday. The video stays up for years.
- No comparisons or reviews. No competitor names, no star ratings, no best seller. Say what it is.
- Landscape. Gallery videos play in a landscape frame; choose 16:9.
Brand Registry is usually required to add video to a listing, and the file format, resolution and length limits are published in Seller Central; they have changed before, so read the current page before you export. Length, format and quality.

## The photo

Your gallery photos are the right input: the product on white, three-quarter, the lid off, the thing in use. Up to four. The product is cleaned into a studio shot on SceneMixer and kept identical in every scene, which is what you want beside gallery images of the same object. People in a lifestyle photo are removed; the person using the product in the video is generated. If scale matters, say it in the line (fits a car cup holder). Which photos work · An ad for a physical product.

## Tone and length

| Tone | Listing video use |
| Show how it works | The default: open, use, close, what is in the box. |
| A day in the life | The product inside a routine; for everyday items. |
| Minimal | The object and light; for design-led products. |
| Avoid | Twist and Funny read as ads, not demonstrations; Epic promises scale a listing cannot prove. |

Thirty seconds covers one product and three facts; sixty covers a product with parts or steps, one per segment. The spoken line should state facts (holds forty ounces, fits a cup holder), not sell. Superlatives are removed from the script anyway. What you actually pay.

## The end card, stripped back

For a listing video the end card is the brand and the product name only: no tagline with an offer, no Shop now (the button is already on the page). Edit the card after the cut and rebuild without reshooting. The same shots can then get a second end card with an offer for use off Amazon, on your own site or social. The end card.

## Variants and markets

A product in five colours is one video with the hero colour, or five short ones from five photos. A product sold on several Amazon marketplaces is the same video with the spoken line in another language: the line is performed in the language you choose, so a German and a Spanish version are two ads from the same photos. A second-language version.

## Make it in SceneMixer, start to finish

- Photos: open the Ad Video tab and upload one to four photos of the same subject. The read is free and tells you what it saw: category, product name, any brand name legible on the label.
- One line: say what it is and who it is for. When the line and the photo disagree, the line wins.
- Pickers: vibe, length (15, 30 or 60 seconds), portrait or landscape, 480P / 720P / 1080P. The price beside the quality is the ceiling you will pay.
- Make my ad: the script, cast, locations and storyboard are written, every segment is generated with its sound, and the end card carries your brand, product name, tagline and call to action.
- Fix, then export: a shot that missed can be regenerated alone; the end card text can be edited and the cut rebuilt without reshooting. The ad progress page.

| Rule for the listing | How the ad meets it |
| Landscape frame | Choose 16:9 Landscape in the pickers |
| No price or promotion | Leave the offer out of the line; end card = brand and product name |
| No competitor names or reviews | The line describes the product; superlatives are removed |
| Product must match the listing | The product is redrawn from your gallery photo and kept identical |
| Format and length limits | Export 1080P MP4; check Seller Central for the current limits |

## FAQ

### Can I use an AI video on an Amazon listing?

You can make a product demonstration video on SceneMixer and upload it as a listing video; Amazon's own rules on content, format and eligibility apply, so check Seller Central.

### What should the video show?

The product working: size, how it opens, what it does, what is in the box. Show how it works is the tone for it.

### Can it mention the price?

Not for a listing video; Amazon excludes prices and promotions. Keep the end card to the brand and product name.

### Landscape or portrait?

Landscape 16:9 for the Amazon gallery. Make a portrait cut separately for social.

### Does the product match my gallery photos?

It is redrawn by SceneMixer from your photo as a studio shot and kept identical in every scene.

### Can I make versions for other marketplaces?

Yes. The spoken line follows the language you choose; make another ad from the same photos.

### How long should it be?

30 seconds for one product and three facts; 60 for parts or steps, one per segment.

### Do I need Brand Registry?

Amazon generally requires it to add video to a listing; the video itself does not depend on it.

Updated November 3, 2026


---

# How to Make ReelShort-Style Dramas with AI | SceneMixer

URL: https://scenemixer.com/guides/make-reelshort-style-dramas/
Five steps to ReelShort-style vertical dramas with AI: format rules, script prep, consistent characters, per-shot video, episode assembly and costs.

Guide

# How to Make ReelShort-Style Dramas with AI

Updated November 3, 2026

Quick answer: ReelShort-style dramas are 9:16 vertical series — 60–120 second episodes, a hook every episode, 40–80 episodes per show. To produce one with AI you need four things: a serialized script with cliffhanger pacing, a cast that stays visually identical across episodes, vertical-native video generation, and per-episode compositing. SceneMixer packages all four into one pipeline: upload a script or novel, and it delivers analyzed episodes, a consistent character library, storyboards, 9:16 segments, and finished episodes.

## The format rules (learned from the hits)

- Vertical, always — 9:16 full-screen phone framing; faces large, two-person close-ups dominate.

- A nail at the end of every episode — identity exposure, emotional suspension, or a reversal. Rotate them; repetition kills retention.

- 3–4 physical locations per episode maximum — the format is conversation- and emotion-driven, not spectacle-driven.

- Series length is the business model — the market (roughly $11B in 2025) monetizes bingeing, so plan 40+ episodes from the same cast.

## The 5-step AI production workflow

- Source a serialized story. Web novels are ideal — they're already written in cliffhanger rhythm. Chinese and English sources both work; SceneMixer can also draft an original script from a one-line idea (try the Inspire Me button).

- Run script analysis. One pass produces the episode structure (with conflict arcs and hooks), the cast list with locked reference art, scenes, and props.

- Localize for your market. Pick a target region — Western, East Asian, Southeast Asian, or Latin American — and names, casting, scenery, and on-screen text adapt. One IP, several markets.

- Generate and review segment by segment. Episodes are broken into 4–15 second shots with editable storyboard scripts. Regenerate only what's wrong; version history keeps every take.

- Composite and publish. One click merges segments into a finished episode. Download and publish to TikTok, YouTube Shorts, Reels, or a short-drama platform.

## What it costs

Three line items on SceneMixer: script analysis priced by text length (4 credits per 3,333 characters; reference images are billed separately, with 15 free per account), storyboard scripts per episode (1 credit per 333 characters, regenerating costs the same) and per-second video generation (from 10 credits per second (≈ $0.06/s at the lowest credit price) depending on the tier (Wan 3.0 · 480P by default; higher resolutions and other models cost more per second), plus 1 credit per second for episode compositing; see the pricing table). Credit packages start at $1.49; Pro membership ($7.99/mo) adds 700 monthly credits; larger credit packs carry top-up bonuses up to 15%. A 90-second episode is a predictable, repeatable cost — which is what makes 40-episode planning possible.

## The one problem you can't ignore: consistency

Viewers forgive AI lighting; they do not forgive the male lead having a different face in episode 7. Whatever tool you use, verify it maintains a persistent character identity across generations — not just within one prompt. SceneMixer generates a full-body reference sheet per character at analysis time and injects face/body crops into every segment featuring that character, across the entire series.

## Frequently asked questions

### Is ReelShort itself an AI tool?

No — ReelShort is a streaming app where audiences watch vertical dramas. To produce ReelShort-style content you need a production tool; SceneMixer is a pipeline that generates that style of serialized vertical drama with AI.

### How long is a ReelShort-style episode?

Typically 60–120 seconds, always 9:16 vertical, ending on a hook. SceneMixer's episode outlines are generated with that pacing, and setting a project to 9:16 makes every segment vertical (the default is 16:9).

### Can one person really produce a whole series?

Yes — that's the point of the pipeline. SceneMixer automates script analysis, art, storyboards, video, and compositing; your job is taste: editing storyboard text and choosing takes.

### Do I own the videos I generate?

Yes, you download and publish your episodes wherever you like. An optional AIGC watermark can be toggled in SceneMixer’s settings.


---

# Make the Next Episode: Add Episodes or Continue Your Story

URL: https://scenemixer.com/guides/make-the-next-episode-add-episodes-or-continue/
Episode 2, adding episodes and second seasons: open the next episode in step 3, add one to the same project, or start a new project that reuses the same cast.

# Making the Next Episode: When It Is Already in Your Script, and When the Story Goes On

Quick answer: There are two kinds of “next episode”. If your script already contains it, it is waiting in step 3: open that episode, press Parse Storyboard, then generate its segments in the editor. If the story ends where your script ends, add the continuation to the same project: at the bottom of step 3, press Add an episode, paste the next part or upload a TXT or DOCX file — or let AI write it from where the story stands and edit the draft before adding. Only the new pages are read; the cast, scenes and props of the earlier episodes are reused, and whatever first appears in the new pages is added to step 2. A second season can still be its own project that brings the cast across with Pick existing.

Open your project and go to step 3, Episode Videos: every episode the reading found has its own card, numbered in order.

## Episodes already in your script

The reading splits your script into episodes once, at the start. All of them are listed in step 3, each on its own card with its title, its characters and scenes, and its own Parse Storyboard button. Episode 2 is made the same way as episode 1: parse its storyboard, press Start Editing, generate the segments, composite. You do not have to finish one episode before starting the next. Generate the storyboard for an episode · Reading the episode cards.

With a long script, later episodes may show Preparing… for a while after the reading; they become ready on their own.

## Adding an episode to the same project

At the bottom of step 3 is an Add an episode card. Paste the script that continues the story, or upload a TXT, DOCX or Markdown file, and press Add and read. Only the new pages are read, priced like a reading of that length; the earlier episodes, their storyboards and clips are not touched. The new episode appears at the end of the list with the characters and scenes it shares with the earlier episodes already attached. If a character, scene or prop first appears in the new pages, a dialog lists it after the reading and offers to generate its sheets, with the price shown; the free sheet allowance applies if you still have it. The new episode's storyboard is priced by its own length and is not covered by the first-episode storyboard allowance.

Let AI continue writes the next part for you: give it a sentence of direction (or none), choose how many episodes, and it picks up right after the last episode with the same cast, in the same language and format as your script. It is priced like AI Create, by the length it writes. You read the draft, change anything, then add it; nothing is read or charged for the reading until you do. Once something has been added, the whole project cannot be read again; to split the same script differently, start a new project with it.

## Starting a second season as a new project

When the next part should stand on its own — a second season, a new style, or a different episode split — start a new project.

- Write the next part of the story, or let AI Create write it from a short brief that recaps where the story stands.
- Start a new project with that text and choose the same visual style, scene style and costume style as the first part.
- In step 2, open each returning character and press Pick existing: choose the character from your first project. The whole entry comes across — name, images, description, voice and state images — and the original is kept, so a replacement can be undone. Reuse a character, scene or voice across projects.
- Check that names match the first part. A character the new script calls by a nickname becomes a new entry; rename it or pick the existing one.

Reuse a character, scene or voice across projects.

## Keeping the second part consistent with the first

Copied sheets carry the faces and costumes across. Scenes can be copied the same way, so a house in season 1 is the same house in season 2. Generate a test segment early in the new project and compare it with the first part before generating everything. Why a whole series stays consistent.

## Joining the parts into one video

Each episode is composited on its own, once every segment in it has a clip. To watch two projects as one film, download each episode and join the files in any video editor. Which segments a composite joins.

| You want | Where | What to do |
| Episode 2, 3… of your script | Step 3 of the same project | Parse Storyboard on that episode's card, then Start Editing |
| More episodes after the end | Step 3 of the same project | Add an episode: paste or upload the continuation, or Let AI continue |
| A second season as its own project | A new project | Write the next part, start a project, Pick existing for the cast |
| A different episode split | A new project | Start a new project with the same script |
| One long video of everything | Your own video editor | Download each episode and join them |

## FAQ

### Can I add an episode to a project I already parsed?

Yes. In SceneMixer, at the bottom of step 3, press Add an episode: paste the continuation, upload a file, or let AI write it. Only the new pages are read.

### Does adding an episode re-read the whole project?

No. Only the appended pages are read and charged on SceneMixer; the earlier episodes keep their storyboards and clips. Once something has been added, the project cannot be re-read as a whole.

### What if the new episode brings new characters?

After the reading, a dialog lists the new characters, scenes and props and offers to generate their sheets, with the price shown. Characters the new pages share with earlier episodes are attached automatically.

### Can I merge the second part into the first project?

No. Projects stay separate; their episodes can be joined after download.

### Is the second project charged again?

Yes. It is read and generated like any project, at the prices on the pricing page.

### Does copying a character cost credits?

No. Pick existing on SceneMixer moves the whole entry over — name, images, description, voice and state images — without a charge, and no plan is needed.

### My next episode says Preparing. Is it stuck?

Long scripts finish preparing later episodes on their own after the reading. Open another episode meanwhile.

Updated November 3, 2026


---

# Making a Multi-Episode AI Series: Order and Continuity | SceneMixer

URL: https://scenemixer.com/guides/making-a-multi-episode-series/
How a series is made: one script read into episodes, one cast shared by all, each episode shot on its own. The order that saves money, and how to add the next episode.

# Making a Multi-Episode Series

Quick answer: A series is one project. The script is read once and split into episodes; the cast, locations and props are made once and shared by every episode. Each episode is then storyboarded and shot on its own, in any order, and composited into its own file. Faces, rooms and props hold across the whole run because they come from the same entries. The order that saves money: finish episode 1 end to end, fix the cast while it is cheap, then run the rest.

A finished sample episode is one of a series made this way.

## One project, many episodes

Upload the whole script. The reading splits it into episodes according to the Type you chose — short episodes with hooks for Short Drama, longer ones for TV Series — and produces one cast list and one set of locations for all of them. A long script keeps preparing its later episodes in the background after the reading. Type · The reading.

## What carries over

Everything in step 2. A character has one sheet, one voice and one set of states, and every episode draws on them; the kitchen in episode 9 is the kitchen from episode 1. Nothing you do inside one episode changes another unless you change the entry in step 2. Why a series stays consistent.

What does not carry over automatically: an image set with Edit (this episode), and per-segment state choices. Those are per episode by design.

## The order that saves money

Storyboard and shoot episode 1 completely first. Watch it. Fix what is wrong at the source: a face in step 2, a voice, a description. Only then parse the remaining episodes. A cast problem found in episode 1 costs one regeneration; found in episode 8 it costs seven episodes of regeneration.

The first episode’s storyboard is free on your first drama, which is another reason to start there. Storyboards.

## Running the rest

Parse several episodes at once from step 3; they run alongside each other. In each episode, generate the segments — several can run at once, up to your plan’s limit — and composite when all are done. Each episode is its own file to download or share. Compositing.

## Continuity between episodes

Within an episode, each segment opens where the last one ended. Between episodes, the cast, rooms and props hold; the story picks up wherever the script does. If episode 4 opens on the aftermath of episode 3, write that state into the first segment’s opening frame and it is honoured. Continuity.

## Adding or removing episodes

The episode split is decided by the reading, but a project can grow at the end: Add an episode at the bottom of step 3 takes the script that continues the story — paste it, upload a file, or let AI write it — and reads only the new pages. The cast and scenes of the earlier episodes are reused, and whatever first appears in the new pages is added to step 2. A second season can also be a new project that reuses the same cast from the asset library. To drop an episode, Delete Episode on its card; the confirmation says its storyboard, videos and composite go with it. Make the next episode · Reusing a cast.

| Step | Do once | Do per episode |
| Read the script | Yes | — |
| Cast, locations, props and their images | Yes | — |
| Storyboard | — | Parse Storyboard |
| Video | — | Generate each segment |
| Composite | — | One file per episode |

## FAQ

### Do I have to generate the cast again for each episode?

No. The cast is made once in SceneMixer's step 2 and shared by every episode.

### Can I shoot episodes out of order?

Yes. Each episode is storyboarded and shot on its own. Finish episode 1 first anyway, to catch cast problems while they are cheap.

### How many episodes can a project have?

As many as the reading finds in the script, within the reading's length limit. A long script keeps preparing later episodes in the background.

### Is there one file for the whole series?

Each episode composites into its own file. Join them in any editor if you want one long file.

### Can I add an episode after the reading?

Yes. Add an episode at the bottom of step 3 in SceneMixer: paste the continuation, upload a file, or let AI write it. Only the new pages are read; the existing cast is reused.

### How do I make the next episode?

If the script already has it, open that episode in step 3 and generate its storyboard. If the story ends where the script ends, press Add an episode and paste the next part, or let AI continue it.

Updated November 3, 2026


---

# Micro Drama vs Short Drama vs Vertical Drama: What Each One Means

URL: https://scenemixer.com/guides/micro-drama-vs-short-drama-vs-vertical-drama/
Micro drama, short drama, vertical drama and mini drama mostly name one format: one-to-three-minute episodes, framed for a phone, sold by the episode. How they differ.

# Micro drama, short drama, vertical drama, mini drama: what each term means

Quick answer: They mostly name the same thing. A micro drama (also short drama, vertical drama or mini drama) is a scripted series of very short episodes — usually one to three minutes, often 60 to 100 of them — framed vertically for a phone, built so every episode ends on a hook, and usually sold episode by episode in an app. The words differ in emphasis: vertical names the 9:16 frame, micro and mini the episode length, short drama is the translation of the Chinese duanju where the format began. An AI micro drama is the same format generated with AI video instead of filmed.

How to write one: micro drama script format, template and example.

## The terms side by side

| Term | What it stresses | Where you hear it |
| Micro drama | Very short episodes | Industry and press in English |
| Short drama | The format as a genre; from Chinese duanju | China-origin apps and their audiences |
| Vertical drama | The 9:16 phone frame | US apps and creators |
| Mini drama | Short episodes, a small production | Some apps and South Asian markets |
| Web series / webisode | Online-first, any length or frame | Older usage; usually horizontal |
| AI micro drama / AI short drama | Made with generated video | Creators and tools |

## What makes a micro drama a micro drama

- Episode length: about one to three minutes; many apps settle near 60–120 seconds.
- Frame: vertical, 9:16, watched one-handed.
- Structure: conflict inside the first few seconds, a turn about every 30 seconds, a hook at the end of every episode.
- Volume: dozens of episodes per title rather than a handful.
- Genres: romance, revenge, billionaire, werewolf and family drama lead. Genre guides.
- Money: usually free first episodes, then pay or watch ads per episode.
The audience is large and growing: Sensor Tower data reported by TechNode put average daily use of short-drama apps at 25 minutes worldwide in April 2026, up 85% from January 2025.

## Micro drama vs a TV series or a film

| | Micro drama | TV series | Film |
| Episode | 1–3 minutes | 20–60 minutes | One piece |
| Frame | 9:16 vertical | 16:9 | 16:9 or wider |
| Pacing | A hook every episode, fast turns | Arcs over an episode and a season | Three acts |
| In SceneMixer, step 1 Type | Short Drama | TV Series | Film |
The Type in step 1 shapes how the script is cut into episodes and how long they run. Short drama, series or film.

## What changes when it is made with AI

An AI micro drama keeps the format and replaces the shoot: a script is read into a cast, locations and an episode outline; each character is drawn once as a reference image; every episode is storyboarded shot by shot; and the shots are generated as video, with the lines spoken on screen, and joined into one file per episode. The hard parts move from the set to the page: faces that stay the same across 80 episodes, and rooms that stay the same rooms. How to make an AI short drama · How consistency works · What a micro drama costs.

## FAQ

### What is a micro drama?

A scripted series of very short episodes, usually one to three minutes each, framed vertically for a phone, with a hook at the end of every episode and often dozens of episodes per title.

### Is a vertical drama the same as a micro drama?

Almost always. Vertical names the 9:16 frame and micro the episode length; in practice both words describe the same app-distributed format.

### What is the difference between a short drama and a micro drama?

Mostly where the word comes from: short drama translates the Chinese duanju, micro drama is the English industry term. The format is the same.

### What is a mini drama?

Another name for the same short-episode format, used by some apps and markets.

### How long is a micro drama episode?

About one to three minutes; many apps settle near 60 to 120 seconds.

### What is an AI micro drama?

The same format made with generated video: the cast, sets and shots are generated from a script instead of filmed.

### Can I make a micro drama myself?

Yes. In SceneMixer, pick Short Drama and 9:16 Portrait in step 1, paste a script or a one-line brief, and each episode is storyboarded, generated and joined into one video.

Updated October 20, 2026


---

# MiniMax H3 (Hailuo) Price per Second: Two Tiers, Their Rules | SceneMixer

URL: https://scenemixer.com/guides/minimax-h3/
MiniMax H3, the Hailuo video model, on this site: what each of its two tiers costs per second, what a segment can hold, and the reference limits that decide what fits.

# MiniMax H3 on SceneMixer: two tiers, priced per second

Quick answer: On this site MiniMax H3 is a two-tier family. MiniMax H3 renders at 2K for 25 credits per second (about $0.16 at the lowest credit price), and MiniMax H3 Fast at 768P for 16 (about $0.10). Segments run up to 15 seconds and always come with sound, and the family’s own reference budget is up to nine images and three voice clips per generation, with a voice clip needing an image alongside it. Design the busy scenes around that, and the rest works like any other tier on the site.

## What sets the family apart

Three things. The picture: the top tier renders at 2K — the larger pixel count of the two — and MiniMax H3 Fast renders at 768P for drafts and phone screens. The cut: a segment can hold several shots with real cuts between them inside one generation, so an exchange and the reaction to it can live in a single render; the storyboard’s shot list is the edit. The sound: every generation includes dialogue and ambient sound — there is no silent mode.

The tiers are picked the usual way: the button in the episode editor’s top bar, or Settings → Model Preferences → Video Model. Whatever you set governs what you render next — a segment already made is untouched, and each version remembers the tier behind it.

## What each tier costs per second

Two tiers on a single SceneMixer bill, charged per second of rendered picture.

| Tier | Price per second, in credits | Cost, at the lowest credit price | Longest run | Resolution |
| MiniMax H3 | 25 | $0.16 | 15 seconds | 2K |
| MiniMax H3 Fast | 16 | $0.10 | 15 seconds | 768P |

What changes between the two is the resolution and the rate; the segment ceiling and the reference budget belong to the family and are shared by both. A draft on MiniMax H3 Fast can be re-rendered as a keeper on MiniMax H3 without touching anything else.

## What you can make with it

Above is a still from The Glass Floor, which this site took from script to composite on Wan 3.0 · 720P. The setting carries the same flow to a MiniMax H3 tier — reference budget permitting.

Where the family earns its place: a short, dense beat — a confrontation that turns, a reveal with its reaction, anything that fits in a single segment and wants the larger picture; and a segment built from a few cuts, since the shots can be cut inside the generation rather than stitched together from separate renders.

There is room for several episodes in a project too: they sit in order, each rendering as its own chain of segments, and any shot inside them can pick its own tier.

## The reference budget, in practice

This family keeps its own numbers: up to nine images and three voice clips per generation, with about 14 s of voice in total. A voice clip also needs an image alongside it — a segment that would carry voices with no image has its voice clips dropped, and each speaker falls back to the written voice description.

When a segment would go over a limit, the site leaves out the least important references first, so the usual failure is a voice changing or a prop losing its sheet rather than a refusal. A segment that names a crowd and attaches a sheet for every face will hit the cap; fewer, more deliberate references carry further here. Segments that need the whole cast attached are better kept on a Wan step, where the budget goes to ten images and five voice clips.

## How to start

Count the cast before the tier switch: a MiniMax H3 segment will take up to nine images and three voice clips, and no more — a scene that wants a sheet for every face may need trimming first, or it belongs on a Wan step. When the count works, Select Model is the switch, and only the segments rendered afterwards move.

Money-wise the family bills like the rest of the site: the read of the script, the reference sheets beyond each account’s free allowance, and the rendered seconds each carry their own charge. The purpose-by-purpose ranking places both H3 tiers, and the H3-or-Wan comparison is the longer read.

## FAQ

### What does MiniMax H3 cost per second?

On SceneMixer: 25 credits per second on MiniMax H3 and 16 on MiniMax H3 Fast — about $0.16 and $0.10 at the lowest credit price. You are charged once the segment lands; a failure sends the credits back, or parks them until support checks the task.

### How many references does a MiniMax H3 segment take?

Up to nine images and three voice clips per generation, with about 14 s of voice in total — and a voice clip needs an image alongside it. When a segment goes over, the site leaves out the least important references first; a character whose voice is dropped speaks from the written voice description instead.

### Can a MiniMax H3 generation run long?

No. Every segment on SceneMixer tops out at 15 seconds on both tiers. For an unbroken longer take, the Wan steps run to 30 seconds — the two families mix happily inside one episode, since each segment keeps its tier.

### Which of the two MiniMax H3 tiers should I pick?

Draft on MiniMax H3 Fast and finish on MiniMax H3 — the pair runs on the same rules (a single generation of up to 15 seconds, the same reference budget, sound in every render) and differs in resolution and rate: 25 credits per second on MiniMax H3 against 16 on MiniMax H3 Fast. Each tier has its own page: the top tier and Fast.

### MiniMax H3 or Wan 3.0?

Come back to Wan for the two things this family cannot do: hold one unbroken take — a Wan step runs to 30 seconds where a segment here closes at 15 — and carry a deep reference list, ten sheets and five voice clips per generation against the nine and three available here. A crowded frame that wants a sheet on every face will shed references on this side; on a Wan step it fits. Stay with MiniMax H3 when the moment is short and dense, when the picture wants 2K, or when the cuts belong inside the segment rather than between renders — MiniMax H3 Fast brings the same texture in at 768P where that is enough. The side-by-side guide works through both directions.

Updated November 3, 2026


---

# MiniMax H3 Fast (Hailuo) Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/minimax-h3-fast-tier/
What MiniMax H3 Fast, the faster Hailuo H3 tier, costs per second on SceneMixer, how it differs from MiniMax H3, and when it is the right place to draft.

# MiniMax H3 Fast: price per second, limits and when to use it

Quick answer: : MiniMax H3 Fast costs 16 credits per second (about $0.10): the 768P sibling of MiniMax H3 with the same 15-second takes, sound, in-generation cuts and reference limits at a lower rate.

The still above is from The Last Envelope, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

MiniMax H3 covers the 2K tier and the reference rules in detail.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

## What it shares with H3

Everything but the resolution: 15-second segments, sound always generated with the picture, several shots with genuine cuts in a single generation, and the reference limits below. A segment you set up on SceneMixer runs on H3 unchanged.

## References it accepts

Per generation, the H3 family’s budget:

## What it costs

16 credits per second, against 25 on H3. The usual pattern: draft the cuts here, regenerate the keeper on H3. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## What it is for

Finding out whether a 15-second segment’s cuts land — the look, the reaction, the insert — before paying the 2K rate. At 768P the blocking and timing are fully visible; fine texture is not the point.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When to switch tier

Up to H3 for the 2K keeper. Across to a Wan or Seedance 2.5 tier for more than 9 image references, or for a 30-second take if the episode’s storyboard was generated with 30-second segments.

| Reference type | Limit on this tier | Note |
| Images | Up to 9 | At least one image is required when any reference is attached |
| Voice clips | Up to 3, about 14 s in total | Voices need an image alongside; with no image the site drops the voice clips and uses the written voice descriptions |
| Over a limit | — | The site leaves out the least important references first; a character whose voice is left out speaks from the written voice description |

## FAQ

### Is Fast a different model from H3?

On SceneMixer it is the 768P tier of the same family, with the same rules and cuts.

### Why did a reference get dropped?

The segment had more than 9 images, more than 3 voice clips or too many seconds of voice, or a voice clip had no image with it.

### Can it render 30-second takes?

No; 15 seconds per segment on SceneMixer.

### Is sound optional?

No. Every SceneMixer render includes sound.

### Can I keep a Fast draft and an H3 keeper of the same segment?

Yes. Every version on SceneMixer is kept; the composite uses the active one.

Updated November 3, 2026


---

# MiniMax H3 (Hailuo) Prompting Guide: Shots, Cuts, References | SceneMixer

URL: https://scenemixer.com/guides/minimax-h3-prompt-guide/
How to prompt MiniMax H3 (Hailuo) on this site: spending the seconds across cuts inside one generation, keeping lines short, and planning the reference budget.

# MiniMax H3 prompting guide: shots, cuts and references

Quick answer: On MiniMax H3 one generation runs up to 15 seconds, and the shots you write inside it are cut together in the same render, so the job is to spend those seconds shot by shot: one framing, one action and one line per shot, with the seconds adding up to the take. Plan the references too — the family takes up to nine images and three voice clips, fewer than the Wan steps.

## What is different about writing for this family

You do not write a prompt for the video model in a separate box. You write the shots of a segment in the storyboard, in the shape every tier shares — marker and seconds, framing and camera, who is in the picture, what happens, the line, what we hear — and SceneMixer turns that segment into one request. What MiniMax H3 changes is the size of that request. A segment tops out at 15 seconds, it always renders with sound, and a handful of cuts can live inside it, so a small exchange and its reaction come back as one piece instead of being stitched from separate renders.

That makes the family a good home for short, dense beats: a confrontation that turns, a reveal with its reaction, an exchange of two lines. It is less comfortable for a scene that wants one long, unbroken take or a large cast, and the sections below say what to do about each.

## Write the shot, then the seconds

The still above is from Rocket Pup, an episode this site took from script to composite on Wan 3.0 · 720P; the shot grammar it was written in is the same one a MiniMax H3 segment uses, and only the size of a take differs.

Each shot opens with its marker and its seconds. The seconds are not decoration: a segment is one generation and its length is the sum of its shots, so the arithmetic has to close before you render. A shot with a line needs room to say it — about three English words to the second — and a shot with nothing but a look can be short. Here is a four-shot segment with room to spare, written the way a segment is written here:

[Shot 1] 4s: Medium shot, fixed camera. @Mara at the head of the table, @Jules on her right. She slides the @Sealed letter across. @Mara (low, steady): Read it before you say no. (voice: 🔊Mara) Rain on the window.
[Shot 2] 3s: Cut to close-up. In frame only: @Jules. He looks down at the letter, jaw tightening. No dialogue.
[Shot 3] 5s: Cut to medium shot, handheld. @Jules pushes back his chair and stands. @Jules (through his teeth): You had no right. (voice: 🔊Jules)
[Shot 4] 3s: Cut to close-up. In frame only: @Mara. She does not look away. No dialogue. Ambient: the clock in the hall.

Notice what each shot does not do: none of them changes framing halfway, none holds two actions, and the shots with no speech say so, so nobody is made to talk. Those habits matter more here than on a take that is never cut, because the cuts are the structure of the render.

## Write the cuts into the shot list

Because the shots of a segment are rendered together, the cut between two of them is made by SceneMixer inside the render, not by you in an editor. Treat each new shot as a cut to a new framing and say so — Cut to close-up:, Cut to wide shot: — and when the same people sit in two shots in a row, change the framing or the angle between them. Without that the join reads as a skip.

The inverse also holds. If you want the camera to stay with one thing, write it as one long shot and give it the seconds it needs; do not chop it into shots just because the family can cut. A single held shot is also how you keep an action unbroken, which is the reason to move a scene to a family with a longer ceiling when the whole beat will not fit in 15 seconds.

## Keep lines short and one speaker per shot

Dialogue and ambient sound are part of every generation on this family; SceneMixer has no silent mode, so a shot with no speech should say No dialogue rather than leave the model to guess. Give each line its own shot when two people are talking and let the reaction be the next shot, instead of packing an exchange into one long line of text. Short sentences also leave less room for a slip.

## Plan the references before you write

This is where H3 segments most often disappoint, and it is arithmetic rather than prompting. A segment takes up to nine images and three voice clips, with about fourteen seconds of voice in total, and a voice clip only works when an image is attached with it. When a segment asks for more, SceneMixer leaves out the least important references first, so the usual symptom is not an error but a voice that changes or a prop that loses its sheet.

So write the segment around the budget. Name only the people who are in the picture. A scene with four speakers has one voice too many — split it into two segments or accept that the extra speaker is voiced from the written description in the delivery note. A crowd does not need a sheet for every face; describe the crowd in words and spend the images on the people the camera will find. The family's reference table lives on its tier page, and MiniMax H3 and its own rules lists it in full.

## When to write this scene for a different family

If one action has to run unbroken for longer than 15 seconds, or the cast is large enough that every face needs its own sheet, the same shots belong on a family with a longer ceiling and a wider reference budget: the Wan steps and the Seedance 2.5 tiers on SceneMixer take up to 30 seconds a generation with up to ten images and five voice clips. The side-by-side is in when to pick MiniMax H3 over Wan; the tier sits next to it as Seedance 2.5 against MiniMax H3.

## FAQ

### How do I write a prompt for MiniMax H3 on this site?

You write shots in the SceneMixer storyboard rather than a freeform prompt: marker and seconds, framing and camera, who is in the picture, what happens, the line and the ambience. The seconds of the shots add up to one generation, which can run up to 15 seconds on MiniMax H3 and MiniMax H3 Fast.

### Can one MiniMax H3 segment hold several shots?

Yes. SceneMixer cuts the shots of a segment together inside a single generation, so an exchange and its reaction can come back as one render. Write each shot as a cut to a new framing and make the seconds add up.

### Why does my voice change between segments on this family?

Usually the reference budget: a segment on SceneMixer takes up to nine images and three voice clips, and when it asks for more the least important references are left out first. Name fewer people in the segment, or split it.

### Does MiniMax H3 Fast take the same shots as MiniMax H3?

Yes. MiniMax H3 Fast is the lower-resolution, lower-price sibling with the same segment ceiling, sound and reference limits, at 16 credits a second on SceneMixer against 25 on MiniMax H3, so a shot list drafted on one renders unchanged on the other.

### When should I write for a longer take instead?

When one unbroken action needs more than 15 seconds, or when the cast needs more than nine images. Both are reasons to move that scene to a family with a longer ceiling.

Updated November 3, 2026


---

# MiniMax H3 (Hailuo) Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/minimax-h3-tier/
What MiniMax H3 (Hailuo) costs per second on SceneMixer, how long one take runs, the references it reads and what sets it apart from the Wan and Seedance tiers.

# MiniMax H3: price per second, limits and when to use it

Quick answer: : MiniMax H3 costs 25 credits per second (about $0.16): takes up to 15 seconds at 2K with dialogue and ambient sound, cuts inside one generation, and reference limits of its own — nine images and three voice clips.

The still above is from The Glass Floor, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

H3 Fast is the 768P sibling with the same rules at a lower rate.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

## What a segment on this tier can be

Length: up to 15 seconds in one generation; the 30-second long-take mode is not available here. Resolution: : 2K, in the project’s aspect ratio. Sound: always on — lines in each character’s voice and ambient sound come out of every render. Cuts: a segment can hold several shots with genuine cuts between them inside a single generation; the storyboard’s shot list is the edit.

## References: the rule that trips people

Two limits to watch: up to 9 images and up to 3 voice clips (about 14 seconds of voice in total). When a segment would go over, SceneMixer leaves out the least important references first. A segment that names many characters and props, plus their voices, can reach the limits; when it does, prefer fewer, better references. A segment keeps at most three voice clips (fewer if they add up to more than about 14 seconds); the other speakers use their written voice descriptions. Voice clips need an image alongside: if a segment would carry voices but no image, the site drops the voice clips and falls back to each character’s written voice description.

## What it costs

25 credits per second. H3 Fast is 16 per second for the same rules at 768P; for comparison the Wan 1080P tier is 38 per second. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## What it is good at, and what to watch

Good at: a 15-second segment rendered sharp enough for a large screen — like the Wan tiers, it can hold an exchange of looks, a reaction and an insert with cuts between them. Watch: the reference cap, the 15-second ceiling, and the fact that this family has its own reference limits, different from the Wan tiers’; a segment that worked on a Wan tier may need its references thinned to run here.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When to switch tier

To H3 Fast for the same rules at a lower rate while drafting. To a Wan or Seedance 2.5 tier when a scene needs more than 9 image references, or an unbroken 30-second take — the latter only if the episode’s storyboard was generated with 30-second segments. Each segment remembers its tier.

| Reference type | Limit on this tier | Note |
| Images | Up to 9 | At least one image is required when any reference is attached |
| Voice clips | Up to 3, about 14 s in total | Voices need an image alongside; with no image the site drops the voice clips and uses the written voice descriptions |
| Over a limit | — | The site leaves out the least important references first; a character whose voice is left out speaks from the written voice description |

## FAQ

### Is 2K higher than the 1080P tiers?

Yes in pixel count. Whether it shows depends on where the clip is watched; on phones the difference is small.

### Why was a reference dropped from my segment?

The segment had more than 9 images, more than 3 voice clips or too many seconds of voice, or a voice clip had no image with it. Name fewer characters or props in that segment and regenerate.

### Can H3 render a 30-second take?

No. On SceneMixer, segments on this tier are up to 15 seconds.

### Does H3 always add sound?

Yes. Every render includes dialogue and ambient sound; there is no silent mode.

### Do H3 segments mix with Wan segments in one episode?

Yes. Each segment remembers its tier and the composite stitches the active versions.

Updated November 3, 2026


---

# My AI Character's Face Changed: A Ten-Point Checklist | SceneMixer

URL: https://scenemixer.com/guides/my-characters-face-changed-checklist/
A shot came back with the wrong face, or a different person altogether? Work down this list: the chip, the unseen name, the missing image, the twin, the stale edit.

# My Character’s Face Changed: A Checklist

Quick answer: A wrong face in one shot almost always has a cause in that shot’s text or in the character’s entry, and almost never in the video tier. Work down the list: is she named with an @ chip? Is someone who is not visible named too? Does she have an image? Do two characters have near-identical descriptions? Did you change her after this shot was made? Is a state or a per-episode image in play? Are there too many people in the frame? Fix the one you find, regenerate that segment, compare the versions.

Why a named character keeps her face explains the rule this checklist enforces.

## 1. Is she named in the shot?

Open the segment and read the shot. The character must appear as an @ chip in the picture text — @Mara turns to the window — not as a typed name and not as “the woman”. If there is no chip, the shot had nothing to hold on to. Add it. When you wrote or edited the segment yourself, the editor also asks before generating if some shots have no linked characters. Who is in frame.

## 2. Is someone unseen named too?

A chip for a character who is not visible — the person she is talking to off screen, someone in the room but out of shot — invites a second face, or a swapped one. Replace the chip with a direction (she looks to frame left) or an off-screen delivery note.

## 3. Does she have an image?

After saving, SceneMixer warns if a named character has no image yet. A character with no image is drawn from her description alone, which is where drift starts. Generate her image in step 2.

## 4. Is there a twin?

Two characters with near-identical descriptions can come out with the same face, and a shot naming one may show the other. Open both in step 2, give each a distinct anchor — a scar, glasses, a hairstyle, an age — and regenerate their images.

## 5. Did she change after this shot was made?

If you edited her description or replaced her image after this segment was generated, the segment still shows the old her. That is not drift; it is an old take. Regenerate it.

## 6. Is a state assigned?

A costume state, a transformed look or an injured version is its own image. Check which state this segment uses; a state generated from a different-looking base can read as a different person. States.

## 7. Is there a per-episode image?

Edit (this episode) on a character in the editor gives that episode its own image. If someone set one, this episode’s shots use it and the rest of the series does not. Check the left panel.

## 8. How many people are in the frame?

Four or more named characters in one shot is where faces start to blend. Split the moment into two shots with fewer people each, or frame on the ones who matter and let the rest be background people in the header.

## 9. Is it a different person altogether?

When the clip shows someone else entirely — another age, another face, another ethnicity — look at her sheet in step 2 first. If the sheet itself is wrong, fix the description or upload a photo, regenerate the sheet, then regenerate the segment. If the sheet is right and the shot names her with a chip, do not keep regenerating: send feedback with the segment number so support can look. When a character looks the wrong age or ethnicity.

## 10. Then regenerate, and only that segment

Fix the one thing you found, save, press Regenerate on that segment. The old take stays in Version History; put whichever is better into the episode. The tier does not enter into it: resolution does not fix a face. Regenerating one segment.

| Check | Where | Fix |
| Named with a chip | The shot text | Add the @ chip |
| Unseen person named | The shot text | Direction or off-screen note instead |
| No image | Warning after save; step 2 | Generate her image |
| Twin description | Step 2, both entries | Distinct anchors; regenerate images |
| Changed after the shot | Step 2 history vs the shot | Regenerate the segment |
| State in play | The segment’s state choice | Check or clear the state |
| Per-episode image | Editor, left panel | Check the override |
| Crowded frame | The shot text | Fewer named people per shot |
| A different person altogether | The sheet in step 2 | Fix the sheet; if it is right, send feedback |

## FAQ

### Will a higher video tier fix the face?

No. Resolution does not change who is drawn. The shot text and the character's entry do.

### Only one shot is wrong. Do I redo the episode?

No. Fix that shot's text and regenerate that segment. The rest is untouched.

### Two characters look alike.

Their descriptions are too similar. Give each a distinct feature and regenerate both images.

### The face is right but the clothes are wrong.

That is a state question, not a face question: assign the right state to the segment.

### I fixed it and it is still wrong.

Regenerate once more; a single take can miss. If it misses repeatedly, reduce the number of people in the shot.

Updated November 3, 2026


---

# "Already Has an Item With That Name": Why Names Are Unique

URL: https://scenemixer.com/guides/name-already-used-why-cast-scene-and-prop-names-must-be-unique/
Every cast member, scene, prop and state has a unique name because storyboards refer to them by name. What the message means, and what a rename changes.

# ‘This Project Already Has a Character, Scene or Item With That Name’: Why Names Are Unique and How Renaming Works

Quick answer: Storyboards refer to cast, scenes and props by name: the tags you insert from the @ menu, the props line, the in-frame list. If two cards shared a name, a shot could not say which one it meant, so a project refuses a second card with a name already in use, and a card refuses a second state with the same name. When a script itself has two characters with one name, the parse gives the second a suffix like ‘(2)’; rename either to something clearer. Renaming a card rewrites every reference to it in this project’s storyboards, so a rename is safe; it does not touch other projects, and it does not change what the clips already made look like.

In Step 2, try to add a character with the name of an existing one: the message appears before anything is created.

## Why names are unique inside a project

Every reference in a storyboard — a character tag in a shot, a location in the header, an item on the props line — is looked up by name to find the card whose sheet becomes a reference image for the clip. A name that matched two cards would attach two sheets, or the wrong one. Uniqueness is per project: two projects can each have a ‘Mei’. The @ Menu in the Storyboard.

## Where the message appears

Adding a character, scene or prop in Step 2 with a name already on another card; renaming a card to a name in use; adding a state to a card that already has a state of that name. (Bringing an entry over from another project is different in one way: it carries its name with it, so if this project already has an entry of that name the replacement is refused before anything changes — rename one of the two and press again.) In each case nothing is created and nothing is charged; choose another name.

## Duplicates in the script itself

Two characters called ‘Guard’, or a scene and a prop both called ‘The Letter’: the SceneMixer parse keeps both and gives the later one a suffix, so no card is lost. Rename them to what the story calls them (‘Guard at the gate’, ‘Guard on the wall’) before generating sheets; the sheets and the storyboards then read cleanly. Re-parsing matches cards by name, so keep the names you settle on.

## What a rename changes

The card’s name and every reference to it in this project’s storyboards, including voice tags; done together so nothing is left pointing at the old name. The old name is not thrown away: unless it is already another card’s name, it moves onto the card as an Also called entry, so text that still uses it keeps matching — nicknames and old names. Not changed: clips already generated (they are pictures), other projects, library copies. Projects created before names were made unique may contain old duplicates; renaming there changes only the card, and the storyboard should be checked by hand.

## Choosing names that work

Short, distinct, as the story says them. Avoid names that are also common words (‘Door’, ‘Night’) for props and scenes; they read ambiguously in shot text. Nicknames can be added in the description; the name is the key.

| You do | Result |
| Add a card with a name in use | Refused; choose another name |
| Rename a card | Card and all references in this project updated |
| Script has two 'Guard' characters | Second becomes 'Guard (2)'; rename both |
| Add a state with a name in use on that card | Refused |

## FAQ

### Can two projects share a name?

Yes. Uniqueness is within one project.

### Does renaming break my storyboards?

No. References are rewritten with the card.

### Do clips change after a rename?

No. They are already generated; only text changes.

### Why did the parse create 'Mei (2)'?

The script has two characters SceneMixer read as 'Mei'; rename one.

### Can a scene and a prop share a name?

No; names are unique across cast, scenes and props in a project.

Updated November 3, 2026


---

# Nano Banana 2.1 and Nano Banana Pro on SceneMixer: Price and Uses

URL: https://scenemixer.com/guides/nano-banana/
Both Nano Banana tiers on SceneMixer, Nano Banana 2.1 and Pro: what each draws, price per image, how to switch them on, and how they fit an AI drama series.

# Nano Banana on SceneMixer: Nano Banana 2.1 and Nano Banana Pro for AI series

Quick answer: SceneMixer offers two Nano Banana models as image tiers: Nano Banana 2.1 (19 credits per image, about $0.12) and Nano Banana Pro (46 credits, about $0.29). They draw the reference images an AI series is matched to — each character on white, each location as an empty plate, each prop, each second outfit or injury — and the promo cover posters; the video tiers then render every shot against those images. Both draw at Google’s 2K setting. Pick either in Settings → Model Preferences → Image Model. Pro keeps more detail from a long English description; 2.1, Google’s October 2026 update of Nano Banana 2, costs far less. The free sheets are drawn on the default model, not on these tiers.

Each tier has its own page: Nano Banana Pro and Nano Banana 2.1; side by side in Pro vs 2.1.

## Which Google models these are

Nano Banana is Google’s nickname for its Gemini image models. The original was Gemini 2.5 Flash Image. Nano Banana Pro is Gemini 3 Pro Image, released at the end of 2025. Nano Banana 2 is Gemini 3.1 Flash Image (preview February 26, 2026; stable May 28, 2026), and Nano Banana 2.1, released on October 6, 2026, is its update, with better prompt following, lettering and consistency across edits according to Google.

SceneMixer offers Nano Banana 2.1 and Nano Banana Pro. In Settings they are listed as Nano Banana 2.1 and Nano Banana Pro.

## What they do on SceneMixer

An AI drama series stays consistent because every shot is generated against the same reference images. SceneMixer reads your script, lists the cast, locations and props, and draws a sheet for each: a full-length casting photo on white for a character, an empty room with its light source readable for a location, the object alone on white for a prop. A second look — the gown at the wedding, the cast on a broken arm — becomes a state image drawn from the main sheet.

The image tier decides how those images are drawn, and it also paints the promo cover — three posters with your title painted in. Nothing else about the project changes when you switch: the same storyboard, the same video tiers, the same price for video.

## How to switch Nano Banana on

Settings → Model Preferences → Image Model, then pick Nano Banana Pro. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

- Open SceneMixer’s Settings from the account menu.
- Go to Model Preferences → Image Model.
- Choose Nano Banana 2.1 or Nano Banana Pro. The price per image is shown next to each tier.
- Back in your project, generate or regenerate the sheets in step 2.

## Pro or 2.1: the short version

- Leads whose faces and costumes carry the story: Nano Banana Pro.
- Supporting cast, locations and props: Nano Banana 2.1, or a cheaper tier.
- Descriptions written in Chinese: a Qwen or SeedDream tier reads them more precisely than the Nano Banana tiers.
- The cover with the title painted in: either works; Google says 2.1 improved lettering. Check the spelling before you pick a candidate.

The full comparison is in Nano Banana Pro vs Nano Banana 2.1.

## Writing descriptions that work on Nano Banana

Both tiers are at their best with plain English sentences that describe what can be drawn. For a character: face shape, brows, eyes, nose, mouth, skin and hair, then one anchor feature nobody else has, then build and posture, then one outfit. Leave out lighting and mood — a sheet is lit flat on purpose — and effect words like beautiful or cinematic, which draw nothing.

Worked examples for a realistic lead, an anime lead and a period costume are in how to write a character sheet description for Nano Banana Pro.

## Price per image

This tier is not covered by the free sheets. Every account’s 15 free sheets are drawn on the default image model, Qwen-Image 3.0, only. On Nano Banana 2.1 each sheet costs 19 credits from the first one (about $0.12 at the lowest credit price), and the price is printed on the button before you generate. Switch back to the default and whatever free sheets you have left are still there. A generation that is refused is not charged, and most failures are refunded automatically; a few kinds are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

Nano Banana Pro is 46 credits per image, the most expensive image tier on SceneMixer. Sheets are a small part of what a series costs; most of the money goes into video seconds, priced on the video tiers.

## What Nano Banana does not do here

- It does not make the video. Shots are rendered on the video tiers; Nano Banana draws the images they use.
- It does not override the project’s visual style. The style is chosen in step 1 and every image tier draws in it.
- It does not change images you already have. Switching tier only affects new and regenerated ones.
- It does not use the free sheets. Those are drawn on the default model, Qwen-Image 3.0.

## Common errors

What you may meet on these tiers:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the description tripped the site’s check, or the model maker’s own filter refused the picture | Edit the description and generate again; nothing was charged |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The model was saturated; the site kept the job queued for a while before giving up | Generate again in a few minutes, or switch tier |
| A sheet that ignores part of the description | Too many details competing in one description | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |
| Credits held for review | A rare failure where the model may already have been paid for the attempt | Nothing to do; support checks it and releases or refunds — see credits held for review |

When Google’s servers are busy, a Nano Banana image can wait in the queue much longer than usual; if it cannot get through, SceneMixer fails it and charges nothing.

## All the image tiers, side by side

| Image tier | Credits per image | In short |
| Qwen-Image 3.0 (default) | 7 | The free sheets; a whole cast cheaply |
| Qwen-Image 3.0 Pro (Recommended) | 18 | Leads in close-up; Chinese descriptions |
| SeedDream 5.0 Lite | 8 | 2K locations and props on a budget |
| SeedDream 5.0 Pro | 25 | 2K, photographic skin, long descriptions read to the end |
| Nano Banana 2.1 | 19 | About 2K; Google’s faster model, the cheapest of the six new tiers |
| Nano Banana Pro | 46 | About 2K; detail kept from long descriptions; the most expensive tier |
| GPT Image 2.5 Flare | 25 | About 2K; OpenAI’s faster general-purpose model |
| GPT Image 2.5 Sunburst | 25 | About 2K; OpenAI’s slower, more precise model |
| Grok Imagine Image 2.0 | 35 | About 2K; xAI’s model, strong on lettering |
| MAI-Image-2.6 | 31 | About 1 megapixel; Microsoft’s model, strict content filter |

## Sources

- Artlist: Nano Banana 2.1, what changed, and which Nano Banana to use — release of Nano Banana 2.1 on October 6, 2026; checked October 25, 2026.
- Google Cloud: Gemini 3.1 Flash Image (Nano Banana 2) — model page and release dates of Nano Banana 2; checked October 20, 2026.
Nano Banana and Gemini are trademarks of Google; they are named here to identify the models offered as tiers. Prices on this page are SceneMixer’s own and are read from the live price list.

| | Nano Banana 2.1 | Nano Banana Pro |
| Shown in Settings as | Nano Banana 2.1 | Nano Banana Pro |
| Google model | Nano Banana 2.1 (update of Gemini 3.1 Flash Image) | Gemini 3 Pro Image |
| Credits per image | 19 | 46 |
| Output size | Google’s 2K setting | Google’s 2K setting |
| Use it for | Supporting cast, locations, props | Leads, detailed costumes and props |
| Free sheets | No — default model only | No — default model only |

## FAQ

### Can I use Nano Banana on SceneMixer?

Yes. Nano Banana 2.1 and Nano Banana Pro are image tiers: choose one in Settings, Model Preferences, Image Model, and every new image is drawn on it.

### How much does Nano Banana cost on SceneMixer?

Nano Banana 2.1 is 19 credits per image and Nano Banana Pro is 46. The live prices are on the pricing page.

### Is Nano Banana free here?

No. Every SceneMixer account's 15 free sheets are drawn on the default image model, Qwen-Image 3.0. On the Nano Banana tiers every image is charged.

### Is Nano Banana 2 still available?

The Nano Banana tier on SceneMixer is now Nano Banana 2.1, Google's October 2026 update of Nano Banana 2.

### What is the difference between Nano Banana Pro and Nano Banana 2.1?

Pro is Gemini 3 Pro Image; 2.1 is the update of Gemini 3.1 Flash Image. On SceneMixer, Pro keeps more detail from long English descriptions and costs more; 2.1 is far cheaper.

### Can I make a whole AI drama with Nano Banana?

Nano Banana draws the sheets and the cover; SceneMixer's video tiers render the shots against those sheets, and the storyboard, dialogue and voices are handled by the rest of the site.

### Does Nano Banana keep characters consistent?

Consistency on SceneMixer comes from drawing each character once and reusing that sheet in every shot. Nano Banana Pro holds a face's anchor details well in close-up, which helps.

### Do I need a Google account or API key?

No. The tiers are part of SceneMixer and billed in SceneMixer credits.

### What resolution do the Nano Banana tiers draw at here?

Both use Google's 2K setting on SceneMixer, for example 1536 by 2752 for a vertical image.

Updated November 3, 2026


---

# Nano Banana 2.1 on SceneMixer: Price per Image and How to Use It

URL: https://scenemixer.com/guides/nano-banana-2-tier/
Nano Banana 2.1, Google's update of Nano Banana 2, for AI series sheets and covers on SceneMixer: where to switch it on, price per image, what changed, when to pick it.

# Nano Banana 2.1 (Nano Banana 2 update) for character, location and prop sheets

Quick answer: Nano Banana 2.1 is Google’s October 2026 update of Nano Banana 2 (Gemini 3.1 Flash Image); on SceneMixer it is the image tier shown as Nano Banana 2.1. It draws a series’ reference sheets — characters, locations, props and states — and the promo cover posters, for 19 credits per image (about $0.12 at the lowest credit price): the cheapest of the six image tiers added in October 2026, at Google’s 2K setting. Use it for supporting cast, locations and props, or a whole English-language cast on a budget; move the leads to Nano Banana Pro if you want more detail kept. The free sheets are not drawn here; they stay on the default model.

Nano Banana Pro is the detailed sibling; Nano Banana on SceneMixer covers both tiers in one page.

## What Nano Banana 2.1 is, and what happened to Nano Banana 2

Google released Nano Banana 2 as Gemini 3.1 Flash Image — a preview on February 26, 2026 and a stable version on May 28, 2026 — and Nano Banana 2.1 on October 6, 2026 as its update. Google lists better prompt following, lettering and consistency across edits, and adjustable thinking levels. Those are Google’s claims; the sources are at the end.

On SceneMixer the Nano Banana Flash-family tier you can pick is now 2.1. If you came here looking for Nano Banana 2, this is the tier that replaces it.

## Where you pick it

Settings → Model Preferences → Image Model, then pick Nano Banana 2.1. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

- Settings, from the account menu.
- Model Preferences → Image Model.
- Pick Nano Banana 2.1; the price per image is next to it.
- Generate or regenerate the sheets in step 2.

## What it is good for here

- Supporting cast and extras on an English-language project, where you want a step up from the default without paying for Pro.
- Locations and props at about 2K.
- Cover posters on a budget: SceneMixer draws the cover on your selected tier, and Google says 2.1 improved lettering. Check the title’s spelling before you pick one.

For the leads’ faces and detailed costumes, Nano Banana Pro keeps more of a long description. For Chinese descriptions, the Qwen and SeedDream tiers read more precisely.

## How to write descriptions for it

Plain English sentences work best: what the person looks like, in drawable terms, in the order face → one anchor feature → build → one outfit. When many decorative details compete, put the identity details first.

Two habits that matter on every tier: describe the cause, not the result (a split lower lip, not looks like she was in a fight), and keep lighting and mood out of the sheet — sheets are plain on purpose.

## What it costs

This tier is not covered by SceneMixer’s free sheets. Every account’s 15 free sheets are drawn on the default image model, Qwen-Image 3.0, only. On Nano Banana 2.1 each sheet costs 19 credits from the first one (about $0.12 at the lowest credit price), and the price is printed on the button before you generate. Switch back to the default and whatever free sheets you have left are still there. A generation that is refused is not charged, and most failures are refunded automatically; a few kinds are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

For comparison: the default Qwen-Image 3.0 is 7 credits per image, GPT Image 2.5 Flare 25, Nano Banana Pro 46.

## When Google is busy

This tier depends on Google’s capacity for the model. When it is tight, an image can wait in the queue much longer than usual — SceneMixer keeps trying for a while rather than failing at once. If it still cannot get through, the job fails and nothing is charged. If you are in a hurry, switch to another tier for that batch; nothing is lost.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the description tripped the site’s check, or the model maker’s own filter refused the picture | Edit the description and generate again; nothing was charged |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The model was saturated; the site kept the job queued for a while before giving up | Generate again in a few minutes, or switch tier |
| A sheet that ignores part of the description | Too many details competing in one description | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |
| Credits held for review | A rare failure where the model may already have been paid for the attempt | Nothing to do; support checks it and releases or refunds — see credits held for review |

## When to switch tier

Up to Nano Banana Pro for the faces and costumes the audience watches most. To GPT Image 2.5 Flare for the same job from another maker. To the default Qwen tier for a large cast on a budget, Chinese descriptions, or to use the free sheets. Nano Banana Pro vs Nano Banana 2.1 goes through the split.

## Sources

- Artlist: Nano Banana 2.1, what changed — release on October 6, 2026 and the changes; checked October 25, 2026.
- Let’s Data Science: Google releases Nano Banana 2.1 — launch coverage; checked October 25, 2026.
- Google Cloud: Gemini 3.1 Flash Image (Nano Banana 2) — release dates of Nano Banana 2; checked October 20, 2026.
Nano Banana and Gemini are trademarks of Google, named to identify the model offered as a tier. Prices are SceneMixer’s own, read from the live price list.

| | Nano Banana 2.1 on SceneMixer |
| Shown in Settings as | Nano Banana 2.1 |
| Google model | Nano Banana 2.1 (update of Gemini 3.1 Flash Image) |
| Used for | Character, location, prop and state sheets; cover posters |
| Credits per image | 19 |
| Output size | Google’s 2K setting |
| Description language | English reads best |
| Free sheets | No — default model only |

## FAQ

### How do I use Nano Banana 2.1 on SceneMixer?

Settings, Model Preferences, Image Model: pick Nano Banana 2.1. Then generate your character, location and prop sheets in step 2 as usual.

### How much does Nano Banana 2.1 cost here?

19 credits per image on SceneMixer, about $0.12 at the lowest credit price.

### Can I still use Nano Banana 2?

The Nano Banana Flash-family tier on SceneMixer is now Nano Banana 2.1, Google's update of Nano Banana 2.

### Is Nano Banana 2.1 the same as Nano Banana Pro?

No. 2.1 is the update of Gemini 3.1 Flash Image; Pro is Gemini 3 Pro Image. On SceneMixer, Pro keeps more detail from a long description and costs much more.

### Is Nano Banana 2.1 free on SceneMixer?

No. The 15 free sheets every account gets are drawn on the default image model only.

### Why is my Nano Banana 2.1 image taking so long?

When Google's capacity for the model is tight the image waits in the queue. If it cannot get through, SceneMixer fails it and charges nothing; switch tier if you are in a hurry.

### What resolution does it draw at?

SceneMixer uses Google's 2K setting, for example 1536 by 2752 for a vertical image.

### Does switching tier redraw my sheets?

No. Images already drawn on SceneMixer stay as they are; only new and regenerated ones use the new tier.

### Which is better for anime styles?

The look is a project setting chosen in SceneMixer’s step 1 and every tier draws in it; switching tier does not switch style. See stylised and anime looks across tiers.

Updated November 3, 2026


---

# Nano Banana Pro Character Sheet Prompts: How to Write Them

URL: https://scenemixer.com/guides/nano-banana-pro-character-sheet-prompts/
Character sheet prompts for Nano Banana Pro on SceneMixer: the order to write a face, the one detail that makes it unique, three worked examples, and what to leave out.

# How to write a character sheet description for Nano Banana Pro

Quick answer: On SceneMixer you do not write the whole image prompt for a character sheet; you write the character’s appearance, and the site draws it as a full-length casting photo on white with plain light. For Nano Banana Pro, write that appearance in plain English, in this order: face (shape, brows, eyes, nose, mouth, skin, hair, age as it shows), then one anchor feature nobody else has, then build and posture, then one outfit, piece by piece. Leave out lighting, mood, camera words and effect words like beautiful or cinematic. Pro rewards specific, drawable detail: it keeps more of a long description than the cheaper tiers.

What the tier is and costs: Nano Banana Pro on SceneMixer.

## Where the description lives

When SceneMixer reads your script it writes an appearance for every character it finds. Open the character card in step 2 and you will see it under the sheet. That text is what the image tier draws from, and it is yours to edit: sharpen the face, add the anchor, fix the outfit, then regenerate that one sheet.

The framing is not your job. Every character sheet on the site is the same kind of picture — one full-length figure on white, lit flat enough to show the bones — because every shot of the series will be matched to it. Why sheets are plain explains the reason.

## The order: face, anchor, build, outfit

| Part | What to write | Example words |
| Face | Shape, brows, eyes, nose, mouth, skin as a person has it, hair line and cut, the age you see | long face, heavy straight brow, deep-set eyes, high-bridged nose, freckled, grey at the temples |
| Anchor | One thing nobody else in the cast has | a scar through the left eyebrow, a chipped front tooth, a white streak of hair |
| Build and posture | Height, frame, how they stand | tall and lean, shoulders forward; short and broad, upright |
| Outfit | One outfit, named piece by piece | navy peacoat over a cream turtleneck, dark trousers, black boots |

Face first matters most on Nano Banana Pro precisely because it reads so much: if the coat is described in eight clauses and the face in one, the sheet frames for the coat. Describe a face so it is not a generic face goes deeper on faces.

## Example 1: a realistic lead

A woman in her late thirties with a long face, a straight heavy brow and deep-set grey eyes; a thin high-bridged nose and a downturned mouth; weathered olive skin with fine lines at the eyes; dark brown hair cut bluntly at the jaw, grey at the left temple. A pale scar runs from the right corner of her mouth toward the ear. Tall and lean, shoulders slightly forward. A navy wool peacoat over a cream turtleneck, dark straight trousers, black lace-up boots.

Why it works: every word can be drawn. The scar is the anchor, the grey at one temple is a second small identifier, the posture gives the figure an attitude without asking for a pose, and the outfit is one outfit with four named pieces — the kind of detail Pro keeps.

## Example 2: an anime lead

A boy of about sixteen with a round face, large upturned amber eyes and thick dark eyebrows; messy black hair with one white streak over the left eye; a small bandage across the bridge of his nose. Short and wiry, stands with his weight on one leg. A school uniform: dark green blazer with a gold crest on the left breast, loosened red tie, grey trousers, scuffed white trainers.

The style (anime, painterly, 3D) is set once for the project in step 1, so the description does not say anime style. It describes the person; the project style decides how he is drawn. The white streak and the nose bandage are anchors that survive a stylised look.

## Example 3: a period costume

A man in his fifties with a square jaw, heavy-lidded brown eyes and a broad nose; close-cropped grey beard; a burn scar on the back of the left hand. Broad and heavy-set, upright posture. A Tang-dynasty official's robe: deep crimson round-collared gown, a black gauze cap with two stiff wings, a jade-plaque belt at the waist, black boots.

Period costume is where Pro is worth its price: a gown, a cap with a specific shape, a belt with a specific material. Name each piece and what it is made of; the costume will be drawn piece by piece rather than as a generic “ancient robe”.

## What to leave out

- Lighting and mood. Dramatic rim light, moody, neon: whatever light is on the sheet is carried into every shot. Write the light per shot in the storyboard.
- Effect words. Beautiful, handsome, striking, masterpiece, 8K, cinematic describe nothing drawable. No magic words explains why they backfire.
- Results instead of causes. A hint of red at the nostrils is painted as a red nose. Write the cause: a cold belongs in the storyboard, not on the sheet. See describe the cause, not the result.
- A second outfit. Usually overalls, a white gown at the wedding gives a gown with straps. One outfit per sheet; the gown is a state. See one outfit per character sheet.
- Real people and known characters. A name or an iconic outfit trips the IP check on every tier.

## Second looks: states, not a longer description

For the wedding gown, the hospital gown, the bandaged arm: add a state on the character card, describe only what is different, and generate it from the main sheet. The face stays the main sheet’s face. Then pick that state for the scenes where it applies. Costume changes and scene states walks through it.

## Fixing a sheet that came out wrong

- Generic face: the face part is too short. Add bones and an anchor; regenerate.
- Costume dominates, face small: the outfit is described at length. Cut the outfit to its main pieces.
- Two outfits blended: move the second to a state.
- Odd light or colour cast: a mood word slipped in. Delete it.
- Blocked by the text check: a word reads as violent, explicit or a real name. Rephrase in neutral visual terms; SceneMixer does not charge a blocked sheet.

To change one thing about a sheet you already like, edit only that thing — write the change, not the whole picture.

## Price

Each Nano Banana Pro sheet costs 46 credits, from the first one: the 15 free sheets every account gets are drawn on the default image model, Qwen-Image 3.0, not on this tier. Regenerating a sheet is a new charge, so a good description the first time is also the cheap one. A practical order: get the wording right on the default tier, where the free sheets are, then switch to Nano Banana Pro and regenerate only the leads.

## FAQ

### What is the best prompt for a Nano Banana Pro character sheet?

On SceneMixer, a plain English appearance in the order face, anchor feature, build, one outfit. The site frames it as a full-length sheet on white, so you do not add layout or lighting words.

### Should I ask for a turnaround or several angles?

Not here. A character sheet is one full-length figure; the video tiers use it from every angle. Second looks go on state cards.

### Do I write 'anime style' in the description?

No. The style is chosen for the whole project in SceneMixer's step 1 and every sheet is drawn in it.

### How long should a character description be?

Long enough to make the person unmistakable, usually three to six sentences. Pro reads all of it, but keep the face ahead of the outfit.

### Can I write the description in Chinese?

You can, but Nano Banana tiers read English more precisely. For Chinese descriptions, a Qwen or SeedDream tier is the better pick.

### Why does my character look generic?

Usually the face is described with effect words. Replace them with drawable features and one anchor detail.

### How do I keep the character consistent across episodes?

Draw the sheet once and keep it. Every shot is matched to it; changes of outfit or injury go on states generated from it.

### Can I upload a photo instead of writing a description?

Yes, with Upload & Replace on the character card. Keep the description in line with the photo, because it still steers how shots are written.

Updated November 3, 2026


---

# Nano Banana Pro on SceneMixer: Price per Image and When to Use It

URL: https://scenemixer.com/guides/nano-banana-pro-tier/
Nano Banana Pro for character, location and prop sheets and cover posters on SceneMixer: price per image, where to switch it on, what it does well, when not to use it.

# Nano Banana Pro (Gemini 3 Pro Image) for character, location and prop sheets

Quick answer: On SceneMixer, Nano Banana Pro is the image tier listed as Nano Banana Pro. It draws the reference sheets a series is built on — characters, locations, props and their states — and the promo cover posters, at 46 credits per image (about $0.29 at the lowest credit price), which makes it the most expensive image tier on the site. It is Google’s detailed model: more of a long English description survives into the picture — a costume with several named pieces, a prop with a specific mechanism, a face that holds its identity in close-up — and it draws at Google’s 2K setting. The free sheets are not drawn on this tier.

Nano Banana 2.1 is the cheaper Nano Banana tier; Pro vs 2.1 compares the two side by side.

## What Nano Banana Pro does on this site

SceneMixer uses an image tier to draw sheets: a full-length casting photo of each character on white, an empty location plate with its light source readable, each important prop alone on white, and a state image for every second look (the wedding gown, the bandaged arm). When a shot is generated, the sheets of everything in it are handed to the video model as “this is what these look like”, and the model copies them closely. That is where consistency across a whole series comes from. The same tier also paints the promo cover — three posters drawn from your leads’ sheets with the title painted in.

So the question for an image tier here is not “how pretty is one picture” but “how much of what I wrote about this person, room or object ends up on the sheet”. That is the axis Nano Banana Pro is strong on.

## Where you pick it

Settings → Model Preferences → Image Model, then pick Nano Banana Pro. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

## Where the extra detail shows

More of the description survives into the sheet. In practice that shows up in three places:

- Costume with many parts. A period outfit described piece by piece — collar, sash, the buttons on one side — keeps more of its pieces.
- Props with a mechanism. A pocket watch with an engraved lid, a folding knife with a brass bolster: the specific part is drawn, not a generic object.
- Faces in close-up. The anchor you gave the face — a scar through the eyebrow, a chipped tooth — holds, and the face reads as the same person from sheet to sheet.

It draws at Google’s 2K setting (1536 by 2752 for a vertical image), so the detail has the pixels to sit in.

## How it reads your description

English sentences, read to the end. Write the character the way a casting director would: face first, then one unmistakable anchor, then build and posture, then one outfit. Chinese descriptions are rendered, but read less precisely than on the Qwen and SeedDream tiers; if your project’s descriptions are in Chinese, those tiers are the better pick.

As on every tier, describe causes rather than results: write a burn scar on the back of the left hand, not a hint of red. And leave mood out: sheets are deliberately plain, white background and even light, because whatever light is on the sheet is carried into every shot. How to write a character sheet description for Nano Banana Pro has worked examples.

## What it costs

This tier is not covered by the free sheets. Every account’s 15 free sheets are drawn on the default image model, Qwen-Image 3.0, only. On Nano Banana Pro each sheet costs 46 credits from the first one (about $0.29 at the lowest credit price), and the price is printed on the button before you generate. Switch back to the default and whatever free sheets you have left are still there. A generation that is refused is not charged, and most failures are refunded automatically; a few kinds are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

For comparison on SceneMixer: Nano Banana 2.1 is 19 credits per image, GPT Image 2.5 Sunburst 25, SeedDream 5.0 Pro 25, Qwen-Image 3.0 Pro 18. A cast of six leads drawn once each on Pro is six images; the money in a series goes into video, not sheets, so the detailed tier is usually worth it for the faces the audience watches most — and only for those.

## Content checks

Every description is checked before the sheet is drawn; a flagged description is not charged. Next to it, an IP check refuses sheets that resemble a real person, a brand mark or a character from a known film, game or comic, on every tier; Google’s own safety filter also runs on this one. See content rules and checks and IP suspected: celebrity, brand or franchise.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the description tripped the site’s check, or the model maker’s own filter refused the picture | Edit the description and generate again; nothing was charged |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The model was saturated; the site kept the job queued for a while before giving up | Generate again in a few minutes, or switch tier |
| A sheet that ignores part of the description | Too many details competing in one description | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |
| Credits held for review | A rare failure where the model may already have been paid for the attempt | Nothing to do; support checks it and releases or refunds — see credits held for review |

## When to switch tier

Down to Nano Banana 2.1 for locations and props where fine detail matters less. To GPT Image 2.5 Sunburst for state images and covers drawn from an existing sheet, at about half the price. To a Qwen tier for Chinese descriptions or the lowest price. Which image tier for your sheets lays the whole choice out in one table.

## Sources

- Artlist: Nano Banana 2.1, what changed, and which Nano Banana to use — where Nano Banana Pro sits next to 2.1; checked October 25, 2026.
Nano Banana and Gemini are trademarks of Google, named to identify the model offered as a tier. Prices are SceneMixer’s own, read from the live price list.

| | Nano Banana Pro on SceneMixer |
| Shown in Settings as | Nano Banana Pro |
| Google model | Gemini 3 Pro Image |
| Used for | Character, location, prop and state sheets; cover posters |
| Credits per image | 46 |
| Output size | Google’s 2K setting |
| Strength | Detail and fidelity to long English descriptions |
| Free sheets | No — default model only |

## FAQ

### Can I use Nano Banana Pro on SceneMixer?

Yes: pick Nano Banana Pro in Settings, Model Preferences, Image Model, and every new sheet and cover is drawn on it.

### How much does Nano Banana Pro cost here?

46 credits per image on SceneMixer, about $0.29 at the lowest credit price. A description blocked by the content check is not charged.

### Is Nano Banana Pro free?

No. The 15 free sheets on every account are drawn on the default image model; this tier is charged from the first image.

### Is Pro higher resolution than Nano Banana 2.1?

Both draw at Google's 2K setting on SceneMixer. Pro's difference is how much of a detailed description it keeps.

### Does Pro keep my character consistent across episodes?

Each character is drawn once and every video shot is matched to that sheet. Consistency comes from the sheet being reused, on any tier; Pro makes the sheet itself closer to what you wrote.

### Can I make video with Nano Banana Pro?

No. It is an image tier for sheets and covers. Video is rendered on the video tiers.

### Should every character be on Pro?

Rarely. Put the leads on Pro and the rest of the cast, the locations and the props on a cheaper tier; switch the setting between groups.

### Can it draw the title on my cover?

Yes. The cover is drawn on your selected image tier and the model paints the title in. Check the spelling before you pick a candidate.

Updated November 3, 2026


---

# Nano Banana Pro vs Nano Banana 2.1: Which to Use for Sheets

URL: https://scenemixer.com/guides/nano-banana-pro-vs-nano-banana-2/
Nano Banana Pro vs Nano Banana 2.1 on SceneMixer: price per image, what Pro keeps that 2.1 drops, and a split that puts the money where the audience looks.

# Nano Banana Pro vs Nano Banana 2.1 (and 2): which tier for which sheet

Quick answer: On SceneMixer, Nano Banana Pro costs 46 credits per image and keeps more of a long, detailed description — many-part costumes, props with a mechanism, the anchor details of a face in close-up. Nano Banana 2.1, Google’s October 2026 update of Nano Banana 2, costs 19 and is the tier for everything that is not a lead. Both draw at Google’s 2K setting, so the difference is in what the picture contains, not how many pixels it has. A good split: Pro for the two to four leads, 2.1 for the rest of the cast, locations and props.

Both tiers on one page: Nano Banana on SceneMixer.

## The two side by side

| | Nano Banana Pro | Nano Banana 2.1 |
| Shown in Settings as | Nano Banana Pro | Nano Banana 2.1 |
| Google model | Gemini 3 Pro Image | Nano Banana 2.1 (update of Gemini 3.1 Flash Image) |
| Credits per image | 46 | 19 |
| Output size | Google’s 2K setting | Google’s 2K setting |
| Strength | Detail from long English descriptions | Price; Google cites better prompt following and lettering than 2 |
| Free sheets | No — default model only | No — default model only |

## What Pro keeps that 2.1 may drop

The difference shows when a description has many specific, drawable parts:

- A costume described piece by piece — the stand collar, the sash knotted on the left, the row of cloth buttons. Pro keeps more of the pieces.
- A prop with a mechanism — the hinged lid of a pocket watch, the brass bolster of a folding knife.
- A face’s anchor detail in close-up — the scar through one eyebrow, the chipped front tooth.

For a plain description — a modern accountant in a grey suit, a kitchen, a coffee cup — the two come out close, and 2.1 costs a fraction on SceneMixer.

## The split that saves money

Take a series with four leads and twenty-four supporting characters, locations and props. Pro for everything: 28 images at 46. Pro for the leads and 2.1 for the rest: 4 at 46 and 24 at 19. The audience spends most of its time on the leads’ faces, so that is where the expensive tier earns its price.

None of these images use the free sheets: every account’s 15 free sheets are drawn on the default model, Qwen-Image 3.0 at 7 credits. If you want to use them, draw the supporting cast there first and move only the leads to a Nano Banana tier.

## How to switch between them

Settings → Model Preferences → Image Model, then pick Nano Banana Pro. From then on every new image is drawn on it: the character, location and prop sheets in step 2, regenerations, state images, and the promo cover posters. Images already drawn keep their pixels; switching never redraws anything. You can switch per group — leads on one tier, background places on another — and switch back.

- Settings → Model Preferences → Image Model: choose Nano Banana Pro.
- Generate the leads in SceneMixer’s step 2.
- Switch the setting to Nano Banana 2.1.
- Generate the rest. Already-drawn leads stay as they are.

## Beyond the two Nano Banana tiers

For state images and the promo cover — both drawn from an existing sheet — GPT Image 2.5 Sunburst (25 credits) is built to hold a reference subject; see GPT Image 2.5 vs Nano Banana Pro. For Chinese descriptions, the Qwen and SeedDream tiers read more precisely than either Nano Banana tier.

## Sources

- Artlist: Nano Banana 2.1, what changed, and which Nano Banana to use — checked October 25, 2026.
Nano Banana and Gemini are trademarks of Google, named to identify the models offered as tiers. Prices are SceneMixer’s own, read from the live price list.

## FAQ

### Is Nano Banana Pro better than Nano Banana 2.1?

The Pro tier on SceneMixer is Google's detailed model and the one to use when a description has many named parts. For plain descriptions the two come out close and 2.1 is far cheaper.

### Which is cheaper?

Nano Banana 2.1, at 19 credits per image on SceneMixer against Pro's 46.

### Does Pro have more pixels?

No. Both use Google's 2K setting on SceneMixer.

### What happened to Nano Banana 2?

The Nano Banana Flash-family tier on SceneMixer is now Nano Banana 2.1, Google's October 2026 update of Nano Banana 2.

### Can I mix them in one project?

Yes. SceneMixer’s setting applies to new images only, so draw the leads on Pro, switch, and draw the rest on 2.1.

### Which is better for anime?

The style is chosen per project on SceneMixer, and both tiers draw in it. Pro helps when an anime lead's costume has many specific parts.

### Which for locations?

Usually Nano Banana 2.1, or SeedDream 5.0 Lite when you want 2K rooms on a budget.

### Do the free sheets work on either?

No. The 15 free sheets per account are drawn on the default image model only.

Updated November 3, 2026


---

# Native Dialogue vs. Dubbing in AI Video: Why It Matters | SceneMixer

URL: https://scenemixer.com/guides/native-dialogue-vs-dubbing-ai-video/
Dubbing an AI drama means re-timing every shot and fighting the lips. Generating the line in the target language avoids both: timing, lip sync and cost.

Guide · Languages

# Native-Language Dialogue vs. Dubbing in AI Video (2026): Why the Cast Should Speak the Line

Published September 12, 2026

Quick answer: Dubbing replaces the audio of a finished clip; the picture was generated for a different line, so mouth shapes, pauses and shot length all belong to the original language. Native-language generation gives the video model the line in the target language before it renders the shot, so the performance, the pacing and the cut points are built for that language. For AI short drama the second approach is cheaper too: there is no dubbing pass, and a wrong line is fixed by regenerating one segment.

## What dubbing has to fight

| Problem | With dubbing | With native generation |

| Mouth movement | Belongs to the original language; lip-sync tools approximate | Generated with the line; audio and picture come from one pass |

| Shot length | Fixed by the original take; a longer translation gets rushed or cut | Computed from the target language's speaking rate before the shot is generated |

| Pauses and breath | Inherited from the original performance | Placed where the target-language line needs them |

| Emotion | Performed by the voice actor or a TTS voice, against an unrelated face | Performed by the same generation that draws the face |

| Cost of a fix | Re-record and re-align the line; sometimes re-cut the shot | Regenerate one segment; only that segment is charged |

## Speaking rate decides the cut

The same sentence is not the same length in every language. SceneMixer's storyboard step derives shot lengths from speaking rate — about 3 words per second for English and about 5 characters per second for Chinese, with the other languages mapped onto those scales — and it does this per dialogue language, before any video is generated. A dubbed workflow cannot: its shot lengths were fixed by the source language, which is why dubbed short drama so often sounds rushed in German and padded in Chinese.

## How native generation is implemented here

- The line is written in the target language at analysis time and copied verbatim into the storyboard; the storyboard step is forbidden from rewriting dialogue, so what you approved is what is performed.

- Voice identity is anchored. Each character is matched to a preset voice sample that is attached to every segment, so the same character sounds the same across episodes regardless of language.

- The video model renders audio and picture together. The models SceneMixer uses generate speech with the shot, so there is no separate TTS or alignment step.

- A guard checks the language before credits are frozen. If a segment's lines drift out of the project's dialogue language, the submit stops and asks you to confirm.

## When dubbing still makes sense

- You already have a finished series and want one more market quickly, and lip-sync is not critical for the platform (audio-off autoplay feeds).

- You need a specific human voice under contract.

- The target language is not among the 16 SceneMixer can generate dialogue in.

For everything else — a new series, a market launch, a pilot — generate the performance in the language. See the 15-language overview and how to set the dialogue language and localize the cast.

Languages

## Native dialogue in 15 languages

Every series below was produced by the same pipeline with the cast speaking that language natively: the interface, the working documents and the spoken lines are all in one language, and nothing is dubbed. Open one to hear it.

- A Carta de LisboaPortuguêsNative dialogue
- 윈터 프라미스한국어Native dialogue
- El Secreto de CostaEspañolNative dialogue
- 浪人の誓い日本語Native dialogue
- Ikrar JakartaBahasa IndonesiaNative dialogue
- Зимняя коронаРусскийNative dialogue
- L’Héritier du DéfiléFrançaisNative dialogue
- Il Tavolo dell'OlivaItalianoNative dialogue
- De Vuurtoren van de FjordNederlandsNative dialogue
- العهد الصحراويالعربيةNative dialogue
- Çantasındaki SözleşmeTürkçeNative dialogue
- Die Istanbul-TäuschungDeutschNative dialogue
- Останній сигналУкраїнськаNative dialogue
- 問劍青雲繁體中文Native dialogue
- The Last EnvelopeEnglishIn your language
See all 15 languages

## FAQ

### Is native-language generation the same as AI dubbing?

No. AI dubbing replaces the audio of a finished clip and then tries to match the lips. Native-language generation gives the video model the line in the target language before the shot exists, so mouth movement, pauses and shot length are produced for that language.

### Why does timing matter so much?

Because the same sentence takes a different amount of time in each language. SceneMixer computes shot lengths from the speaking rate of the chosen dialogue language before generating video; a dubbed clip is stuck with the source language's shot lengths.

### Does the character keep the same voice across languages?

Each character is matched to a preset voice sample that is attached to every segment, so the character's voice identity is stable across episodes. Across languages the timbre stays recognisable while the performance is generated in the new language.

### What does it cost compared with dubbing?

There is no dubbing pass to pay for. Segments on SceneMixer cost the same credits in any language, and a wrong line is fixed by regenerating that one segment.


---

# No Magic Words: Why 'Masterpiece, 8K, Cinematic' Hurts AI Video

URL: https://scenemixer.com/guides/no-magic-words-masterpiece-8k-cinematic/
Keyword lists came from older image tools. Today's models read natural sentences; stacked tags dilute the description and sometimes get painted. What to write instead.

# No Magic Words: Why “Masterpiece, 8K, Cinematic, Volumetric” Do Nothing Here (or Worse)

Quick answer: The models behind SceneMixer read descriptions as sentences, not as tags. “Masterpiece, best quality, 8K, ultra-detailed, cinematic lighting, volumetric fog” adds nothing the model can place in the frame, pushes the words that matter further down where they are read less carefully, and occasionally gets painted literally — fog appears because you asked for it. Write what is in the scene in plain sentences; pick the look with the visual style preset; say what a real cinematographer would say: where the light is, what the lens sees.

Search your descriptions for the words masterpiece, 8K, HDR, cinematic, volumetric, subsurface; delete them and regenerate one sheet to compare.

## Where the words came from

Early image tools were steered by tags, and communities collected lists of words that seemed to help: “masterpiece”, “trending”, “8K”, “subsurface scattering”. The models used here were trained on captions and scripts written in ordinary language and are steered by sentences. The tag habit survives on forums; it does not survive contact with these models.

## Three ways the tags hurt

They dilute. Descriptions have a practical length; the model reads the first part most carefully and the tail loosely. Twenty tags in front push the scar and the lamp to the tail. They get painted. “Volumetric fog” brings fog; “lens flare” brings a flare across the face; “bokeh” blurs the room you wanted consistent. They fight the preset. The visual style preset already says live action or anime or 3D. “Photorealistic” inside an anime project asks for two things at once and gets a muddle. How to Choose Visual Style Settings.

## What to write instead

The cinematographer’s vocabulary, in sentences: the light source and its side, the lens feel (“a long lens, background soft”), the framing (“medium shot, she is frame left”), the surface (“wool coat, matte”, “skin with visible pores, no retouching”). Every one of these is a thing the model can place. The full principle: Describe the Cause, Not the Result.

## Negative lists do not work either

“No people, no text, no watermark, no extra limbs” is a list of nouns, and a list of nouns is an invitation. The models attend to what is named, not to the “no”. Say the positive: “an empty street”, “a plain wall behind her”, “only the two of them in frame”. The site’s own instructions are written this way for the same reason.

## Where quality actually comes from

On SceneMixer, resolution comes from the video tier you pick; look comes from the style preset; identity comes from the sheets; consistency comes from the storyboard. Tags touch none of these. Which tier for which purpose.

| Tag | What actually happens | Write instead |
| masterpiece, best quality | Nothing | Nothing; delete it |
| 8K, ultra HD | Nothing; resolution is the tier | Pick 720P or 1080P in the selector |
| cinematic lighting | Nothing, or random contrast | A single window on the left; a practical lamp on the desk |
| volumetric fog / god rays | Fog and rays appear | A layer of smoke in the room, one small window |
| no people, no text | People and text more likely | An empty platform; a blank sign |

## FAQ

### Do the tags at least not hurt?

They dilute the description and can be painted literally. Deleting them is a free improvement.

### What about 'photorealistic'?

The style preset already decides that. Inside a live-action project it is redundant; inside a stylised one it fights the preset.

### Can I ask for a specific camera or film stock?

Say what it looks like: shallow depth of field, soft highlights, slight grain. Brand names add little.

### Are Chinese tag lists any different?

Same effect. '杰作、超高清、电影感' does what 'masterpiece, 8K, cinematic' does.

### Why does the site not just strip these words?

Some are harmless and some are yours on purpose; SceneMixer trims what it can and leaves your text yours.

Updated November 3, 2026


---

# Non-Human Characters in AI Video: Fruit, Animals, Robots | SceneMixer

URL: https://scenemixer.com/guides/non-human-characters-fruit-animals-robots/
How to make a fruit drama, an animal lead or a robot: what to write so the sheet is a creature and not a person in a costume, which styles suit it, and how voices work.

# Non-Human Characters: Fruit, Animals, Robots

Quick answer: A non-human character is a character like any other: it has a card in step 2, a sheet, a voice and an @ chip. What makes it a creature rather than a person in a costume is the description. Say what it is — a ginger tabby, a talking orange with a face, a battered delivery robot — and whether it has a face, hands or clothes. A real animal is drawn as an animal, on all fours, no hands; a fruit with a face is drawn with a face. Pick an animation style for a fruit or toy drama; live action for a real animal. Dialogue and voices work the same way as for people.

Browse the samples for animated projects; a fruit drama uses the same steps as any of them.

## The fruit drama

The format that is everywhere at the moment: fruit with faces living out a romance or a rivalry, a reality show in a bowl. On SceneMixer it is an ordinary project. Write the script with the fruit as characters — names, personalities, lines — and pick an animation style in step 1 (3D Toy Cartoon, Family 3D Animation or Anime suit it). The reading casts each fruit as a character and gives it a card.

## What to write on the card

The description decides what the sheet shows. Say what the thing is and how much of a person it has. A plump orange with large round eyes, a small mouth, thin stick arms and no legs; sits on the counter gives you a cartoon orange with a face. A ginger tabby cat, green eyes, white bib, thick winter coat gives you a cat — drawn as a real cat, on four legs, no hands, no clothes, because the description says nothing else. If you want the cat in a waistcoat, say so.

In SceneMixer, clothing goes in the Clothing field only if the creature wears any. Leave it empty for a real animal; an empty field on a person is a problem, on a cat it is correct.

## Real animals

A dog, a horse, a hawk: describe the breed or species, the colouring, the size, one mark. The sheet shows the animal as itself, standing naturally, seen from a useful angle. It is still portrait-shaped like every character sheet. In the storyboard, name it with @ like anyone; say what it does in plain words — @Biscuit jumps onto the table, tail up.

## Robots and objects that talk

Describe the machine: shape, size, material, colour, what moves, whether it has a face or a screen, how it gets around. A robot has whatever hands you give it in the description. A talking object — a lamp, a phone — is a character if it speaks and a prop if it does not; cast it accordingly.

## Voices and lines

Every character gets a voice in step 2, creature or not: matched from a voice description, picked from the library, or uploaded. Write a voice description that fits — small, bright, childlike for the orange; flat, metallic, unhurried for the robot. Lines are written the same way as for people, with a voice tag; a creature with a mouth moves it, one without does not. Voices · Lines.

## Mixed casts

A human and a talking cat in the same shot is fine: name both with @, say who is where. The scene’s automatic AI 3D staging places them like any two characters. Keep the creature’s size in the description — knee-high, fits in a palm — so the shot scales it right.

| You want | Write | Style |
| A cartoon orange with a face | The fruit, the face, the limbs it has, where it sits | 3D Toy Cartoon, Family 3D Animation, Anime |
| A real cat | Breed, colouring, size, one mark; Clothing empty | Digital HD or any animation |
| A robot | Shape, material, what moves, face or screen, how it moves | Any |
| A talking lamp | Cast it as a character; describe the lamp | Any |

## FAQ

### Can I make a fruit drama on SceneMixer?

Yes. Write the fruit as characters, pick an animation style, and follow the same steps as any drama.

### Will my cat character be drawn as a person in a cat costume?

Not if the description says it is a cat and gives it no clothes. The sheet follows the description.

### How does a fruit get a voice?

Like any character on SceneMixer: a voice description in step 2, then match, pick or upload.

### Can a creature and a person share a shot?

Yes. Name both with @ and give the creature a size in its description.

### Does a non-human character cost more?

No. SceneMixer prices images and video the same for every character.

Updated November 3, 2026


---

# The Notes Under a Storyboard Segment: What Each One Means

URL: https://scenemixer.com/guides/notes-under-a-segment-what-each-one-means/
Six notes can sit under a segment: no character image, crowd image, state images, video edited, shot somewhere else, in frame. Which block a video and which only inform.

# The notes under a storyboard segment, and what to do about each

Quick answer: Under a segment’s storyboard text, above the buttons (Edit, AI edit, History, Insert), the editor can show up to six one-line notes. Three of them ask for an image before the video will be right — no character image yet, crowd image not generated, and state images not generated; only the last of those stops the video. Three are information: Video edited, In frame, and Shot N is somewhere else, which does stop the video. Nothing on this page costs credits just by being shown; every button on it shows its price first.

Time and weather are shown in the header above the shots, not here: the segment header.

## Where the notes are

Open an episode and select a segment. The storyboard card shows the header and the numbered shots; right under the text, in a thin strip, come the notes. If nothing is wrong there is no strip at all. A note disappears by itself once its cause is fixed — for images, as soon as the picture has been drawn.

The State images, In frame, Video edited and Shot N is somewhere else notes belong to the current storyboard layout and do not appear on older segments or in ad and remake projects. The no character image and crowd image notes can appear on any segment. Which layout is mine?

## At a glance

| Note | Blocks the video? | What to do |
| In this segment but no character image yet: … | No | Press Generate images (it takes you to step 2) |
| This segment declares a background crowd, but the crowd reference image for its scene has not been generated yet. | No | Press Generate crowd image, or ignore it |
| This segment needs state images that haven’t been generated yet. Generate them before making the video: | Yes | Generate the listed images, then the video |
| Video edited | No | Nothing: later segments continue from the edited clip |
| Shot N is somewhere else | Yes | Rewrite the shot, or give it its own segment |
| In frame · Shot 1 … | No | Read it; change the text if it is wrong |

## No character image yet

In this segment but no character image yet: Mara, Leo. The video can only go by the text description, so the face won’t match the character.

A character who is in the picture has no image to copy. The usual causes: the reading judged them “voice only” and the batch skipped them; they were added by a later episode; or their image failed and nobody retried. The video can still be made, but it will invent a face from the words. Press Generate images: it opens step 2, where the price, the free allowance and the member gates are all shown. Nothing is generated from the note itself. Leads, supporting and voice-only characters · Face changed? Checklist.

## Crowd image not generated

This segment declares a background crowd, but the crowd reference image for its scene has not been generated yet. with a Generate crowd image button and a price tag.

A scene that has a crowd in it gets a second reference picture, the scene with people in it. Until it exists the video falls back to the empty scene plus the words, so the crowd may be thin or missing. It is a suggestion, not a block. The button uses the same generator and the same free allowance as step 2; once the picture lands, the note goes away on its own. Background people in a scene.

## State images not generated

This segment needs state images that haven’t been generated yet. Generate them before making the video: followed by the list, each with a price and a generate button.

A state is a character’s or scene’s other look: the torn coat, the night version of the room. When the segment uses a state that has no picture yet, the video would have nothing to copy, so it is refused with These state images haven’t been generated yet: … Generate them in the asset’s states first — no credits were charged. Generate the listed images (they run through the same generator and the same small state-image allowance as step 2) and press Generate again. Use a character state in one scene · State image messages.

A state needs its character’s main image first; if the main image is still being drawn, the list waits for it.

## Video edited

Video edited — The storyboard text wasn’t changed — later segments continue from the edited video.

This note means the segment’s current clip was trimmed with Trim or changed with AI visuals. What you see no longer matches the storyboard text word for word, and that is on purpose: the next segments carry on from what is on screen, not from the text. A new voice (redub) or a quality enhancement does not trigger it, and switching back to the original take removes it. Edit this clip.

## Shot N is somewhere else

The one note that also blocks the video. A segment on SceneMixer is one place and one timeline; a shot that cuts to another place or time has to be rewritten or moved to its own segment. The full guide.

## In frame

In frame · Shot 1 Mara, Leo · Shot 2 Mara · Shot 3 Leo

For each shot, who the video will show, worked out from where everyone is in the room, how wide the shot is and who it is on. It appears when the scene has been built as a 360° room, and it is the same list the video is generated from, so it is the place to catch a shot that will show the wrong person. If a name is missing, make the shot’s subject clearer in the text: say who it is on, and how wide. If a name is there that should not be, make the shot tighter. A dash means nobody named is in that shot. Who is in frame · The 360° room.

## FAQ

### Where are the notes under a segment?

Right under the storyboard text of the selected segment, above the Edit and AI edit buttons. If nothing is wrong, there is no strip.

### Which notes stop me from generating the video?

Two: state images that have not been generated, and a shot that is somewhere else. The others are information or a suggestion.

### Does a note cost credits?

No. Showing a note is free on SceneMixer. A button on a note shows its price before it does anything, and the generate buttons use the same allowance as step 2.

### The note will not go away.

It goes when its cause is gone: the image is drawn, the shot is rewritten, the clip is switched back to the original. For images, wait for the picture to land; the note checks again by itself.

### Why is my character flagged with no image when I generated everything?

A character added after you generated, a character first read as voice only, or an image that failed. Press Generate images to go to step 2.

### What does Video edited do to later segments?

They continue from the edited clip. Redub and quality enhancement do not count; switching back to the original removes the note.

### Why does In frame not show on my segment?

It needs a segment in the current layout and a scene built as a 360° room. Older segments, ads and remakes do not show it.

### Can I hide the notes?

No, but every one of them goes away once you do what it says.

Updated November 3, 2026


---

# One Outfit per Character Sheet: Where the Second Outfit Goes

URL: https://scenemixer.com/guides/one-outfit-per-character-sheet/
A clothing field listing two outfits gets a sheet with half of each. Why, how the site splits labelled outfits, and how to add a second look as a state.

# One Outfit per Character Sheet: Why Two in One Description Fails and Where the Second One Goes

Quick answer: The clothing text of a character sheet must describe one outfit. Given ‘early episodes: work overalls; wedding: white gown’, the model paints both at once — a gown with overall straps, or a figure with three arms — because nothing tells it these are two moments. The site helps: when the parse writes labelled sections (Early: … Later: …), the first is kept on the sheet and the others become state cards; labelled sections you type in later are trimmed so the sheet draws only the first. Do not rely on it; write one outfit, and add the others as states, each with its own image. Keep the sentence short: a long coat described in detail makes the model frame for the coat and shrink the face.

Open a character the parse gave a second look: the state card beside the sheet holds the second outfit, and the sheet shows only the first.

## Why two outfits in one description go wrong

The model draws everything the text names, in one picture. Two outfits are two sets of shapes competing for the same body; the result blends them, and the blend is what every later shot is matched to. The same applies to ‘usually X but sometimes Y’, ‘in flashbacks she wears’, and ‘after the fire his clothes are burnt’: each of those is a state, not part of the base description.

## What the site does with labelled outfits

Text of the form Label: outfit. Label: outfit. is split when the parse saves the character: the first outfit stays, each further labelled outfit becomes a state card (a second look with its own image, selectable per scene). Typed into the clothing field later, the extra outfits are only left off the sheet; no state card is made, so add those states yourself. Labels that name a part (Accessories:, Footwear:) are treated as part of the same outfit, not a new one. This is a safety net for scripts that describe several looks at once; the reliable way is to write the states yourself.

## Adding a second look as a state

On the character card, add a state: name it (wedding), describe only what differs (the gown, hair up, no glasses), generate its image from the main sheet. Then pick that state for the scenes where it applies; other scenes keep the main outfit. Costume Change and Scene States, Use a State in One Scene.

## Keep the outfit sentence short

One sentence: what it is, its cut, its colour, one distinguishing detail. A paragraph on the coat’s embroidery makes the sheet a fashion shot: full-length, small head, the face — the thing that has to stay consistent — lost. Face first, clothing last. Describe a Face.

## What goes in the storyboard instead

Costume details that matter in a particular scene (the torn sleeve, the medal pinned on) go into that segment’s text, where the clip is generated; the sheet stays the neutral base everything is matched to.

| Clothing text says | Sheet shows | Do instead |
| Overalls; wedding: white gown | A gown with straps, or two half-figures | Overalls on the sheet; 'wedding' as a state |
| Usually a suit, sometimes casual | A suit jacket over a hoodie | Suit; 'casual' as a state if it appears |
| Long embroidered coat (three sentences) | Small head, coat fills the frame | One sentence on the coat |
| After the fire, burnt clothes | Burnt clothes from episode one | Clean clothes; 'after the fire' as a state |

## FAQ

### The parse wrote two outfits into one character; is that my fault?

No; some scripts describe both. Check the state cards: the split may already be done. If not, move the second outfit to a state.

### Do states cost credits?

Each state image on SceneMixer is a generation at the sheet price; a free allowance applies in the first project.

### Can a state change more than clothes?

Yes: hair, injuries, a transformation. Anything that changes the look, not the position.

### Which outfit do the clips use?

The state chosen for that scene, or the main sheet if none is chosen.

### My character's coat hides the face on the sheet.

Shorten the clothing sentence and regenerate; the face is the priority.

Updated November 3, 2026


---

# "Output Was Cut Off": When a Parse or Storyboard Stops Halfway

URL: https://scenemixer.com/guides/output-was-cut-off-parse-or-storyboard-failed/
A text model has a fixed output budget; a very long or dense script can run past it. The site retries once; if that fails too, the credits are held for review.

# “Output Was Cut Off” or “Invalid JSON”: When a Parse or a Storyboard Stops Halfway

Quick answer: Parsing and storyboards are written by a text model that has a fixed budget for how much it can write in one go. A very long script, or an episode dense with dialogue, can run past that budget and the answer arrives incomplete. The site detects this and runs the same job once more. If the second answer is also incomplete, the task is marked failed and its credits stay held: the model was called, so support checks it and then refunds or settles it; it is not refunded automatically. Before running again: for a parse, trim the script or split it into separate projects.

After a failed parse or storyboard, open your credit history: that task’s line says Under review until support settles it.

## What 'cut off' means

The model writes a structured answer — the cast list and episodes for a parse, the segments and shots for a storyboard. When it reaches its output limit mid-answer, the structure is incomplete and cannot be used; the message reads cut off, truncated or invalid JSON. Nothing partial is kept; your script and any earlier storyboard stay as they were. What each task status means: Task statuses.

## What the site does first

SceneMixer retries every cut-off answer once automatically with the same input; many second attempts complete. For a storyboard, the better of the two drafts is saved if it is usable at all. Only when neither attempt gives a usable answer does the task fail, and then the held credits wait for review instead of coming back on the spot. Refund rules.

## A parse that keeps cutting off

Long scripts are read in two passes precisely to avoid this; the threshold is automatic. A parse that still cuts off is usually a script with an unusually large cast and many locations. Trim it, or split it into two projects. Long Scripts Are Read in Two Passes and the caps.

## A storyboard that keeps cutting off

An episode is written as segments and shots; an episode whose share of the script is very long, or wall-to-wall dialogue, produces more shots than one answer can hold. An episode cannot be split or its outline edited on SceneMixer, so generate its storyboard once more (a later run often completes); if the same episode keeps failing, write to support with the project and episode. In your next project, keep each episode’s share of the script moderate. Storyboard Checks: Two Drafts.

## Rarer relatives

Field type mismatch: the model wrote a number where a word belonged; the site repairs what it can and retries the rest. Wrong language: the model ignored the project language; retried once with a firmer instruction. All of these end the same way — a second attempt, and if that also fails, a failed task whose credits are held for review — and none of them changes what you already have.

| Task | Usual cause of a cut-off | What to do |
| Parse | A script with a very large cast and many locations | Trim it or split it into two projects |
| Storyboard | A very long or dialogue-dense episode | Generate once more; write to support if it keeps failing |
| AI Create | A target length past the parse cap | Pick a shorter target |

## FAQ

### Were my credits refunded?

Not automatically. After a second failed attempt SceneMixer keeps the credits held; support reviews the task and refunds or settles it.

### Do I lose my script?

No. The script, the summary and any existing storyboard are untouched.

### Will running it again just fail again?

A parse with nothing changed often does; trim or split first. A storyboard often completes on a later run.

### Why does a longer script cost the same but fail?

Price is by length; the model's output budget is a separate limit, which is why long scripts read in two passes.

### Can I keep the half that was written?

No. A cut-off answer is structurally incomplete and is discarded whole.

Updated November 3, 2026


---

# Pay with Alipay, WeChat Pay or Your Local Currency (Airwallex)

URL: https://scenemixer.com/guides/pay-with-alipay-wechat-pay-or-local-currency/
Since October 2026 checkout has a third tab, processed by Airwallex: local methods such as WeChat Pay and Alipay, and one membership period that does not auto-renew.

# Pay with Alipay, WeChat Pay or in Your Local Currency

Quick answer: Since October 2026 the payment window on scenemixer.com has a third tab next to card and PayPal, processed by Airwallex. It is the one showing local payment logos, such as WeChat Pay and Alipay. Prices on the site stay in US dollars; on the Airwallex form you can pay in your local currency, and any conversion fee is charged by Airwallex. For credits, press Continue to payment and pick a method. For a membership there are two rows: the local-method row buys one period that does not renew; Card / Google Pay is a subscription that renews automatically. Which methods appear depends on where you are, and WeChat Pay is not offered on phones and tablets.

Open Buy Credits from your balance at the top right, choose a pack, and look at the tabs in the payment window: the third one carries the local-method logos. Nothing is charged until you finish paying.

## The three tabs in the payment window

The payment window on the international site has a row of tabs, each shown by its logos:

- Card, Apple Pay or Google Pay — card brands and wallets, in an embedded form.
- PayPal.
- Airwallex — the logos of the local methods available to you, for example WeChat Pay or Alipay. The bottom of the tab reads Processed by Airwallex.
Apple Pay is only on the card tab; the Airwallex tab never offers it. The Chinese site, scenemixer.cn, has its own payment methods and none of these tabs. Buy credits, step by step.

## Which methods you will see

The Airwallex tab shows only the methods that can take a payment from you: a method appears when it accepts US dollars or your local currency, judged from the country you are connecting from. As of October 2026:

| Where you are | Local methods on the Airwallex tab |
| Australia | Alipay (mainland China and Hong Kong), WeChat Pay |
| Mainland China | Alipay, WeChat Pay |
| Hong Kong | AlipayHK, WeChat Pay |
| Elsewhere | WeChat Pay |
- On a phone or a tablet, WeChat Pay is not offered. Paying it means scanning a code, which you cannot do with the same phone; use a computer, or another method.
- Other local methods can appear later for other regions; the tab always shows the ones that work for you at that moment.
- When buying credits, the Airwallex form can also offer a card and Google Pay alongside the local methods.

## Buy credits with a local method

- Open the window. Click your balance at the top right, then Buy Credits, and choose a pack. The summary shows what you pay, in US dollars.
- Choose the Airwallex tab. Before anything loads it shows a short summary and a note: The next page shows the local currencies and payment methods available where you are. Any currency conversion fee is charged by Airwallex.
- Press Continue to payment. The order is created now, and the form loads: Loading the secure payment form…
- Pick a method. The local methods come first. The form shows the amount in the currency you will pay.
- Pay. WeChat Pay shows a QR code to scan with the WeChat app on your phone; Alipay follows its own confirmation.
- Wait for the confirmation. The window says Payment received. Confirming…, then shows your new balance. If confirming takes longer: Airwallex is still confirming this payment. Credits arrive automatically once it completes; check your order history in a moment.
The order appears in Order History on the billing page with Airwallex as its channel.

## A membership: one period, or a subscription

Choose a plan and a cycle on the membership page and press subscribe; on the Airwallex tab the window lists payment methods as rows:

| Row | What you buy | Renews? |
| The local methods, e.g. Alipay / WeChat Pay | One period of the plan and cycle you chose, paid once | No. When the period ends the plan ends; buy again to continue |
| Card / Google Pay | A subscription; under it, a line such as (price)/month · Auto-renews, cancel anytime | Yes, until you cancel |
One period, no renewal: pay with the local method in the form. The window says Payment received. Activating your membership…; if it takes longer, Airwallex is still confirming this payment. Your membership activates automatically once it completes. The plan’s credits arrive like any other membership. Nothing is charged later; there is nothing to cancel.

Subscription: Renews automatically through Airwallex. Cancel any time on the membership page; after cancelling, access continues until the current period ends. Once you start an Airwallex subscription checkout for a plan, the window stays on that tab: An Airwallex subscription checkout has been started for this plan. To pay another way, close this window and choose the plan again.

The one-period row only appears when a local method is available to you; on a phone outside the regions where Alipay is offered there may be none, and only the subscription row shows. Subscribe, upgrade or cancel a plan.

## One provider renews your plan

A membership renews through one provider at a time. If your plan renews automatically through Airwallex, plan changes go through Airwallex and the other tabs say so: Your membership renews through Airwallex. Please use Airwallex for plan changes. The same rule applies the other way round when your plan renews by card or PayPal. To switch, cancel the renewal, let the period end, then subscribe with the other method. Payment methods by site.

## Currency, fees and refunds

- SceneMixer sets the price in US dollars. The Airwallex form converts it to your currency; the conversion fee, if there is one, is Airwallex’s, as the tab says before you continue.
- Your statement shows the amount in the currency you paid.
- Refunds follow the same rules as any other payment and go back to the method you paid with. Refund rules.
- Card details typed into the Airwallex form go to Airwallex: Card details are encrypted and never reach our servers.

## If the payment does not go through

| You see | What to do |
| The payment form didn’t load. Please try again. with Try again | Press Try again; if it keeps failing, use the card tab or PayPal. |
| Couldn’t open the Airwallex payment page. Please try again. | Close the window and start again. |
| Airwallex couldn’t complete the payment. Please try again or use another payment method. | Try again, or another method or tab. |
| The Airwallex payment wasn’t completed. Choose the plan again to retry. | Pick the plan again on the membership page. |
| This Airwallex payment page has expired. Please start again. | Start a new payment. |
| Couldn’t confirm the Airwallex payment just now. Please try again in a moment. | Wait a moment; if you paid, it is confirmed automatically. Check Order History before paying again. |
| One-time purchase isn’t available right now. Choose auto-renew or another payment method. | Use the subscription row, or another tab. |
Never pay twice for the same thing while one payment says it is still being confirmed. Payment did not go through · Previous payment still being processed.

| You want | Use |
| Credits, paid with WeChat Pay or Alipay | Airwallex tab → Continue to payment → the local method |
| One month of a plan with no renewal | Membership → Airwallex tab → the local-method row |
| A plan that renews by itself | Card tab, PayPal, or Airwallex → Card / Google Pay |
| Apple Pay | The card tab |
| WeChat Pay on a phone | Not offered; use a computer or another method |

## FAQ

### Can I pay with Alipay or WeChat Pay on the international site?

Yes, since October 2026, on the Airwallex tab of the SceneMixer payment window, where your region allows it. WeChat Pay is offered on computers, not on phones or tablets.

### Can I pay in my own currency?

Prices are set in US dollars. On the Airwallex tab the form converts the amount to your local currency; any conversion fee is charged by Airwallex.

### Can I buy one month of membership without a subscription?

Yes. On the Airwallex tab of the membership checkout, the local-method row buys one period of the plan and cycle you chose, with no automatic renewal. When it ends, buy again if you want to continue.

### Why is there no WeChat Pay on my phone?

WeChat Pay here works by scanning a QR code, which cannot be done with the same phone, so it is not offered on phones and tablets. Use a computer, or another method.

### Why do I only see WeChat Pay and not Alipay?

Methods appear by the country you are connecting from. As of October 2026, Alipay is offered to customers in mainland China, Hong Kong and Australia; elsewhere the tab shows WeChat Pay.

### Where is Apple Pay?

On the card tab. The Airwallex tab does not offer Apple Pay.

### My plan renews by card. Can I pay the next month with Alipay?

Not while the card subscription is renewing: plan changes go through the provider that renews the plan. Cancel the renewal, let the period end, then buy the next period on the Airwallex tab.

### The window says the payment is still being confirmed. Should I pay again?

No. Credits or the membership arrive automatically once Airwallex confirms. Check Order History on the billing page after a moment.

### Is Airwallex available on the Chinese site?

No. scenemixer.cn has its own payment methods.

Updated November 3, 2026


---

# Payment Failed? Declined Card, PayPal, Missing Credits | SceneMixer

URL: https://scenemixer.com/guides/payment-did-not-go-through/
What to do when the card form does not load, the card is declined, PayPal shows pending, or you paid and the credits are not there, and how to resume the same order.

# When a Payment Does Not Go Through

Quick answer: First look at your balance at the top right. If it went up, the payment landed, whatever the window said. If the card form never appeared, the window says so after a few seconds with a retry that reopens the same order. A declined card is the bank’s decision: try another card, PayPal, or a local method on the Airwallex tab. A PayPal payment that shows still clearing arrives by itself when it completes. An order you did not pay waits under Pending Orders for a short while and then expires without charging anything. If money left your account and the credits did not arrive within an hour, send the order number to support.

Buying credits, step by step is the happy path; this page is every way it can stall.

## The form did not load

The card form is loaded from the payment provider. If it has not appeared after fifteen seconds, the window says The payment form didn’t load. Check your network connection and try again with a retry that reopens the same order — no second order is created. An ad blocker or a strict privacy extension can block the form; allow the site or try another browser. The PayPal button has the same message; the Airwallex tab says The payment form didn’t load. Please try again. with Try again.

## The card was declined

The decline comes from your bank, and the form shows the bank’s reason where it gives one. Usual causes: a card that does not allow online or foreign-currency payments (prices are billed in USD), a spending limit, or a mistyped number. Try another card, a wallet if your device shows one, PayPal, or a local method such as Alipay or WeChat Pay on the Airwallex tab (how it works). Nothing is charged on a decline.

## PayPal

Cancelled. Nothing was charged: you closed the PayPal window. PayPal couldn’t complete the payment: pick another funding source inside PayPal. PayPal is still clearing this payment: some PayPal payments take longer to settle; the credits arrive automatically once it completes, with no need to pay again. The Airwallex tab has the same case: Airwallex is still confirming this payment. Credits arrive automatically once it completes; check your order history in a moment.

## I paid, the credits are not there

Check the balance at the top right and Credit History: a Credit purchase line means it landed. The success window can fail to draw even when the payment went through, so the balance is what counts. If nothing appears within an hour and the money has left your account, write to support with the order number from Order History on the credits page; the payment record is matched by hand. Never pay a second time to check.

## The price changed while the window was open

If a pack or plan price changed between opening the window and paying, the window says The price is now ✦ N. Tap again to continue and nothing is charged until you do. The same applies to an upgrade whose difference changed.

## Unpaid and duplicate orders

An order you did not pay sits under Pending Orders for a short while so you can resume it at the same price. Reopening the purchase window resumes the same order rather than creating a second. Unpaid orders on SceneMixer expire on their own and never charge. If you see two paid orders for one intended purchase, write to support: duplicates are refunded. The automatic renewal of a plan is a different kind of payment failure: it puts a Payment failed chip on your plan card, covered in Renewal Payment Failed: How to Restore Your Plan.

| You see | It means | Do |
| Form did not load | Provider script blocked or slow | Retry in the window; check ad blocker |
| Card declined | The bank refused | Another card, wallet, PayPal or the Airwallex tab |
| PayPal still clearing / Airwallex still confirming | Settling | Wait; credits arrive by themselves |
| Cancelled, nothing charged | You closed the window | Reopen and pay |
| Paid, balance unchanged after an hour | Needs matching | Support with the order number |
| Price is now ✦ N | Price changed meanwhile | Tap again at the new price |

## FAQ

### The card form never appears.

Wait for the retry the window offers; it reopens the same order. Ad blockers and privacy extensions are the usual cause.

### My card was declined. Was I charged?

No. SceneMixer charges nothing for a decline. Try another card, PayPal, or a local method on the Airwallex tab.

### PayPal says the payment is still clearing.

Some PayPal payments settle later. The credits arrive automatically when it completes.

### I paid and have no credits.

Check the balance and Credit History first. If nothing shows within an hour, send the order number to support; do not pay again.

### Will an unpaid order charge me later?

No. It waits briefly under Pending Orders so you can resume it, then expires.

### The card number is right, but it says the card is invalid.

The card may not be accepted for this kind of payment. Nothing was charged. Try another card, or pay with PayPal or on the Airwallex tab.

Updated November 3, 2026


---

# Payment Methods by Site: Card, PayPal, Alipay, and Order Records

URL: https://scenemixer.com/guides/payment-methods-by-site-card-paypal-alipay-wechat/
The international site takes cards and PayPal in US dollars; the Chinese site takes Alipay in yuan. Which renew memberships and where your order records are.

# Payment Methods by Site: What You Can Pay With, in Which Currency, and Where the Receipt Comes From

Quick answer: On the international site you pay by card (in an embedded checkout), with PayPal, or on a third tab processed by Airwallex that offers local methods such as Alipay and WeChat Pay. Prices are in US dollars; on the Airwallex tab you can pay in your local currency. All three work for credit packages and memberships; a subscription renews through whichever provider you subscribed with, while a local-method membership on the Airwallex tab is one period that does not renew. On the Chinese site you pay with Alipay, in yuan; memberships there do not renew automatically and simply end when the term does. A membership can be managed by only one provider at a time: if it renews through your card, plan changes go through the card, and the same for PayPal. The site’s order history on the billing page shows every order, its number and its state; the payment provider may also send its own confirmation.

Open the buy-credits window: the methods shown are your site’s; the price is in your site’s currency and does not switch with the interface language.

## What each site offers

International: card (Visa, Mastercard and others accepted by the checkout), PayPal, and a third tab processed by Airwallex with the local methods available where you are, such as Alipay or WeChat Pay; prices in USD, and on the Airwallex tab you can pay in your local currency (any conversion fee is charged by Airwallex). Pay with Alipay, WeChat Pay or in your local currency. Chinese site: Alipay, paid by scanning a code; CNY. Methods and currency follow the site you are on, not your location or interface language. Step by step: Buy Credits.

## Memberships and renewal

International: a card, PayPal or Airwallex subscription renews each cycle until you cancel; cancelling keeps benefits to the end of the paid term. Paying with a local method on the Airwallex tab buys one period that does not renew. Chinese site: a membership is a one-time payment for a term; when it ends you subscribe again if you like; there is nothing to cancel. Subscribe, Upgrade, Cancel.

## One provider holds the subscription

If your membership renews through your card, upgrades and cancellation go through the card; PayPal is refused for plan changes with a message saying so, and the reverse; the same holds for an Airwallex subscription. This keeps one renewal, not two. To switch providers, let the current term end, then subscribe with the other.

## Unfinished payments

Closing the checkout on SceneMixer leaves the order pending for a while; reopening within half an hour for the same package resumes the same order rather than opening a second one. A payment that needs your bank’s confirmation returns to the site to check the result before showing success; if it shows ‘still processing’, wait a moment and check the order history. Payment Did Not Go Through.

## Receipts and invoices

The order history on the billing page lists every order with its number, amount, state and what it granted; PayPal also shows the payment in your PayPal account. For company invoices, contact support with the order number.

| Site | Methods | Currency | Membership renewal |
| International | Card, PayPal, Airwallex (local methods) | USD; local currency on the Airwallex tab | Automatic until cancelled; a local-method purchase is one period, no renewal |
| Chinese | Alipay | CNY | None; ends with the term |

## FAQ

### Can I pay in my own currency?

Prices on SceneMixer are in US dollars. With a card or PayPal, your bank or PayPal converts. On the Airwallex tab you can pay in your local currency; any conversion fee is charged by Airwallex.

### Can I switch my subscription from card to PayPal?

After the current term ends; one provider at a time.

### Where is my receipt?

The order is in your billing history on SceneMixer; PayPal payments also appear in your PayPal account.

### Does the Chinese site charge automatically?

No. Memberships there are one-time payments per term.

### Can I pay on my phone?

Yes — the payment window works in a phone browser the same way; there is nothing to install, and no separate store purchase path.

### Can I pay with Pix or another local method?

The payment window lists the methods available for your order. The Airwallex tab shows the local methods that work where you are, such as Alipay or WeChat Pay; if a method is not listed there, it cannot be used for that payment.

Updated November 3, 2026


---

# Photo to Video AI: Turn Photos Into a 15 to 60 Second Ad

URL: https://scenemixer.com/guides/photo-to-video-ai/
Turn photos into a short ad: which four photos to take, how each kind of subject is treated, which vibe and length to pick, what the end card says and what it costs.

# Photo to Video AI: From Photos to a Finished Ad

Quick answer: Photo to video AI works best when you know what the photos are for. Here, up to four photos of one product, dish, shop, car, property or app screen become a 15, 30 or 60 second ad with a script, spoken lines, music written for the ad and an end card carrying your brand. The photos are read for free before anything is charged, the first one is the subject, and the one price on the button is a cap, not a bill. This guide is the whole path: a shoot list for the photos, what the read does to each kind of subject, which vibe and length to choose, and where to fix a shot afterwards.

## A shoot list for four photos

| Photo | Show | Why it helps |
| 1. The hero | The thing itself, filling most of the frame, in focus | It is the subject: the ad and the end card are built around it. For a product, dish or garment it is the one redrawn as a clean shot |
| 2. Another angle | The same thing from the side or above | Extra photos become the places and details the shots are made in |
| 3. A detail | A label, a texture, steam, a stitch | A detail the shots can be made around |
| 4. In use or in place | The shelf, the table, the street it lives on | Becomes a location, so the ad happens somewhere |

Put the hero first. SceneMixer takes up to four JPG or PNG files, each under 10 MB, of something you have the rights to. If the read cannot make out what is being advertised, it says so on the spot and nothing is charged: too blurry, too tiny, or too ambiguous are the usual reasons. Which photos work for an ad.

## What the read does to each kind of subject

| Subject | What happens to the photo | Why |
| Product, dish, garment | Redrawn as a clean studio shot on white | So it can be lit the same way in every segment |
| Shop, restaurant, car, house | Kept as it is; any people in it are removed first | Cutting a place out of its surroundings destroys the signage, depth and scale; a model given a face copies it, so bystanders are removed |
| App or website screenshot | Used as it is, never redrawn; one screen only, no real face | Redrawing turns readable interface into mush |

Then a dialog says what SceneMixer’s read understood: a category (Product, Food & drink, Apparel, Vehicle, Store / property, App, Website or Service) and a name field you can correct. If you typed one line about what it is and who it is for, the line wins when it disagrees with the photo: “our family restaurant” makes the photo a place, not a product. Nothing is charged until you choose how to make the ad. How to make an ad from one photo.

## Pick the vibe for the subject

| Vibe | What it does | A fit worth trying |
| Epic | Scale, spectacle, a big reveal | A launch, a car, a place |
| Funny | A joke with the product as the punchline | Everyday products |
| Warm | People, feeling, a small moment | Family business, gifts |
| Minimal | The product, light, silence, one line | Skincare, jewellery, design |
| Twist | It looks like one thing, then it is another | Anything with a reveal |
| Mouth-watering close-ups | Food and drink up close: steam, pour, cut | Dishes, drinks |
| Show how it works | Use it, see the result | Apps, services, gadgets |
| A day in the life | Someone living with the product | Shops, clothes, a room |

The vibe also sets the look of the ad and the colours of its end card. It does not change the price on SceneMixer; length and quality do. Vibe, length, format and quality.

## Length: one, two or four segments

SceneMixer makes an ad in 15-second segments joined in the final cut with the end card at the end. 15 seconds is one segment and a single beat: the product and one idea. 30 is two. 60 is four, and is a small story. Every segment carries at least one spoken line, and the tagline is said out loud before the final freeze. Nothing legible is written into the picture, because video models cannot write readable text; the words that have to land are spoken or put on the end card. Format is landscape for sites and screens or portrait for feeds and stories, and it does not change the price. Quality is 480P, 720P or 1080P, with the all-in price of each shown in the menu.

## What the script is not allowed to say, and the end card

Superlatives and absolute claims such as “the lowest price anywhere”, “number one in sales” or “cures everything” are stripped from the script before it is made, and a brand or product name that is itself a claim is softened. The ad ends on a card with your main image and up to four lines: Brand, Product name, Tagline and Call to action. Any line can be blank. The brand is copied letter for letter from a name printed on a sign or package, or from what you typed. To change a line later, use Edit end card and Assemble: the final cut is rebuilt without reshooting. The card is in the ad’s dialogue language, and the music is written for the ad and mixed under the dialogue, with nothing for you to pick. The end card · Where the music comes from.

## One-click or expert, and what you pay

| | One-click generate | Expert mode |
| Who writes the script | AI | AI first draft; you edit or rewrite it |
| Who runs cast, sets and shots | AI | You, in the editor dramas use |
| Price | One all-in price, a cap | A smaller start price, then each step on its button |
| Time | About ten minutes | As long as you take |
| Script editable afterwards | No; the shots are | Until it is broken down |

SceneMixer’s all-in price is the most the ad can cost: each stage freezes and charges its own actual price, so the total lands at or below it, and a stage that fails is refunded. Reading your photos is free. Ads never use the free first project. You cannot switch a one-click ad into expert mode or back; each is a fresh ad. One-click or expert mode · The one price is a cap.

## After it is made

If one shot is wrong, open the finished ad in SceneMixer’s editor, fix that shot and press Rebuild final cut; the other shots are not touched. A different length or format is a new ad. Cancelling works before shooting starts; credits spent on stages that began are not refunded. For the same photos in a story instead of an ad, see AI image to video. For the engine behind it, AI ad video generator.

## FAQ

### How do I turn photos into a video?

Open SceneMixer’s Ad Video tab, add up to four photos of one subject with the best one first, add a line about what it is, and choose a vibe, a length and a quality. An ad of 15, 30 or 60 seconds is made.

### How many photos can I use?

Up to four of the same subject, JPG or PNG, under 10 MB each.

### Is reading my photos charged?

No. The SceneMixer read is free and tells you what it thinks the product is, so you can correct it before you pay for anything.

### Will my product be redrawn?

SceneMixer redraws a product, dish or garment as a clean studio shot. A shop, car, house or app screenshot is kept as it is.

### My shop photo has customers in it.

That is fine. People are removed and the place is kept.

### Will my logo or price appear in the video?

Not reliably in the shots, because video models do not draw legible text. The brand, product name, tagline and call to action are on the end card, and the tagline is also spoken.

### How much does a photo ad cost?

It depends on the length and quality you pick. The all-in SceneMixer price is shown on the button and is a cap, so the final charge is at most that.

### Can I fix one shot afterwards?

Yes. Fix the shot in SceneMixer’s editor and rebuild the final cut. The other shots are not touched.

Updated November 2, 2026


---

# How Character Voices Work in SceneMixer | SceneMixer

URL: https://scenemixer.com/guides/pick-character-voices/
Voices are described in words by default and can be matched to a real preset you can listen to. How the two modes differ, and when to upload your own.

# How Character Voices Work

Quick answer: Every character comes out of the script read with a voice described in words — age, weight, pace, where it sits. By default that description is what the video model performs, and it does a reasonable job. If you want a specific, repeatable voice instead, match the character to a preset from the built-in library: you can play every one before choosing, and once attached it is used as an audio reference on every shot that character speaks in.

Voice descriptions are visible in the character panel of any project. Open a sample, click a character, and the voice line sits under the appearance description.

## Mode 1 — a voice written in words (the default)

The script read gives each character a short voice description, the same way it gives them an appearance. That line travels into the video prompt, and the model performs it. No sample file is attached; nothing is generated in advance.

This is the default because it is free, it is instant, and for most characters it is enough. The cost is consistency: two shots of the same character can land slightly differently.

## Mode 2 — matched to a preset you can hear

Open a character and ask for a voice match. A short, free pass reads the description and picks a voice from the built-in library, preferring one that speaks the right language. You get a sample to play, and a picker to choose a different one if the match is not right.

Once attached, the sample becomes an audio reference on every shot that character speaks in, which is what makes the voice hold across an episode.

The library is a fixed set of system voices with a pre-rendered sample for each, so matching costs nothing to try and nothing to change your mind about.

## Mode 3 — upload your own

You can upload a clip and use it as the character's voice. It is trimmed to a short reference and used the same way a matched preset would be. An uploaded voice is a sticky choice: later matching passes will not overwrite it. Uploading files is a member feature on SceneMixer — the first two modes are free for everyone.

Only upload a voice you have the right to use. A recording of a real person is that person's, and using one without their permission is not something the platform will help with.

## What the voice does and does not control

The voice covers timbre and delivery. It does not decide what is said — the lines come from the storyboard and are spoken word for word. It also does not carry across to narration: lines marked as off-screen or voice-over are handled as off-screen audio, and the people in frame keep their mouths closed.

If a shot has no dialogue, no voice is used at all — what you hear is the ambient track written into that shot.

## Changing your mind

Rematching, picking a different preset, and clearing back to a text description are all free and take effect on the next shot you generate. Segments you have already shot keep the voice they were shot with; regenerate a segment if you want it to pick up the new one.

| Mode | Cost | Consistency |
| Description in words | Free, automatic | Good, varies slightly per shot |
| Matched preset | Free to match and change | Holds across the episode |
| Your own upload | Member feature | Holds; not overwritten by rematching |

## FAQ

### Do I have to pick voices before generating video?

No. Without a matched voice the description is performed directly, and that is the default for every project.

### Can I hear a voice before committing to it?

Yes — every preset in your SceneMixer library has a sample you can play in the picker.

### Does matching a voice cost credits?

No. The matching pass on SceneMixer is free, and so is changing your mind.

### Will the voice be used for narration and inner monologue?

Those lines are treated as off-screen audio: they are heard, but the people on screen do not move their lips.

### Can I clone a specific person's voice?

Members can upload a clip they have the right to use. We do not offer cloning of a named person, and uploading someone else's voice without permission is not permitted.

### Can I give a character an accent or a dialect?

Yes — write it into the voice description, which is where most characters already are: every line that character speaks is performed as described, so a Taiwanese or Sichuan accent written there carries through the episode. It works while the character has no audio attached; a matched preset or an upload takes over and the description is not what gets performed. Give a Character an Accent or Dialect.

Updated November 3, 2026


---

# Pitch Deck for a TV Show: Slide-by-Slide Template and What AI Adds

URL: https://scenemixer.com/guides/pitch-deck-for-a-tv-show/
A slide-by-slide pitch deck template for a TV show, and what SceneMixer can make for it: a pilot, a pitch trailer, a drawn cast and a promo cover. It makes no slides.

# Pitch deck for a TV show

Quick answer: A TV show pitch deck is about ten slides: the title, the logline, the world, the characters, the tone, the pilot, the season arc, the engine, the trailer and the ask. SceneMixer does not make slides and does not write the deck; it makes the pictures and moving pieces a deck points at: a drawn cast, a pilot episode, a one-minute pitch trailer and a promo cover for the title slide. Write the deck in your own slide tool and put the video link on it.

Write your logline first, then make the cast and a one-minute trailer: they are the two things a reader will remember from the deck.

## The ten slides, and what SceneMixer can supply

| # | Slide | Put on it | SceneMixer can supply |
| 1 | Title | Show title, tagline, format (8 episodes of 3 minutes) | A promo cover: three poster options with your title painted in |
| 2 | Logline | One sentence | Nothing: you write it |
| 3 | The world | The place, its rules, why it is fresh | Location images drawn from your descriptions |
| 4 | The characters | Three to five leads and what each wants | Full-figure character sheets, downloadable |
| 5 | Tone and look | Three words and a reference look | Frames from the trailer in the look you chose |
| 6 | The pilot | One paragraph and its hook | The pilot episode itself |
| 7 | Season arc | What changes by the finale | An episode outline from the read, as a starting point |
| 8 | The engine | What repeats every episode | A second episode that repeats it |
| 9 | The trailer | A link to a one-minute video | The pitch trailer |
| 10 | The ask | What you want the reader to do | Nothing: you write it |
AI movie pitch generator · Logline examples · Story bible template · Make a promo cover poster.

## Sample slide text for an original show

THE NIGHT COUNTER — a series about a 24-hour pharmacy in a harbour town.

| # | Slide | Text |
| 1 | Title | THE NIGHT COUNTER. 8 episodes of 3 minutes, vertical. Everything that matters happens after midnight. |
| 2 | Logline | A pharmacist working the 24-hour counter in a harbour town learns that every customer after midnight is collecting something nobody prescribed. |
| 3 | The world | One pharmacy, one harbour, one night shift. The door is never locked, the till is never empty, and no one comes twice with the same story. |
| 4 | The characters | INES CALDER, 38, night pharmacist, wants to leave and cannot. TOBIAS, 19, her trainee, wants to stay. THE REGULAR, who has never been seen to enter. |
| 5 | Tone and look | Deadpan, warm and cold at once. Neon on wet streets. |
| 6 | The pilot | Ines fills a prescription with no doctor's name, for a boat that left the harbour a year ago. |
| 7 | Season arc | From Ines trying to leave to Ines taking over the counter. Episode 8 answers who the Regular is. |
| 8 | The engine | Each episode: one customer, one unfilled need, one small cost. |
| 9 | The trailer | A one-minute video of the five best moments. |
| 10 | The ask | Read the pilot; a development meeting. |

## Make the pieces in this order

- The cast. Draw the leads and the main place; the first 15 setup images on a SceneMixer account are free.
- The trailer first. A one-minute script of five to eight moments with the same cast is the cheapest thing that sells the show. A pilot or pitch trailer · A movie trailer · A sizzle reel.
- The pilot, if the pitch calls for an episode. Set the type to TV Series; an episode runs up to 300 seconds.
- The cover. Open the Cover art tab in Step 2 and generate three options.
- The deck, in your own tool, with the character sheets and cover downloaded from the asset library and a link to the video. The asset library.

## Showing the video in the deck

A finished episode can be shared by a link that opens a real page with no sign-in and no expiry, or downloaded at full quality and embedded. A video shared through a SceneMixer link always shows the AI label, even on a paid plan; your own download is unaffected. Compositing and sharing · Share links.

## Say plainly what it is

- Call it previs or a proof of concept, not a finished broadcast. What is previs? · AI drama clips as previs.
- No famous faces. Real people cannot be depicted, so there is no casting a known actor in the trailer.
- Say it is AI-made. Do AI videos need a label?
- Follow the recipient's rules. Whoever you pitch to has their own submission format and their own view of AI-made material; check before you send.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Street Life, Neon Nights; Costume Style Smart Casual, Professional Uniforms; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw Ines, Tobias and the Regular, and the pharmacy and the harbour street as locations; they are reused in the pilot and the trailer. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Piece | Length | What it proves |
| Pitch trailer | About 60 seconds | The tone, the cast and the hook |
| Pilot episode | Up to 300 seconds | That the first episode works |
| Cast sheets | A page of images | That the characters have faces |
| Promo cover | One image, two shapes | That the show has a title and a look |

## FAQ

### How do I make a pitch deck for a TV show?

Write about ten slides: title, logline, world, characters, tone, pilot, season arc, engine, trailer and ask. SceneMixer makes the pictures and video that go on them.

### Does SceneMixer make slide decks?

No. It makes a cast, a pilot episode, a one-minute trailer and a promo cover. You write the deck in your own tool.

### How do I pitch a TV show?

Lead with a logline you can say in one breath, show the cast and a one-minute trailer, and send what the recipient asks for, in the format they ask for.

### How long should a pitch trailer be?

About a minute, with five to eight moments.

### Do I need a pilot episode?

Not always. A trailer proves the tone; a pilot proves the story works. Make the trailer first.

### Can I use famous actors in the trailer?

No. Real people cannot be depicted.

### Can I download the character sheets?

Yes. Every sheet, scene image and video can be downloaded from your SceneMixer asset library.

### Is it broadcast quality?

Treat it as previs: tone, cast and pacing. Higher resolution is available on SceneMixer for the best clips.

Updated November 3, 2026


---

# Post Your AI Drama to TikTok, YouTube Shorts and Reels

URL: https://scenemixer.com/guides/post-your-ai-drama-to-tiktok-youtube-shorts-and-reels/
Get an AI short drama ready for TikTok, YouTube Shorts and Reels: vertical frame, which file to download, the AI label, captions, music and the first three seconds.

# Posting Your AI Drama to TikTok, YouTube Shorts and Instagram Reels

Quick answer: Make the project vertical from the start: choose 9:16 Portrait in step 1, because the frame cannot change later. Composite each episode, download it (the download is the full-quality file), and upload that file to the platform rather than posting a SceneMixer link. Turn on the platform’s own “made with AI” label when you post. Add captions in the platform’s editor or your own; the file already carries the episode’s background music, so a replacement track is only needed if you want one. And open each episode on a hook: the first seconds decide whether people keep watching.

Composite one finished episode, download it, and play it on your phone full screen before you post: that is what viewers will see.

## Vertical from the first step

Phone feeds are vertical. Pick 9:16 Portrait on the step 1 settings screen; it sets the frame of every video in the project and cannot be changed after you continue. A landscape project can still be posted, but it shows with bars above and below. Vertical or landscape.

## Which file to post

Composite the episode once every segment has a clip, then download it. The download is the full-quality original; the player and share pages use a lighter copy. Upload the downloaded file natively. A SceneMixer share link is fine for sending a preview to a friend, but feeds favour videos uploaded to the platform itself. Composite and share an episode · Download versus play version.

For a cleaner picture, draft at a low tier and generate the final segments at a higher one; the composite is only as sharp as the tier you used. Draft at 480p, finish at 1080p.

## The AI label

Downloads from an account without a paid plan carry a small SceneMixer AI label in the corner. On a paid plan, downloads are label-free by default; videos shared through a SceneMixer link always carry it. Separately, TikTok, YouTube and Instagram each have a setting to mark a post as made with AI; turn it on. Platforms can limit or remove realistic AI content that is not labelled. The AI label switch · AI short drama compliance.

## Captions and music

The clip has dialogue, ambient sound and the episode’s own background music, but no subtitles burned in, which suits these platforms: add captions with the platform’s auto-captions or in your editor, so people watching on mute can follow. The downloaded cut already carries the score; if you would rather publish with a different track, turn the Music switch off, download the version without music, and lay your licensed track (or the platform’s library track) under it, well below the voices. Background music in an AI short drama.

## The first seconds

Viewers decide almost at once. Open on the conflict, a line that raises a question, or a striking image, not an establishing shot; end each episode on an open question so the next one gets watched. You can shape this in the script before it is read, or by editing the first shot of an episode. Write hooks and cliffhangers into your script.

## Length and series

Platforms set their own maximum lengths and change them from time to time; check the current limit before posting a long episode. Post a series in order and name the parts in the title or the first frame of the caption, so a new viewer can find episode 1.

| Step | Do |
| Before the reading | Choose 9:16 Portrait in step 1 |
| Final segments | Generate on the tier whose sharpness you want to publish |
| Export | Composite the episode and download the file |
| Post | Upload the file natively; turn on the platform's AI label |
| Finish | Add captions in the platform or your editor; the cut already has its own score |

## FAQ

### Can I post a landscape project?

Yes, with bars. For phone feeds, start a vertical project; the frame on SceneMixer cannot change after step 1.

### Does the SceneMixer label stay on the file?

On accounts without a paid plan, yes. Paid plans can download without it. Platform AI labels are separate and still apply.

### Should I post the share link?

Upload the file instead; links are for previews.

### Can I add subtitles inside SceneMixer?

No. Add them in the platform's editor or your own editing app.

### Is the video file a standard format?

Yes, SceneMixer delivers an MP4 that every platform accepts.

Updated November 3, 2026


---

# "Your Previous Payment Is Still Being Processed": What It Means

URL: https://scenemixer.com/guides/previous-payment-still-being-processed/
Closed the checkout and opened it again, or switched from card to PayPal? Before starting a new payment, SceneMixer checks the last one, so you are never charged twice.

# "Your Previous Payment Is Still Being Processed": Switching Payment Methods Safely

Quick answer: When you close a checkout and start another — to switch from card to PayPal, say — SceneMixer first asks the payment provider what happened to the previous one. If it went through, your credits are added and you see the success message instead of a new checkout. If it is still being processed, you see Your previous payment is still being processed. Please check the order status in a moment. and no second order is opened, so you cannot pay twice. Wait a minute and look at your order history.

If you see the message, do not pay again elsewhere. Refresh your order history after a minute; the order will show Paid or become available to retry.

## Why the site checks first

A payment can finish a moment after you close the window: a bank confirmation, a PayPal approval that is still being completed. If the site simply cancelled your order and opened a new one, you could end up paying for both. So before any new checkout, the previous order is checked with the provider, and only cancelled if the provider confirms nothing is going through.

## The three outcomes

- It was paid — the order is completed, credits are added, and you get the success message. No new checkout opens because none is needed.
- It is still processing — Your previous payment is still being processed…. Nothing new is opened. Give it a minute.
- Nothing went through — the old order is cancelled and the new checkout opens as usual.

If the provider cannot be reached for a moment, the site tries again after a second or two before carrying on.

## PayPal specifics

After you approve on PayPal, SceneMixer completes the payment on its side. During those few seconds the order cannot be cancelled. If you close the PayPal window, the site checks the order before saying Cancelled. Nothing was charged, so you do not see “cancelled” for a payment that actually went through. Payment failed or still clearing.

## Retrying an unpaid order

An order that is waiting for payment can be retried from your order history; the retry reuses the same order number rather than creating a new one. Payment methods and order records.

## If you were charged twice anyway

Two payments for the same package within a few minutes are flagged to the SceneMixer team automatically. You can also write through feedback with both order numbers. Refund rules.

## Account deletion waits too

While a payment on your account is being processed, the account cannot be deleted; try again once the order shows its result. Delete your account.

| You see | What happened | Do |
| Payment success, credits added | The earlier payment went through | Nothing; do not pay again |
| Your previous payment is still being processed… | The provider has not finished | Wait a minute, check order history |
| A new checkout opens | Nothing went through before | Pay as usual |
| Cancelled. Nothing was charged | Checked with PayPal: not paid | Pay again if you want |

## FAQ

### How long does 'still being processed' last?

Usually a minute or two. Refresh your order history; it will show Paid or be retryable.

### Should I pay with another method meanwhile?

No. Wait for the result so you are not charged twice.

### I was charged but have no credits.

Check the order history after a few minutes, then send feedback with the order number if it still shows unpaid.

Updated September 30, 2026


---

# "Price Changed, Confirm Again" and "Tier No Longer Exists"

URL: https://scenemixer.com/guides/price-changed-please-confirm-again-and-tier-no-longer-exists/
Where the site checks a price you already saw before charging, why the check can refuse a click, and what a retired video tier means for a segment.

# ‘The Price or Selected Model Has Changed’ and ‘That Tier No Longer Exists’: What They Mean

Quick answer: ‘The price or selected model has changed’ comes from the few places that promise you a price before charging: an ad’s one-price and a remake’s quote. If the current price is higher than the one you confirmed, the click is refused and the new price is shown; nothing is charged until you confirm it. A card or PayPal membership upgrade works the same way with its own confirmation. Sheets, storyboards and clips are different: they are charged the price at the moment you click, with no second check, so a rate or allowance that changed while the page was open changes the charge. ‘The selected video tier no longer exists’ means the tier you picked was retired: the click is refused, nothing is charged, and you pick another tier.

Start an ad or a remake, leave the page open for a while, then confirm: if the price moved in the meantime, you are shown the new number before anything is charged.

## Where a price is promised in advance

Ad one-price: fixed when you start; before each paid stage the site checks the current price, and if a stage would now cost more the ad stops and asks. Remake quote: shown before the read and checked when you confirm. Membership upgrade by card or PayPal: the amount is shown first, and the charge is refused if it changed since you saw it. In each case you see the new number and decide. Subscribe, Upgrade, Cancel.

## Everything else is priced at the click

A sheet, a storyboard or a segment holds the price computed when you press the button. What can move that price while a page sits open: a tier’s rate adjusted, the account’s free sheets or the first project’s free storyboard used up elsewhere, a longer segment. Refresh the page before a large batch if it has been open for a long time. Why Video Costs More.

## Tier no longer exists

Model tiers come and go as providers change. A segment submitted on a retired tier is refused with this message and nothing is charged; the selector no longer lists that tier. Refresh, pick another tier in the model selector in the editor’s top bar and generate again. The segment keeps its text and versions. Which Model Is Best.

## What to do

For an ad or a remake, read the new number and confirm if you still want it. For a retired tier, pick another. For a surprise on an ordinary generation, check your credit history: each task’s line shows what it took.

| Message | Cause | Do |
| Price or model changed, confirm again | An ad or remake now costs more than you confirmed | Read the new price, confirm again |
| The selected video tier no longer exists | Selected tier retired | Refresh and pick another in the selector |
| Ad stopped: price changed | A later stage now costs more | Confirm the new price on the ad page |
| Confirm the upgrade amount | Upgrade amount changed since preview | Confirm again |

## FAQ

### Was I charged the old price or the new?

Neither; the click was refused. The next confirmation charges the price shown then.

### Does a normal generate button check the price again?

No. SceneMixer charges the price at the moment you click.

### Can I keep the old price?

No. For an ad or remake you confirm the new one or stop.

### The tier I liked is gone; will it come back?

Tiers follow provider availability; the selector shows what is offered now.

### Was anything charged when the tier was gone?

No. The request was refused before anything was held.

Updated November 3, 2026


---

# Progress Bar Stuck or Page Closed: Is the Generation Still Running?

URL: https://scenemixer.com/guides/progress-bar-stuck-page-closed-does-generation-continue/
Generations run on the server, not in your browser. What the bar and estimate mean, why a bar sits still for minutes, what closing the tab does, slow versus stuck.

# The Progress Bar Is Not Moving, or You Closed the Page: What Is Actually Happening

Quick answer: Every generation runs on the server: closing the tab, losing the connection or switching devices does not stop it, and the result is attached to the segment when it arrives. The page only watches: it keeps a live connection for updates and reconnects by itself if it drops; a refresh always shows the true state. The bar is not a countdown. Video is produced by the provider in one piece and reports few intermediate steps, so a bar can sit at the same place for several minutes and then finish. The time shown is the recent average for that tier, from tasks like yours. A job that has really stalled is handled for you after a fixed wait. A video clip moves to checking and the site asks the provider for the result; if it comes, it is attached. A sheet or a text job that times out is marked failed. Whenever the outcome cannot be confirmed, the credits stay held until support reviews them. Nothing needs a click from you.

Start a generation, close the tab, come back after a few minutes: the card shows the clip or the current stage.

## What the progress bar is showing

Stages, not percent of work: queued (waiting for a free slot), running (sent to the provider), checking (provider took longer than expected; the site is asking it for the result), done or failed. Within running there is little to report — a clip is rendered in one go — so the bar stays still and then jumps. The estimate next to it is the average duration of recent jobs on the same tier. The full list of states: What Each Task Status Means.

## Closing the page, losing the connection

The server does not know or care whether your page is open. The page keeps a connection for live updates; if the network drops, it reconnects on its own, and if it cannot, the next refresh reloads the real state. Two devices logged into the same account see the same cards. If a card looks frozen for a long time, refresh once before concluding anything.

## Queued: why it has not started

Your SceneMixer account runs a limited number of generations at once (3 free, 5 Pro, 8 Studio, 10 Enterprise); the rest wait in your queue and start as slots free up. That is the most common reason a card says queued for minutes. Generating Many Segments at Once.

## Slow versus stuck

Slow: running for less than the estimate plus a few minutes; providers vary. Long: a clip’s card moves to checking by itself, and the site asks the provider for the result and attaches it when it comes; a sheet or text job that times out is marked failed. When the outcome cannot be confirmed, the credits are held for review rather than refunded on the spot. You never need to resubmit a running task; doing so would create a second, charged job. What can and cannot be stopped: Stop a Generation.

## What to do if it really is wrong

A card that has said running or checking for far longer than any estimate, across a refresh, is worth a report: the Ask window’s feedback attaches the project and time. Credits for a task that fails are settled by the rules: When a Generation Fails.

| You see | Means | Do |
| Queued for minutes | Your concurrent slots are full | Wait, or cancel lower-priority queued tasks |
| Running, bar not moving | Provider rendering in one piece | Wait; refresh once if in doubt |
| Estimate passed | Recent average exceeded; normal variance | Wait a few more minutes |
| Checking | Site is asking the provider for the result | Nothing; it resolves by itself |
| Failed | Settled by the refund rules | Read the message on the card |

## FAQ

### I closed the laptop; is the clip lost?

No. It is attached to the segment when it arrives; open the editor to see it.

### The bar has been at the same place for five minutes.

Normal for video; the provider reports little until done.

### Can I speed a queued task up?

Cancel other queued tasks, or a higher tier's larger slot count.

### Should I submit again if it looks stuck?

No. A second submission is a second charged job. Refresh; then report if it is truly stuck.

### Why did the estimate change while waiting?

It is recomputed from recent jobs; busy hours lengthen it.

Updated November 3, 2026


---

# Project Cards: Resume, Rename, Delete, Project Limit | SceneMixer

URL: https://scenemixer.com/guides/project-cards-resume-rename-delete/
What a project card shows, what Step n/3 means, how resuming works, how to rename or delete a project, what deletion removes and does not refund, and the project limit.

# Project Cards: Resume, Rename, Delete

Quick answer: Every project is a card under Recent Projects on the home page and under Projects in the sidebar: a thumbnail, the name, and Step n/3 for how far it got. Click the card and the project reopens exactly there — nothing is lost by closing the browser, because every step saves itself. Hover for Rename and Delete Project. Deleting removes everything in the project — cast, sets, episodes, videos — permanently, and credits already spent are not refunded. Plans have a project limit; when it is reached, delete an old project or upgrade.

What is inside a project is what the card is a door to.

## What the card shows

The thumbnail is the first cast sheet or a frame from a video; the name is what you typed in step 1 or what the read suggested. Step 1/3: the script is being or has been read. Step 2/3: the cast, sets and props. Step 3/3: episodes and the editor. Ad projects and remakes carry their own labels. View all opens the full list, newest first.

## Resume

Click the card. Step 1 and 2 reopen in the wizard where you were; step 3 reopens the episode list, and from there the editor for the episode you were in. Every action saves as it happens — a generation you started keeps running on SceneMixer while you are away, and the card reflects it when you return.

## Rename

Hover the card, Rename, type the new name, Save. The name is only a label: it changes nothing inside the project. The project’s title in the summary is separate and stays.

## Delete

Hover, Delete Project, then Confirm Delete. The dialog says it: every character, scene, item, episode and video in the project is removed, permanently. Credits spent on them are not refunded; a generation still queued in the project is stopped and refunded, one already generating is charged. Files are purged after a short delay; share links stop at once. Where files live.

## The project limit

Each SceneMixer plan allows a number of projects; the Membership page lists it per plan. At the limit, creating a new project is refused with Project limit reached — delete an old project or upgrade your plan. Deleted projects free their slot at once. What a paid plan changes.

## Sample projects

Featured Projects are not yours: they cannot be renamed or deleted and open read-only. Learning from a sample.

| Action | Where | Reversible? |
| Resume | Click the card | — |
| Rename | Hover → Rename → Save | Yes |
| Delete | Hover → Delete Project → Confirm Delete | No; credits not refunded |
| See all | View all, or Projects in the sidebar | — |

## FAQ

### I closed the browser. Is my work gone?

No. On SceneMixer, every step saves itself. The card reopens the project where you stopped.

### What does Step 2/3 mean?

The project is at the cast, sets and props step. Step 3/3 on SceneMixer means episodes and the editor.

### Are credits refunded when I delete a project?

No. Only a generation still queued inside it is stopped and refunded on SceneMixer.

### Why can't I create a new project?

Your SceneMixer plan's project limit is reached. Delete an old one or upgrade.

### Does renaming change anything inside?

No. It is a label on the card.

Updated November 3, 2026


---

# Promo Codes: Where to Enter One, What It Does, Why It Fails

URL: https://scenemixer.com/guides/promo-codes-how-they-stack/
Promo codes apply to credit packages. Where the field is, what a code adds to the order, the limits on a code, and why one is refused.

# Promo Codes: Where to Enter One, What It Does to Your Order, and Why a Code Is Refused

Quick answer: Enter a promo code on the credits (billing) page before paying; it is checked on the spot. Codes apply to credit packages only. Percentage discounts from a code and a membership together are limited by a site-wide cap, which is currently zero, so today a code works by adding bonus credits to the order. A code has a validity period, a total number of uses and a per-account limit, and some need a minimum order. Type it exactly as issued: capitals matter.

Open the credits page, pick a package and apply a code: the page tells you straight away whether it is valid.

## Where to enter a code

On the credits page (Credits in the avatar menu), next to the package list, before you choose how to pay. The quick buy-credits window does not have a code field; open the full page if you have one. Membership orders do not take a code. How to buy credits.

## What a code does

On SceneMixer, a code can carry bonus credits, a percentage discount, or both. Percentage discounts, from a code and from a membership added together, are limited by one cap for the whole site; that cap is currently zero, so at the moment codes act through their bonus credits and the price you pay is the package price. The bonus arrives with the package’s credits when the payment lands.

## Limits on a code

Each code has a start and end date, a total number of uses and a maximum per account; some are limited to a minimum order amount or to credit purchases. The code is checked when you apply it and again when the order is created, so a code can run out between the two.

## Why a code is refused

Not found or expired: a typo, or outside its dates; SceneMixer matches codes exactly, including capitals. Used up: the total number of uses is reached. Already used: you have reached the per-account limit. Not for this order: the code is for another kind of purchase, or the order is below its minimum. If a code from us still fails, write to support with the code.

## Codes and unpaid or refunded orders

A use is counted when the order is created. If you cancel the order or let it expire unpaid, the use is released and the code can be used again. A paid order that is later refunded does not give the use back.

| Situation | Result |
| Valid code on a credit package | Bonus credits added with the package |
| Code outside its dates, or mistyped | Refused, nothing charged |
| Per-account limit reached | Refused |
| Order cancelled or expired unpaid | The use is released |
| Order paid, then refunded | The use is not released |

## FAQ

### Can two codes be used on one order?

No. One code per order.

### Why did the price not go down?

SceneMixer caps percentage discounts site-wide, currently at zero; a code works through its bonus credits.

### Do codes work on memberships?

No. The code field on SceneMixer is for credit packages.

### Does a code change the first-purchase bonus?

No. The first-purchase bonus on SceneMixer follows its own rules.

### Is a code case-sensitive?

Yes. Type it exactly as issued.

Updated November 3, 2026


---

# Props and Locations in a Shot: Right Object, Right Room | SceneMixer

URL: https://scenemixer.com/guides/props-and-locations-in-a-shot/
Where a segment's location is set, how a shot gets a specific prop instead of a look-alike, what belongs in Set dressing, and what to do when the object comes out wrong.

# Props and Locations in a Shot

Quick answer: Each segment happens in one place: the Location field of its header holds a scene chip, and every shot in the segment is set there. Objects that matter are props: list them in the header’s Props line and name them with an @ chip in the shot where they are handled, and the shot gets that object — the same one every time. Things that are simply in the room go in Set dressing as plain words. A new place means a new segment.

Newer storyboards. In storyboards written or regenerated since late October 2026 the segment header has only Location and a Time · Weather line: there is no Props or Set dressing line. A prop appears when a shot names it with @ (Props tab) and stays in the hand it is in until the text says otherwise; the room comes from the scene image. Older storyboards, ads and remakes keep the lines described below. The segment header, field by field · What the system keeps track of between segments.

Open a finished episode and read a segment header: Location first, Props further down, and the same prop names again inside the shots.

## The location

The first line of a segment header is Location, and its value is a scene chip from the project’s Scenes. That is the room, street or car the whole segment happens in, and the shots are generated in it — from any angle, but always that place. Why it is the same room every time.

To change where a segment happens, put the cursor in the Location field, type @ and pick another scene. To move the story to a second place mid-segment, do not: split it. Insert a segment after this one and give the new segment the new location. One segment, one place.

## Props: the objects that must be the same one

A prop is an entry in the project’s Items with its own image: the letter, the sword, the phone. A shot that names it with a chip — @Sealed letter is slid across the table — gets that object. Why a named prop stays the same.

Name it twice: once in the header’s Props line, so the segment knows what it needs, and again in each shot where the object is handled or clearly visible. A prop listed in the header but never mentioned in a shot is dropped from the header when you save.

## Set dressing: everything else

The table, the lamp, the rain on the window, the cars parked outside: these are the room, not props. Describe them in plain words in the Set dressing line. They do not need chips and do not need to be the same object twice. The same line is where background people are declared. The header, field by field.

If an ordinary object suddenly matters — the cup that will be poisoned in episode 4 — promote it: add it as a prop in step 2, generate its image, and name it from then on.

## Where the object is

Say it. @Sealed letter in @Mara’s left hand; the @Sealed letter on the table between them; @Jules picks up the @Sealed letter. A shot with the object in a hand and a shot with it on the table are two shots; write which one you mean.

## When it comes out wrong

Wrong object. Was it named with a chip in that shot? If it was typed by hand, the shot does not know which one you mean. Right object, wrong state — sheathed when it should be drawn, whole when it should be broken: that is a state of the prop, assigned to this segment. States. Too many objects. A shot asking for five named props usually loses one; split the action across two shots. Then regenerate the segment — every version is kept.

| Thing | Where it lives | How you write it |
| The place the segment happens | Header, Location field | One scene chip |
| An object that must be the same every time | Header Props line + the shots that use it | Prop chips |
| Furniture, weather on the glass, clutter | Header, Set dressing | Plain words |
| A crowd or passers-by | Header, Set dressing | Background people: … |
| A second place in the same scene | A new segment | Insert segment after, set its Location |

## FAQ

### Can one segment have two locations?

No. A segment on SceneMixer happens in one place. Split it into two segments and give each its own Location.

### Do I have to name a prop in every shot?

Name it in the shots where it is handled or clearly visible. Where it is out of the picture, leave it out.

### What is the difference between a prop and set dressing?

A prop must be the same object every time and has its own image. Set dressing is the rest of the room, described in words.

### The prop was listed in the header but disappeared after saving.

It was not mentioned in any shot, so the header dropped it. Name it in a shot and it stays.

### How do I add a prop the script did not have?

Add it under Items in SceneMixer's step 2 (or with Add assets in the editor), generate its image, then name it with @.

Updated November 3, 2026


---

# Props: How to Describe Them and When They Get a Reference Image

URL: https://scenemixer.com/guides/props-how-to-describe-them-and-when-they-get-a-reference/
Prop sheets are drawn on white from your description. What to write, how a sword or a broken watch gets states, and the line that decides which props ride into a segment.

# Props: How to Describe an Object, Which State to Draw, and When Its Sheet Is Used as a Reference

Quick answer: A prop sheet is the object by itself on a white background, lit only to show what it is made of. Describe the thing: shape, size against a hand, material, wear, one identifying mark. Draw it in its carried state — a sword in its scabbard, a watch intact; a structurally different form (blade drawn, glass cracked, blade snapped) is a separate state with its own image, chosen per scene. Where the prop is, who holds it and how it is lit are words in the storyboard, not states. In a segment, only props named in the header’s props line ride along as references, and a prop that the segment text never mentions is dropped from that line automatically. One kind of prop never gets a reference: a photograph or portrait of a character, because the models would copy the face on it into the scene.

Open a segment and read its header: the props line lists the sheets that travel with it; remove a prop from the text and the line drops it.

## What a prop sheet is

One object, white background, three-quarter view, light used to reveal material (brushed steel, cracked leather, lacquer). No hand, no table, no scene: the sheet is a reference the models look at when the object appears, and anything else in it would appear too. Same Prop in Every Hand.

## How to describe it

What it is, then size (fits in a palm; as long as a forearm), then material and finish, then wear (nicked edge, worn grip), then the one mark that makes it this object (a brass plate with initials). Skip what it means to the story; the sheet cannot draw meaning. Describe the Cause, Not the Result.

## States: drawn, broken, transformed

If the object’s structure changes during the story, each form is a state: the sheath and the drawn blade; the watch whole and the watch with a cracked face; the letter sealed and the letter opened. The main sheet is the everyday form; states are generated from it and chosen per scene. Being held, lying on a table, catching the light, covered in mud: those are described in the shot text and do not need a state. Use a State in One Scene.

## The props line in a segment

Each segment header carries a props line naming the objects that appear in it; those sheets are attached as references when the clip is generated. The line is filled from the storyboard text and cleaned up before saving: a prop that the shots never mention is removed. So if a prop is missing from the clip, check that the shot text names it. Props and Locations in a Shot.

## Photos and portraits of characters

A framed photograph of the heroine, a wanted poster with the villain’s face, a portrait over the fireplace: these are props whose image would carry a face into the scene, and the models copy faces from references literally. They are kept as props in the storyboard but are not attached as reference images; describe them in the shot text instead. Project limits on SceneMixer: up to 100 props per project by default.

| Change to the object | Handle as |
| Sword drawn from scabbard | State |
| Watch face cracked | State |
| Sword held in the left hand | Shot text |
| Watch lying on the nightstand in lamplight | Shot text |
| Letter opened | State |
| Letter covered in mud | Shot text |

## FAQ

### My prop's sheet has a hand in it.

Remove any mention of holding from the description; the sheet is the object alone.

### Which form should the main sheet show?

The carried, everyday form; other forms are states.

### The clip did not show the prop.

Check that the shot text names it; the props line drops unmentioned props.

### Why is the portrait of my character not used as a reference?

Its face would be copied into the scene; portraits of characters are described, not attached.

### Do states cost credits?

Each state image on SceneMixer is a generation at the sheet price; the first project has a free allowance.

Updated November 3, 2026


---

# "Generation Service Hit a Temporary Problem": When to Retry

URL: https://scenemixer.com/guides/provider-busy-not-charged-what-now/
The provider was overloaded and turned the request away. The site already kept trying; your credits are back. When to retry, and how it differs from queueing.

# “The Generation Service Hit a Temporary Problem. You Weren’t Charged”: What It Means and When to Retry

Quick answer: The model provider refused the request because it was overloaded. Before giving up, the site kept trying: when a text tier or a Qwen image tier was rate-limited it resubmits for up to about a quarter of an hour, and for every task it then tries once more on another account (or after a short pause); when that also failed, the task was marked failed and the held credits went straight back. Nothing is wrong with your segment. Wait a few minutes and press Retry; if it keeps happening, generate other segments meanwhile or come back later.

Press Retry on the failed card; the price shows again because it is a fresh task, held the moment you press it, including while it waits in your queue.

## What the message means

Every generation goes to a model provider. Providers have capacity limits, and at busy hours a request can be turned away with a “busy” or “rate limited” answer before any work is done. Because no work was done, nothing was billed, and the site returns the held credits immediately. The card says so in one sentence; there is nothing to report. The families of failure messages: When a Generation Fails.

## What the site already did

Before marking the task failed, the site kept trying. Work on a text tier or a Qwen image tier that was refused as rate-limited is resubmitted with growing pauses for up to about a quarter of an hour; then, for every kind of task, the site tries once more on a second account for the same tier when one exists, otherwise on the same account after a short pause. Most busy moments clear in that time and you never see the message. When you do see it, the provider was busy for a while; try again a few minutes later. SceneMixer only retries video tasks within their first 30 minutes.

## Busy is not the same as queued

Queued is your own account’s line: you started more tasks than your plan runs at once, and the rest wait with their credits held and start by themselves. Busy is the provider’s line: the task was refused and its credits returned, and it will not start by itself — you press Retry. Generating Many Segments at Once and Why a Generation Takes Time.

## If it keeps happening

Two or three busy failures in a row within a few minutes means the provider is congested. Options: generate other segments meanwhile, come back later, or switch that segment to MiniMax H3, which runs on separate capacity; switching to another tier on SceneMixer usually does not help otherwise. Sheets and storyboards rarely hit this; when they do, the same rule applies.

## When it is not the provider

A message that names your segment — invalid input, moderation — is a different family and retrying unchanged will fail again. Invalid Input; Content Rules and Moderation.

| You see | What happened | Do |
| The generation service hit a temporary problem. You weren’t charged | Provider overloaded; nothing generated | Retry in a few minutes |
| Queued | Your plan's concurrency reached | Wait; it starts by itself |
| Generating (long) | Normal; video takes minutes | Wait |
| Busy three times in a short while | Provider congested | Come back later, or try MiniMax H3 for that segment |

## FAQ

### Were my credits taken?

No. The card says refunded, and the task’s line in your SceneMixer credit history is tagged Refunded too.

### How long did the site keep trying?

If a text tier or a Qwen image tier was rate-limited, SceneMixer keeps trying for up to about a quarter of an hour, then once more on another account; in other cases, once more on another account. Waiting longer would only keep your credits and a slot tied up.

### Is there a better time of day?

Congestion varies by hour. On SceneMixer, MiniMax H3 runs on separate capacity from the other video tiers.

### Does Retry cost more?

No. Retry costs the same as any generation on SceneMixer, held the moment you press it.

### Will my storyboard change?

No. Retry on SceneMixer re-sends the same segment.

Updated November 3, 2026


---

# Qwen-Image 3.0 Pro: The Recommended Image Tier for Faces | SceneMixer

URL: https://scenemixer.com/guides/qwen-image-3-pro-tier/
Qwen-Image 3.0 Pro on SceneMixer: why it carries the Recommended badge for character sheets, what it improves over the default, its price, and why it sometimes queues.

# Qwen-Image 3.0 Pro: the recommended tier

Quick answer: : Qwen-Image 3.0 Pro is the image tier marked Recommended in Settings: the default tier’s way of reading a description, with more identity in a face and more reliable texture, at 18 credits per sheet (about $0.11 at the lowest credit price). It is the sensible tier for character sheets that must hold up in close-up. One habit: at busy moments it queues for a short while, because it accepts fewer requests per minute than the default tier.

The default tier is the cheaper sibling; this page is about the step up.

## Where you pick it

Settings → Model Preferences → Image Model. The choice applies to every sheet you generate from then on — character, location and prop images in step 2, regenerations, state images — until you change it. Sheets already made keep their pixels; nothing is re-rendered by switching.

In the list it carries the Recommended badge. The badge is advice: new accounts still start on the default tier.

## What improves over the default

Faces carry more of what the description says — the shape of the eyes, the set of the mouth, the one feature that makes the character recognisable — and skin and hair keep texture rather than smoothing out. For a cast that will be seen in close-up across many episodes, that identity is what keeps the same face coming back shot after shot.

The reading habits are the family’s: identity first in a long description, and write what to draw rather than what not to. State images are edited from the base sheet; a redraw from your own photo is always charged and does not use the free sheets.

## Why it sometimes queues

This tier accepts fewer requests per minute than the default. When many sheets are generated at once — yours, or everyone’s at a busy hour — some wait briefly and the site retries them for you. A sheet that seems slow on this tier is usually queued, not stuck; give it a minute before regenerating.

## What it costs

Every account has 15 free sheets in total (across projects, any type) on a project whose script has been read, drawn on the default image model, Qwen-Image 3.0, only. On this tier, SceneMixer charges each sheet from the first one: 18 credits (about $0.11 at the lowest credit price). A rejected generation is not charged, and most failures are refunded automatically; a few kinds of failure are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

Against the default tier at 7 per sheet and SeedDream 5.0 Pro at 25.

## Content checks

Same as every image tier: the description is checked before rendering and a flagged description is not charged; edit the wording rather than retrying. See content rules and checks.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the character or scene description tripped the check | Edit the description and generate again |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Generate again in a few minutes |
| A sheet that ignores part of the description | Long description; this tier follows the start more reliably than the end | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |

## When to switch tier

Down to the default for locations and props, where identity matters less than layout. To SeedDream 5.0 Pro when a description is long enough that its end matters, or for the most photographic skin. Existing sheets are untouched by switching.

| | This tier |
| Role | Recommended for character sheets |
| Price per sheet | 18 credits |
| Faces | More identity and texture than the default |
| Busy hours | May queue briefly; the site retries |
| State images | Edited from the base sheet |
| Free sheets | 15 per account, default tier only |

## FAQ

### Is Recommended the same as default?

No. Recommended is the badge on the tier SceneMixer suggests for most character sheets; the default for new accounts stays the cheaper tier.

### Why is a Pro sheet taking longer?

It accepts fewer requests per minute; at busy moments a sheet waits in a short queue and the site retries it. It is not stuck.

### Should I use Pro for locations too?

Usually not; the default tier lays out a room or a street just as well for less. Pro earns its price on faces.

### Does Pro read the whole description?

It follows the start more reliably than the end, like the default. Keep identity first.

### Can I regenerate one character on Pro and leave the rest?

Yes. Switch the preference, regenerate that sheet, switch back if you like.

Updated November 3, 2026


---

# Qwen-Image 3.0: The Default Image Tier for Sheets | SceneMixer

URL: https://scenemixer.com/guides/qwen-image-3-tier/
What the default image tier does on SceneMixer: sheets at the lowest price, how it reads long descriptions, edit mode with a reference, and when to step up.

# Qwen-Image 3.0: the default image tier

Quick answer: : Qwen-Image 3.0 is the tier every account starts on for character sheets, location images and prop images. It is the cheapest of the main tiers at 7 credits per sheet (about $0.04 at the lowest credit price), it follows a description written in plain language well, and it draws state images by editing the base sheet. Its two habits to know: it follows the start of a long description more reliably than the end, and it does not reliably follow instructions phrased as what not to draw.

What you get for free covers the free sheets every account has before this tier starts charging.

## Where you pick it

Settings → Model Preferences → Image Model. The choice applies to every sheet you generate from then on — character, location and prop images in step 2, regenerations, state images — until you change it. Sheets already made keep their pixels; nothing is re-rendered by switching.

If you have never changed the preference, this is the tier your sheets come from.

## What a sheet on this tier looks like

Character sheets are full-length portraits on a plain background at 9:16; location and prop images are 16:9. The tier renders at a resolution suited to reference use rather than print: sharp enough for the video tiers to read a face, a costume and a room, with about half the pixels of the SeedDream tiers. Skin and fabric read as real; very fine texture is where the Pro tiers pull ahead.

## How it reads your description

Two habits worth knowing. First, it weights the beginning of a description more than the end: put the things that identify the character — face shape, eyes, hair, the one feature nobody else has — before clothing and mood. Second, it does not reliably follow negatives: no hat can even pull a hat into the picture, while bare head, hair tied back works. Write what you want to see. Write the cause, not the result has more on phrasing that these tiers follow.

## Drawing from a reference

Two different things happen when a reference image is involved. A state image — the same character in another costume, a location after the fire — is drawn from the base sheet with a short edit instruction, so it keeps the face or the room and changes only what the state describes. Your own photo works differently: the site reads the photo, combines what it sees with the description from your script, and draws a proper reference sheet from scratch, on a plain background in the standard framing. That redraw is always charged at this tier’s price and does not use the free sheets.

## What it costs

Every account has 15 free sheets in total (across projects, any type) on a project whose script has been read; after that each sheet on this tier costs 7 credits (about $0.04 at the lowest credit price). A rejected generation is not charged, and most failures are refunded automatically; a few kinds of failure are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

For comparison: Qwen-Image 3.0 Pro is 18 per sheet, SeedDream 5.0 Pro 25.

## Content checks

The description is checked before rendering. A flagged description is not charged; edit the wording rather than retrying the same text. See content rules and checks.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the character or scene description tripped the check | Edit the description and generate again |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Generate again in a few minutes |
| A sheet that ignores part of the description | Long description; this tier follows the start more reliably than the end | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |

## When to switch tier

To Qwen-Image 3.0 Pro when a face needs more identity in close-up. To SeedDream 5.0 Pro for the most realistic skin and for descriptions so long that the end matters. Switching does not touch sheets already made; regenerate the ones you want on the new tier.

| | This tier |
| Role | Default for character, location and prop sheets |
| Price per sheet | 7 credits |
| Long descriptions | Start followed more reliably than the end |
| Negative phrasing | Unreliable; write what to draw |
| State images | Edited from the base sheet |
| Redraw from your photo | Always charged; not from the free sheets |
| Free sheets | 15 per account, on this tier only |

## FAQ

### Why is the default the cheapest tier?

Because sheets are references for the video tiers, not final art; this tier gives the video tiers what they need at the lowest cost.

### Do the free sheets only work on this tier?

Yes. Since October 24, 2026 SceneMixer draws the free sheets on this default tier only. Sheets on another image tier are charged at that tier's price and do not use the allowance; switch back and whatever is left is still there.

### Can I use one tier for characters and another for scenes?

Not at the same time; the preference is one setting. Switch before generating each group.

### Does switching tiers regenerate my existing sheets?

No. Existing images stay. Regenerate the ones you want on the new tier.

### Why did my sheet ignore a detail at the end of the description?

This tier follows the start more reliably than the end. Move the detail up, or use a tier that reads the whole description.

Updated November 3, 2026


---

# Qwen3.8 Flash: A Low-Cost, Fast Text Tier | SceneMixer

URL: https://scenemixer.com/guides/qwen3-8-flash-text-tier/
Qwen3.8 Flash on SceneMixer: a low-cost, fast text tier, what it is good for (quick drafts, short scripts), where it falls short on storyboards, and its prices.

# Qwen3.8 Flash

Quick answer: : Qwen3.8 Flash is a low-cost text tier at 10% of the base rate, and quick to answer. It is the tier for a quick read of a short script or a draft storyboard to see an episode’s shape. Dense episodes are where it is most likely to miss the storyboard format; when that happens, the site’s retry and then a higher tier are the path.

Qwen3.8 Max is the full-size sibling for hard scripts.

## Where you pick it

Settings → Model Preferences → Text Model. The choice applies to the three text jobs that cost credits — reading a script, writing a script with AI, and generating an episode’s storyboard — from the next task on. Translations of what you see on screen and the small helper calls do not follow it. Nothing already generated is redone by switching.

## What it is good for

Speed and cost. Reading a short script to see its cast and episodes in seconds; a first storyboard of an episode to check the shot count and pacing before you commit a pricier tier; AI-written drafts you expect to rewrite. SceneMixer itself uses this tier for the quick style suggestion it makes when you create a project.

## Where it falls short

A long or crowded episode is where a fast, low-cost tier is most likely to miss the strict storyboard format — a shot without its duration, a character named who is not in frame. The site checks every draft and asks for one more before reporting a failure; on this tier that second draft is needed more often. Treat a failure as a signal to step up for that episode.

## What it costs

On this tier, reading a script costs 1 credit per thousand characters (a thousand equivalent characters for alphabet-based scripts, see the pricing page), a storyboard costs 1 credit per hundred characters of the episode’s share of the script, and AI writing costs 1 credit per thousand characters written. That is 10% of the base rate. Most failed tasks are refunded automatically; a few kinds of failure hold the credits until support checks the task.

The default tier is at 50%.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| A long script or a dense episode fails | The reply ran past what one call can return; the site already ran it once more | The credits stay held until support checks the task. Trim the script or split it before running again |
| Storyboard generation failed | Neither of two drafts came back as a usable storyboard | Nothing was saved or charged; the credits stay held until support returns them. Regenerate, or switch tier |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Try again in a few minutes |
| The storyboard came back in the wrong language | The tier ignored the language setting | The site redoes it once automatically; if it still arrives wrong, regenerate or switch tier |

## When to switch tier

To the default for anything you will keep. To Qwen3.8 Max for hard Chinese scripts. A script is read once, on the tier you had at the time; for a better read of the same script, switch tier and start a new project with it.

| | This tier |
| Role | Cheap, fast drafts and short scripts |
| Rate | 10% of base |
| Long scripts | Read in two passes automatically |
| Format checks | Retried once before a failure is reported |

## FAQ

### Why is its rate so low?

Tier rates follow what each tier costs to run, and this one costs little.

### Can I use it for a whole series?

You can. Expect more storyboard retries on long or crowded episodes; switch those episodes up.

### Why might a storyboard fail the format check here?

The format is strict, and a dense episode asks a lot of a fast tier. SceneMixer retries once; if it still fails, step that episode up.

### Does the site use it anywhere by itself?

For the style suggestion when you create a project, where speed matters and the stakes are low.

### Is a Flash storyboard usable for video?

If it passed the checks, yes. The video tiers read whatever storyboard you approve.

Updated November 3, 2026


---

# Qwen3.8 Max Text Tier: For Hard Chinese Scripts | SceneMixer

URL: https://scenemixer.com/guides/qwen3-8-max-text-tier/
Qwen3.8 Max on SceneMixer: a text tier above the base rate for dense Chinese scripts and crowded storyboards, its prices, and when it is worth it.

# Qwen3.8 Max

Quick answer: : Qwen3.8 Max is a text tier at 133% of the base text rate. It is a choice for Chinese-language scripts and storyboards with many characters, and it reads long instructions carefully. Very long scripts are read in two passes on this tier as on every text tier — the site does that for you. Pick it for a hard Chinese script when the base-rate tiers have failed twice; its cheaper sibling, Qwen3.8 Flash, is for speed.

Qwen3.8 Flash is the fast, cheap sibling; this page is the full-size tier.

## Where you pick it

Settings → Model Preferences → Text Model. The choice applies to the three text jobs that cost credits — reading a script, writing a script with AI, and generating an episode’s storyboard — from the next task on. Translations of what you see on screen and the small helper calls do not follow it. Nothing already generated is redone by switching.

## What it is good at

Chinese scripts with period vocabulary, large casts and interlocking scenes; storyboards where every shot must name who is in frame and what they hold; instructions that run long.

## Long scripts

Each reply has a ceiling, on this tier as on the others. A long script is read in two passes: the cast and world first, then the episode outlines. You see one read; the split is automatic. A storyboard for a single episode fits comfortably.

## What it costs

On this tier, reading a script on SceneMixer costs 6 credits per thousand characters (a thousand equivalent characters for alphabet-based scripts, see the pricing page), a storyboard costs 2 credits per hundred characters of the episode’s share of the script, and AI writing costs 6 credits per thousand characters written. That is 133% of the base rate. Most failed tasks are refunded automatically; a few kinds of failure hold the credits until support checks the task.

Qwen3.8 Flash is at 10%; the default and Pro tiers sit between the two.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| A long script or a dense episode fails | The reply ran past what one call can return; the site already ran it once more | The credits stay held until support checks the task. Trim the script or split it before running again |
| Storyboard generation failed | Neither of two drafts came back as a usable storyboard | Nothing was saved or charged; the credits stay held until support returns them. Regenerate, or switch tier |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Try again in a few minutes |
| The storyboard came back in the wrong language | The tier ignored the language setting | The site redoes it once automatically; if it still arrives wrong, regenerate or switch tier |

## When to switch tier

Down to the default for routine episodes. To Evolving or Pro for a hard English script at the base rate. To Flash when you want a quick, cheap draft of a storyboard to see the shape of an episode.

| | This tier |
| Role | For hard Chinese scripts and dense storyboards |
| Rate | 133% of base |
| Long scripts | Read in two passes automatically |
| Format checks | Retried once before a failure is reported |

## FAQ

### Is it better than Pro for Chinese scripts?

It is a strong choice for dense Chinese scripts; for ordinary ones Pro and the default are cheaper and comparable.

### Why two passes on long scripts?

Every text tier has a length ceiling per reply; SceneMixer splits the read so nothing is cut off.

### Is it good for English scripts?

Yes. For English scripts the base-rate tiers, Evolving and Pro, are the usual step up from the default.

### Can I use it for one episode only?

Yes. The preference applies per task.

Updated November 3, 2026


---

# Reading Step 2: The Tabs, Badges and Shapes | SceneMixer

URL: https://scenemixer.com/guides/reading-step-2-tabs-and-badges/
Characters, Scenes and Items: what each tab holds, what the card badges mean, the progress line, and why character images are tall while scene ones are wide.

# Reading Step 2: Tabs, Badges and Why Scenes Are Wide

Quick answer: Step 2 is three tabs — Characters, Scenes, Items — each a grid of cards, one per entry the reading found. A card shows the entry’s image or its status (Pending, Generating, Failed, Uploaded), and on characters a role badge (Lead, Supporting, Villain) or Off-screen for a voice-only part. Character images are tall, scene and item images are wide, whatever your video ratio. Click a card to open its details.

Open a finished episode and look at the left panel: it is the same three lists, filtered to that episode.

## The three tabs

Characters: everyone the reading put in the cast, leads and bit parts alike. Scenes: every place the story happens in. Items: the props the story depends on. Switching tabs changes the grid and the Generate All button above it, which always applies to the tab you are on. Generate All.

## Badges on a card

Lead / Supporting / Villain: the role the reading assigned. It is a label, not a setting; edit it in the card’s details if it is wrong. Off-screen: a voice-only part — a narrator, a caller — that gets a voice but no image (Voice only — no image generated). Pending: no image yet. Generating: on its way. Failed — click to retry. Uploaded: your own image; generation never overwrites it. State: a card with a dashed border is a state of another entry, not an entry of its own. States.

## The progress line

Above the grid, a line counts the batch: Characters generating — 12/20 done. Another shows the free allowance on SceneMixer: N/15 free character, scene and prop images left, counted across the account. When something in the tab has no image yet, a hint says how many and offers to generate or upload.

## Why the shapes differ

Character images are tall: a full-body sheet on a plain background, made to be seen at every distance. Scene and item images are wide: a place seen as a whole, an object seen clearly. Neither follows the video’s aspect ratio, and changing the ratio never regenerates them. Aspect ratio.

A character card also carries two automatic crops once the image exists: Headshot and Half body. You do not make them; they are cut from the full-body image.

## Opening a card

Click a card for its details: the name, the description (appearance, clothing), the image with Regenerate, Upload & Replace and AI Edit Image, the role, and for characters the voice. Edit any of it and Save. Regenerating one sheet.

## Adding an entry

The Add button on each tab creates an entry the reading did not: name it, describe it, Create, then generate or upload its image. To bring an entry over from another project, press Edit on a card and choose Pick existing — the whole entry moves over, free on any SceneMixer plan. Reusing entries across projects.

## Leaving step 2

You do not have to finish. Next works with images still generating or missing; a dialog counts how many have no image yet and offers Generate all or Skip & continue. Entries without an image are drawn from their description alone, which is where faces drift, so give the cast its images before you generate video.

| Badge | Means | Do |
| Lead / Supporting / Villain | Role assigned by the reading | Edit in details if wrong |
| Off-screen | Voice-only part, no image | Nothing; give it a voice |
| Pending | No image yet | Generate, or Generate All |
| Generating | Image on its way | Wait |
| Failed | Generation failed | Click to retry |
| Uploaded | Your own image | Nothing; it is never overwritten |
| State (dashed card) | A state of another entry | Open to assign or generate |

## FAQ

### Why are my scene images wide when my video is vertical?

Scene and item images describe a place or object, not a frame. They are always wide; character images are always tall.

### What does Off-screen mean on a character?

A voice-only part. It gets a voice but no image is generated for it.

### Do I have to generate every image before moving on?

No. Next offers to generate the missing ones or skip. Give the cast images before generating video, though.

### What are Headshot and Half body?

Automatic crops from the character's full-body image. You do not make them.

### Can I change a character's role badge?

Yes. It is a label; edit it in the card's details.

Updated November 3, 2026


---

# Reading Your Credit History: Held, Usage, Refund | SceneMixer

URL: https://scenemixer.com/guides/reading-your-credit-history/
What each line in Credit History means: Held when a task starts, Usage when it delivers, Refund when it fails, Reversed after a chargeback, plus top-ups and grants.

# Reading Your Credit History

Quick answer: Click your balance at the top right: the panel shows Available Credits and Frozen, then your latest entries under Credit History; View all at the bottom opens the full Credit History on the billing page, grouped by day. Since October 2026 each task is one line — a video segment, a sheet, a storyboard, a fix — with what it was, where it was (project, episode, segment) and the credits it took. A line still running says In progress; a failed one shows its amount struck through and Refunded; Partly refunded and Under review mean what they say. Credit purchase, Membership credits, Signup bonus, Promotional bonus and Referral reward are lines of their own, with a plus sign. Purchase reversed only appears after a refund or chargeback on a payment.

Where credits go in one episode shows the same numbers from the other end.

## One line per task

Pressing any priced button holds the credits, and a line appears straight away: the task’s name — Video generation, Character image, Storyboard, AI visuals, Trim, Enhance clarity and so on — with the project, episode and segment underneath, and the amount with a minus sign. While it runs it is tagged In progress, and the same credits show under Frozen in the balance panel. When the task delivers, the tag goes and the line stays as spent. When it fails before anything was made, the amount is struck through and tagged Refunded: those credits are back in your balance. Which failures refund.

Partly refunded means part of the hold came back and the rest was spent; the amount shown is what you actually paid. Under review means the task ended in a state that needs checking; the credits stay on hold, and the page notes: Credits marked Under review are on hold while our team checks the task. After review they’ll be refunded or settled. Add up the amounts on the page and you get the real change in your available credits; a refunded line is struck through and does not count. Stopping a task refunds a hold that had not started.

## Credits coming in

Credit purchase: a pack you bought, base and bonus together. Membership credits, with the plan’s name: a plan’s period grant, one line per month. First subscription bonus and Promotional bonus: extras on a first subscription, first top-up or a limited-time offer. Signup bonus: the 50 credits a new SceneMixer account starts with. First video share reward: the reward for sharing your first video. Referral reward: your share when someone you referred tops up. Credits you were given rather than bought are shown in gold. The referral program.

## Reversed and clawed back

Purchase reversed lines, marked After a payment refund or dispute, appear only when money went back: a refund on an order, or a chargeback. The credits that payment bought are taken back, as far as your available balance covers them; for a membership, what could not be taken back is settled from later membership grants first (credits to be offset on the plan card). Referral reward reversed follows when the referred person’s payment was reversed. Credits restored puts them back if a chargeback is decided in your favour.

## Adjustments

Added by support / Adjusted by support: a manual correction by support; the reason is not printed on the line, so ask support if you need it. When support refunds a task after checking it, or releases credits that were stuck on hold, no new line appears: the task’s own line changes to Refunded or Partly refunded. A refund that belongs to no task shows as Credits returned.

## Finding what a line belongs to

Each task line names the project, the episode and the segment it was for, or the character, scene or item for an image, as they are named now: rename a project and its older lines show the new name; a deleted project’s lines keep only the task’s name. There is no task number on the line; open that segment in the editor to see the clip itself. Task statuses.

| Line | When | How it shows |
| A task (Video generation, Storyboard…) | A priced button is pressed | Minus the price; In progress while it runs |
| The same line, Refunded | The task failed, or was stopped before starting | Amount struck through; not counted |
| The same line, Partly refunded | Part of the hold came back | Minus what was actually spent |
| The same line, Under review | The task needs checking | On hold until settled |
| Credit purchase | A pack is paid | Plus |
| Membership credits | A plan period starts | Plus |
| Purchase reversed | A payment is refunded or charged back | Minus |
| First video share reward | You share your first video and claim the reward | Plus |

## FAQ

### What are the Trim, AI visuals and Redub lines?

Fixes made with Edit this clip under the player. Each is one line named after the fix (Trim, AI visuals, Redub, Enhance clarity) with the segment it was made on, held while it runs and spent when it lands; one that does not go through shows Refunded.

### Why are some credits frozen?

They are held by tasks still running; those lines say In progress. Each is spent when its task delivers, or shows Refunded when it fails. A line marked Under review stays on hold while the task is checked.

### A task failed. Where is my refund?

On the task's own line: the amount is struck through and tagged Refunded, and the credits are back in your balance. If the line says Under review, the task is being checked first.

### Why is there a Purchase reversed line?

A payment was refunded or charged back, so the credits it bought were taken back.

### Can a reversed payment make my balance negative?

No. SceneMixer takes back only what is available. For a membership, the rest is settled from later membership grants first, shown as credits to be offset on the plan card.

### Where do I see which task a line belongs to?

Under the task's name on the line: the project, episode and segment, or the name of the image. The line does not show a task number.

Updated November 3, 2026


---

# Recast a Remake: Swap Faces, Voices and Locations | SceneMixer

URL: https://scenemixer.com/guides/recast-a-remake/
A remake's cast, sets and props are entries you rewrite before an image is made: change who plays whom, give them voices, move the story, keep the storyboard.

# Recasting a Remake: New Faces, Voices and Places

Quick answer: After the read, the people, places and props from your clip sit in step 2 as entries with descriptions and no images. That is the moment to recast: rewrite a character’s appearance, upload a face, pick a voice, rename a place or replace it with your own photo, then generate the images. Uploading an image or a voice sample is a paid-plan feature. The storyboard read from the clip stays as it is and names the same entries, so the new cast walks through the same shots. Change the world as a whole with Target Region in step 1; change one person at a time in step 2.

Why a recast face stays the same in every shot applies to remakes exactly as to dramas.

## What the read gives you

Here is what we found: the shots, the characters, the scenes, the props, the lines. Each character has a name (yours if you gave them in the one line, invented otherwise), a description of what the read saw, and a voice description. Each scene has a description of the place. Nothing has an image yet, and that is the point: the images are made from whatever the entries say when you generate them.

## Recast a character

In step 2, open the character. Rewrite Appearance and Clothing to the person you want — age, face, hair, build, what they wear — and Save, then generate the image. Or Upload & Replace with a photo (a paid-plan feature on SceneMixer) and optimise it into a sheet. The name can change too; the storyboard follows a rename. Writing a description · Using a photo.

To reuse a character you already have in another project, open its card in step 2 and press Pick existing — the whole entry moves over, voice included, free and on any plan. Reusing a cast.

## Give them voices

Each character has a voice description from the read. Match a voice from it, pick one from the library with a preview, or upload your own (a paid-plan feature). The lines the read heard are then spoken in that voice. Character voices.

## Change the places

A scene read from the clip is a description of what the read saw. Rewrite it, or upload a photo of your own location (a paid-plan feature), and generate. The storyboard’s segments keep pointing at the same scene entry, so the whole scene moves with it. To move the entire story to another market at once, use Target Region in step 1 before you press Next. Region in a remake.

## Change the words

The storyboard is editable like any other: open the episode in step 3, edit a shot’s line, save, and shoot. The read’s shot structure — who is in each shot, what they do — is what you keep; the dialogue is yours to change. Editing the storyboard.

## What stays from the clip

The structure: the cuts, the framing, who is on screen when, what happens, what is said (as text). Nothing visual from the source is reused: not the faces, not the footage. That is what makes a remake a new picture and keeps you clear of the source’s images; it is not a licence to its story. Rights.

| You want to change | Where | The storyboard |
| Who a character looks like | Step 2, Appearance and Clothing, then generate | Unchanged; follows the entry |
| A character’s voice | Step 2, the voice on the card | Unchanged |
| A character’s name | Step 2, the name field | Updated to the new name |
| A location | Step 2, the scene’s description or photo | Unchanged; points at the same scene |
| The whole world | Step 1, Target Region, before Next | Names kept; descriptions relocated |
| A line of dialogue | Step 3, the editor | That shot |

## FAQ

### Can I put my own actors in a remake?

Yes. In step 2, rewrite each character's description, or upload a photo (a paid-plan feature), then generate the images.

### Will the new cast follow the same shots?

Yes. The storyboard read from the clip names the entries, not the faces; whatever the entries look like walks through it.

### Can I move the story to another country?

Yes. On SceneMixer, Target Region in step 1 relocates ethnicity, sets and props; names stay as read.

### Can I change what they say?

Yes, in the editor, shot by shot, like any storyboard.

### Does anything from the original video appear in the result?

No. SceneMixer reads only the structure; faces and footage are never reused.

Updated November 3, 2026


---

# Reference Image Flagged as a Real Person's Face: Refund and Fix

URL: https://scenemixer.com/guides/reference-image-flagged-as-a-real-persons-face/
A segment fails because a character sheet or an uploaded photo looks like a real person, often a famous actor. You are refunded; what to change before you try again.

# "A Reference Image Shows a Real Person's Face": Why, the Refund, and the Fix

Quick answer: The video provider checks the reference images of every segment, and refuses a generation when a face in them looks like a real person — an actor’s likeness, or a photo of someone. The segment shows The provider’s face check flagged a reference image as showing a real person’s face (an actor’s likeness, or a photo). Regenerate that character’s image with different facial features, or upload a different image, then retry. The credits are refunded automatically. Retrying without changing the image fails the same way, and repeated blocks pause video generation for a while.

Look at the character sheets used in the failed segment: the one that could pass for a known actor, or that is a real photo, is the one to change.

## Which image it is

Usually a character sheet in a realistic style whose face came out close to a well-known actor — common when a character is named after, or written like, a famous role — or a real photograph you uploaded as a character. It is almost always one of the character images attached to that segment.

## The refund and the pause

The segment is marked failed and its credits come back on their own; there is nothing to request. Like other content blocks, it counts toward the pause: after 12 blocked video attempts within a day, SceneMixer pauses video generation for 1 hour. So do not press retry on the same segment again and again: fix the image first. The pause rules.

## What to change

- Regenerate the character’s sheet with features that are clearly their own: a different hairline, a scar, a broken nose, freckles, glasses. Changing words that do not change the face (a different coat) will not help. Describe a face so it is not a generic face.
- If the name itself is a famous role, the face tends to drift back toward that actor however you describe it; a distinct look helps more than a longer description.
- If you uploaded a photo of a real person, replace it with an image of a character who is not a real, identifiable person. Using your own photo.
- Consider a stylised visual style for the project — anime, 3D or illustration faces sit far from any real person and are rarely flagged.

Then generate the segment again. Real people, brands and franchises.

## Character sheets can be refused too

The same check on SceneMixer can stop a character image from being made, for example when you redraw a sheet from a photo of a real person. That is refunded as well and does not count toward the video pause.

| Cause | Fix |
| Sheet looks like a famous actor | Regenerate it with distinct features |
| Character named after a famous role | Give them a distinctive look, not just more words |
| Uploaded photo of a real person | Replace it with a non-identifiable image |
| Realistic style keeps drifting to real faces | Try a stylised preset |

## FAQ

### Was I charged?

No. SceneMixer returns the credits automatically.

### Why does retrying not help?

The same image is checked again and refused again, and each block on SceneMixer counts toward the pause.

### Is my project in trouble?

No. It is a check on one image; change that image and carry on.

### Can I appeal?

Not on SceneMixer. The check is the provider's; changing the image is the way through.

Updated November 3, 2026


---

# SceneMixer Refunds: What Comes Back, What Waits, What Doesn't

URL: https://scenemixer.com/guides/refund-rules-what-comes-back/
Credits are held on submit, charged on success, returned on failure. Which failures refund themselves, which wait for review, what packs and plans refund, how to ask.

# Refunds on SceneMixer: What Comes Back Automatically, What Waits for Review, What Does Not

Quick answer: Credits are held when you submit, charged when the task delivers and returned when it fails. Most failed generations refund themselves; a few are held until support checks what happened; a delivered generation is never refunded. Credit packs and memberships are non-refundable except for a duplicate charge, a purchase we could not deliver, or a purchase by a minor without a guardian’s consent. To ask, use Still not solved? → Send feedback inside the Ask window, or email support with the order number, or the project, episode and segment; we reply within 7 business days.

Click your balance in the top bar, then View all: each task in your Credit History is one line, and its tag says which of the outcomes below it reached.

## How a task shows in your credit history

Each task is one line. Held: the moment you press generate, parse or storyboard, the credits move from available to held, and the line says In progress. Nothing is spent yet. Charged: the task delivered and the hold became a charge; the tag goes. Refunded: the task failed, we confirmed the provider was never billed, and the hold went back to available; the amount is struck through and tagged Refunded. A task held for review is tagged Under review. The rarer lines (a reversal, a support adjustment) are explained in Reading Your Credit History.

## Generations: three outcomes

Refunded automatically, no request needed. The task’s line in your SceneMixer credit history shows Refunded. This covers the cases where the model never received the request or the provider confirmed it did not bill us: the provider was busy or rate-limited, or the task timed out while still waiting its turn; the input was refused as invalid (for example too many reference assets); a provider-confirmed moderation block on a video, up to 12 blocks per account in 24 hours; a sheet description stopped by the check (these do not count towards that number); an AI Create script that exceeded the parse limit and never became a project; a failed remake read; Stop on an AI Create, parse or storyboard job that had not been sent to the model yet; and, when you delete an episode or project, work that had not started or had not reached the model. When your balance is short, the buy-credits window opens and nothing is submitted, so there is nothing to refund.

Held for review. The task’s line in your credit history says Under review. The request may have reached the model but the result cannot be confirmed: the service was interrupted mid-generation, a request timed out after it was sent, a parse or storyboard answer came back incomplete twice, the delivery step failed after the clip was made, the provider cannot report a result, or a moderation block past the daily number. Refunding blindly would pay the provider twice, so a person looks. Support reviews held tasks; the line then turns into Refunded or a normal charged line. Do not resubmit the same segment in the meantime — it would hold a second amount.

Charged, not refunded. A delivered generation is paid for, even when you dislike it: a drifting face, the wrong action, lip sync that is off. The fix is to edit the storyboard or the description and regenerate, which is a new charge — see When a Shot Comes Out Wrong. Also charged: Stop pressed on a storyboard, parse or AI Create job after it reached the model (a clip that is generating cannot be stopped at all); work already sent when you delete an episode, and clips already sent when you delete a project that the provider then delivers (other work already sent in a deleted project is held for review); free allowances (the account’s free sheets and the first project’s first storyboard) once used, even if you delete and rebuild; and the AI badge on a clip, which is not a defect — on a paid plan, new clips come without it by default.

## Credit packs

Credits count as delivered the moment they land, so SceneMixer packs are non-refundable, with three exceptions: the same order charged twice (the extra charge goes back to the original payment method); a purchase you could not use for more than 24 hours because of a technical fault on our side (tell us within 7 days); and a purchase by a minor without a guardian’s consent, refunded on the guardian’s request net of the value already used. Refunds go to the original payment method or to credits, normally within 7 business days of approval. Repeated buy–use–refund cycles are declined. The full wording is in the Refund & Cancellation Policy.

## Memberships

Card and PayPal subscriptions can be cancelled any time from the Membership page; access runs to the end of the paid period and the current period is not refunded pro rata. An upgrade takes effect immediately and charges the prorated difference to the same card or PayPal account; the page shows the amount before you confirm. When a membership payment is refunded in full, everything that order granted is taken back, including the first-subscription bonus, and the subscription behind it is cancelled; any shortfall is recorded and deducted from later membership grants. A partial refund puts the membership on hold until support reviews it. How cancelling and upgrading work step by step: Subscribe, Upgrade or Cancel a Plan.

## Chargebacks

Contact us before opening a dispute with your bank; most issues are resolved within days. A chargeback on a delivered service reverses the credits and may suspend the account until the matter is resolved.

## Moderation, repeated blocks and suspended accounts

SceneMixer refunds a single blocked generation when the provider confirms the block; past the daily number it is held for review instead. More than 12 blocks within 24 hours pause video generation for 1 hour, and the count runs for 24 hours from the first block, so each further block in that time pauses again. Content rules are zero tolerance: deliberate or repeated violations, getting around moderation, tampering with the AI label or using several accounts to dodge a block lead to removed content, restricted features and account suspension; a balance left on a suspended account is handled under the Terms (sections 03 and 09). What is not allowed and how to appeal: Content Rules and Moderation.

## How to ask, and how long it takes

1. Find the task’s line in your credit history (it names the project, episode and segment) or the order number (packs and memberships; it is on the receipt). 2. On the site: Ask AI (the speech-bubble icon in the top-right group, left of the settings gear; also in the avatar menu) opens the Ask window; if that does not solve it, Still not solved? → Send feedback, category Billing & credits. Include the project, episode and segment (or the order number), the time and the message you saw, with a screenshot if you can. 3. Or email support from your account address; the address is in the footer under Contact. 4. We reply within 7 business days. A refund made by support turns the task’s line into Refunded. How to get help.

| What you see | Where the credits went | What to do |
| A line tagged Refunded in your credit history | Back in your balance | Fix and retry; for a moderation block, change the text first |
| A line tagged Under review | Held until support settles it | Wait; do not resubmit the same segment |
| Card shows Done | Charged | Regenerate for a new take (a new charge) |
| The buy-credits window opened instead | Nothing was held | Top up and press the button again |
| Stop on a parse, storyboard or AI Create job not yet sent | Back in your balance | Edit and run it again |
| Stop after the job reached the model | Charged | — |
| A clip that is generating | Cannot be stopped | Let it finish, then regenerate if needed |

## FAQ

### Was my moderation failure refunded?

The card on SceneMixer only shows a short reason; hover it for the original message. Your credit history answers it: Refunded on the task's line means yes, Under review means a person is checking.

### I generated six segments at once and they all failed. Where are the credits?

Any you could not afford were never submitted: the buy-credits window opened instead. The others follow the table: refunded, held for review, or charged if they had delivered.

### I switched model tiers and regenerated. Do I pay again?

Yes. Every generation on SceneMixer is its own charge, whichever tier it runs on.

### Is the parse fee refundable?

A parse that fails before reaching the model refunds itself; one whose answer came back incomplete is held for review. After a successful parse, regenerating sheets or storyboards on SceneMixer is a separate charge.

### I deleted the project. What about the held credits?

Work that had not started, or had not reached the model yet, is refunded. SceneMixer settles clips already sent by what the provider reports; other work already sent is held for review, as is any task already held.

### Can I get money back instead of credits?

Only for the pack exceptions above (duplicate charge, unusable purchase, minor). SceneMixer settles everything else in credits.

Updated November 3, 2026


---

# Re-parse an Episode: What You Lose and What You Keep | SceneMixer

URL: https://scenemixer.com/guides/regenerate-a-storyboard-what-you-lose/
Re-parse rewrites an episode's storyboard from its outline. What it discards (segments, videos, versions, hand edits), what survives, what it costs, and when to use it.

# Re-parsing an Episode: What You Lose and What You Keep

Quick answer: Re-parse on an episode card writes that episode’s storyboard again from its outline. The confirmation says what goes: every storyboard segment in the episode, including videos you have already generated, which are not kept and cannot be recovered; credits already spent are not refunded. What survives: the cast and their images, every other episode, and any composite you already made. It costs the storyboard fee again. Use it when the breakdown is wrong throughout; when one shot is wrong, edit that shot instead.

Open a finished episode: ten segments, ten videos. Re-parse would replace all of them at once.

## Where the button is

Step 3, on the card of an episode that already has a storyboard, where Parse Storyboard used to be. Pressing it opens Confirm Re-parse, which spells out the loss before you confirm.

## What is discarded

Every segment of the episode: the storyboard text, including every edit you made by hand. Every video generated for those segments, and their version histories. Per-segment choices such as which state image a segment used. The credits spent on those videos are not refunded. Re-parse is refused while any segment is still rendering; wait or stop those first.

## What survives

The cast, locations and props in SceneMixer’s step 2 and their images. Episode-level choices such as an image you set with Edit (this episode). Every other episode. An episode composite you already made — it is a separate file and stays downloadable. The outline the new storyboard is written from.

## What it costs

The storyboard fee for that episode, priced by its script length — first and repeat generations cost the same. The videos you then generate are priced as usual. Storyboard pricing.

## When to use it

When the breakdown itself is wrong: the wrong scene order, the wrong people in the wrong places throughout, a segment length you want to change, a type or costume setting you changed after the first parse. It is the right tool when most segments would need rewriting anyway.

When one shot is wrong, or three, edit them. Editing on SceneMixer is free, regenerating a segment is priced per segment, and every version is kept. Fixing one shot. When a segment needs to become two, insert one. Inserting a segment.

## Before you press it

Download any video you want to keep; it will be gone. Copy any storyboard text you wrote by hand. If you changed the segment length or a style setting, check that it is set the way you want, because the new storyboard is cut to the current settings.

| | Discarded | Kept |
| Storyboard text | All segments, including hand edits | The outline |
| Videos | All, with version histories | — |
| Credits spent on videos | Not refunded | — |
| Cast and images | — | All |
| Episode composite | — | Stays as a file |
| Other episodes | — | Untouched |

## FAQ

### Does Re-parse refund the videos it discards?

No. SceneMixer does not refund credits already spent; the confirmation says so.

### Can I get a discarded video back?

No. Download what you want to keep before re-parsing.

### Does it affect other episodes?

No. Only the episode whose card you pressed.

### Why is Re-parse refused?

A segment in the episode is still rendering video. Wait for it or stop it first.

### Can I re-parse to get a different segment length?

Yes. Set Segment length first, then Re-parse; SceneMixer cuts the episode again at the new length.

Updated November 3, 2026


---

# Regenerate One Sheet and Write a Description That Works | SceneMixer

URL: https://scenemixer.com/guides/regenerate-one-sheet-and-describe-it/
Don't like a character's look? Redo one character, scene or prop image: what to write in Appearance and Clothing so the change shows, and what happens to states.

# Regenerate One Sheet, and Write a Description That Changes It

Quick answer: Open the card, edit Appearance and Clothing, Save, then Regenerate. The image is made from those two fields, so write what can be seen: face shape, hair, eyes, age, build, a distinguishing mark; one outfit with its cut, colours and materials. Adjectives about mood do nothing. After you edit, the card says Text edited — image is from before until you regenerate. Each regeneration is priced like any image; the free allowance covers it while it lasts.

Open a finished episode and click a character in the left panel to read a description that produced a usable sheet.

## Opening the card

In step 2, click the entry. The details show Name, Role, Appearance (the description), Clothing, the image with Regenerate, Upload & Replace and AI Edit Image, and for characters the voice. Scenes and props have a description instead of appearance and clothing.

## What to write in Appearance

What a camera would see, in order: face shape, skin, eyes and brows, nose and mouth, hair colour and cut, age, height and build, and one thing that makes this person recognisable at a glance — a scar, glasses, a streak of grey, a gap in the teeth. Concrete beats pretty: “square jaw, heavy brows, cropped grey hair, sixty, broad” gives the image something to draw; “handsome and imposing” gives it nothing.

Leave out mood and backstory. Leave out what the character is doing. Those belong in the storyboard, not on the sheet.

## What to write in Clothing

One outfit: the pieces, the cut, the colours, the materials, the accessories. “Charcoal wool overcoat over a white shirt, no tie, black boots, a steel watch.” If the story dresses her differently later, do not put both outfits here; the sheet would try to show both. Give the second outfit a state instead. States.

If the field reads Appearance generated, but clothing was not detected — please fill it in, do: SceneMixer dresses a character with no clothing line at random.

## Scenes and props

A scene description names what is in the place and what it is made of: the room’s shape, the furniture that matters, the floor and walls, the windows, where the light comes from. No people. A prop description names the object’s shape, size, material, colour and markings, and nothing about who holds it. Locations · Props.

## Save, then Regenerate

Save stores the text. The card now says Text edited — image is from before: the image still follows the old text until you press Regenerate. Regenerating replaces the image on the card; the headshot and half-body crops are cut again from the new one. If the new image is worse, regenerate again — each try is a fresh draw from the same text, and the price is on the button.

## What happens to states

State images made from the old main image are marked Main image changed — regenerate. They keep working until you do, but they will not match the new face. Regenerate them after you are happy with the main image, not before.

## Two other ways to change an image

AI Edit Image keeps the same character and applies a short change you describe — a navy shirt, short curly hair, a smile — without rewriting the description. AI edits. Upload & Replace puts your own image on the card. Using your own photo.

| Field | Write | Do not write |
| Appearance | Face, hair, eyes, age, build, one recognisable mark | Mood, backstory, actions, “beautiful” |
| Clothing | One outfit: pieces, cut, colours, materials, accessories | A second outfit (make a state) |
| Scene description | Space, furniture, materials, windows, light source | People, events |
| Prop description | Shape, size, material, colour, markings | Who holds it, what it does in the plot |

## FAQ

### I changed the description but the image is the same.

Saving stores the text; the image changes only when you press Regenerate. The card says so until you do.

### Does regenerating cost credits?

SceneMixer prices it like any image, on the button. The free allowance covers it while it lasts.

### Can I keep the old image?

Regenerating replaces the image on the card. If you may want the old one back, download it first.

### Why did my character come out in the wrong clothes?

Either the Clothing field is empty, or it describes two outfits. Write one; give the other a state.

### What is the stale notice on my states?

The main image changed, so states made from the old one no longer match. Regenerate them once the main image is right.

Updated November 3, 2026


---

# How to Remake a Video with AI, Step by Step | SceneMixer

URL: https://scenemixer.com/guides/remake-a-video-with-ai/
Upload a clip, let SceneMixer read every shot, then reshoot the same beats with your own cast. What the read-through produces, what you can change, what it will not do.

# How to Remake a Video with AI

Quick answer: Upload an MP4 or MOV (within the length and size limits shown on the upload box), and SceneMixer watches it shot by shot: where the cuts are, who is on screen in each one, what they do, and what they say. It turns that into a cast list, a set of locations, a prop list and a finished storyboard — then hands you the normal editor to reshoot it with your characters. The source clip is never given to the video model, so the faces that come out are the ones in your own character sheets.

Everything below is shown on a real project you can open: the sample drama in the editor · the finished episode. No sign-in needed.

## What you need before you start

One video file, MP4 or MOV, within the length and size limits shown on the upload box. It can be something you shot yourself, something you have the rights to, or a clip you are using as a reference for blocking and pacing. Nothing else — no script, no character descriptions, no shot list. Producing those is the whole point of the read-through.

A fast-cut clip has more shots to read than a slow one of the same length, so it takes longer — but it does not cost more. SceneMixer prices the read-through by the length of the clip alone, rounded up to the next minute; the number of cuts never enters the price.

## Step 1 — Upload the clip

Open the home page, switch to the Remake tab, and pick your file. The upload starts the moment you choose it, so the price appears before you commit to anything. If you change your mind, remove the clip and the storage is handed straight back.

You can also type a sentence about the clip. It is a hint, not an instruction — something like “the woman in grey is the lead, the two men are brothers”. The model cannot tell that from pictures alone, and where the sentence disagrees with what is on screen, what is on screen wins.

## Step 2 — Watch the read-through

This is the part that takes a few minutes. It runs in four passes, and the progress screen names each one as it happens: find the cuts, identify the cast, read every shot, then check and save.

What comes out the other end is a complete project: every person who appears becomes a character with a written appearance and a voice description, every place becomes a location, every prop that matters becomes an item, and every shot becomes a numbered beat inside a segment — with its framing, its camera move, who is in frame, and the line that is spoken.

You can see exactly what that looks like in the sample episode: the left column is the cast the read-through produced, the middle column is the storyboard, and each coloured chip is a real link back to a character, a location or a prop.

## Step 3 — Choose how it should look

The read-through tells you what happens. You decide what it looks like. On the first step of the wizard you pick the render style, the art direction, the wardrobe, and a target region if you want the cast moved to a different part of the world. Ethnicity, faces, architecture and street signs follow the region you pick; names do not — they are already written into the storyboard the read-through produced, so renaming a character is a manual edit in the next step.

This is the point where a remake stops being a copy. The beats are the source clip's; the people, the clothes, the architecture and the language are yours.

## Step 4 — Generate and composite

From here it is the normal pipeline: generate the character sheets, then shoot segment by segment, then composite the episode. Each button shows what it costs before you press it. Note that the free allowance a new account gets on its first project does not apply to remakes — SceneMixer charges a remake at list price from the first setup image.

Segments are short on purpose — four to fifteen seconds each. That is one generation call, and it is also the window inside which a video model can hold a face steady.

## What a remake will not do

It does not copy the original footage. The source clip is read and then set aside; it is never handed to the video model as a reference frame. That is deliberate: a model given a reference frame copies the faces in it, and copying a real person's face is the one thing we will not build.

It does not reuse the original audio. The lines are heard, then rewritten in your project's language and spoken by your own cast's voices.

It is not a licence. Reading a clip does not give you the rights to its story. Use footage you own, footage you have cleared, or footage you are only using to study timing.

| What | Limit |
| Source file | MP4 or MOV, within the limits shown on the upload box |
| What it produces | Cast, locations, props, one episode of storyboard |
| Segment length | 4–15 seconds each |
| Video price | from $0.06 per second of finished video |
| Compositing | 1 credit per second |
| Free allowance | Not applicable — the first-project allowance is for ordinary dramas |

## FAQ

### Does SceneMixer copy the faces from the video I upload?

No. The clip is read into a written storyboard and then set aside. It is never passed to the video model as a reference image, so the faces in the finished episode come from your own character sheets, not from the source.

### What happens to the clip after the read-through?

It stops being needed as soon as the read-through has been saved, and the storage it used is released back to your quota automatically.

### Can I remake something in a different language?

Yes. The lines are re-heard and written in the language your project is set to, and the finished episode is spoken by the voices attached to your own characters.

### How long does the read-through take?

A few minutes for a short clip. It scales with the number of cuts rather than with running time, so a fast-cut clip takes longer than a slow one of the same length.

### What if the read-through gets something wrong?

Everything it produced is editable. Rename a character, rewrite a shot, delete a segment, change the location on one beat — the storyboard is a working document, not a fixed output.

Updated November 3, 2026


---

# Remake: How Dialogue and Music Are Read From Your Clip

URL: https://scenemixer.com/guides/remake-how-dialogue-and-music-are-read-from-the-clip/
A remake read listens as well as looks. How each shot is checked for speech, why heard lines are quoted, what happens to music, and what to do when a line is missing.

# Remake: How Spoken Lines and Music Are Read From the Clip, and What Ends Up in Your Episode

Quick answer: Every shot of the clip is listened to twice: first by a speech recogniser that answers ‘is anyone speaking, and what did they say’, then by the model that describes the shot, which is given the recognised words to work from. Lines are therefore quoted, not guessed from lip movement, and a shot where nobody speaks is written as silent. If the clip has music, the read notes its character, and that becomes the tone the episode’s own score is written in: it is planned from the episode’s storyboard — usually one to three pieces — and mixed in under the voices, as in any drama. The dialogue language is chosen on the upload screen and fixed once the read starts: lines are written in it, translated if the source speaks another. A line the recogniser could not make out is simply missing; compare with your clip and add it to the shot in the editor once the storyboard is generated.

After a read, open the episode in Step 3 and read its outline: each scene block lists its lines in speaker-colon form.

## How lines are heard

Speech recognition runs on each shot’s own audio. What it hears is handed to the shot description as ‘these words were spoken here’; the describer then names the speaker from the faces and writes the line into the shot. Music under speech is handled; shouting over a crowd, heavy accents and very quiet lines are where it fails, and those are left out rather than invented. Lines then appear in the outline like a script’s: Write Dialogue So It Is Picked Up.

## Why quote, not guess

Left to itself, a vision model asked ‘what is said here’ will sometimes hear and sometimes not, on the same shot, and will invent plausible lines when unsure. Recognised words pinned into the description stop both; the same words are used every time the shot is described.

## Music

The read asks once whether the clip has music and what kind. If it does, that description becomes the tone the episode’s score is written in, so your remake sounds like the same kind of film; the source track itself is never used. The clips are then generated without music of their own and the score is mixed in under the voices, exactly as in a drama. When an episode is not being scored — an older episode, or one whose plan could not be made — the video model gives those clips their own music instead, as before. Ambient sound from the clip is noted but not carried. Background music in an AI short drama.

## Dialogue language

Chosen on the upload screen, defaulting to your interface language (the Chinese site is always Chinese), and fixed once the read starts. Lines heard in another language are translated into it. It cannot be changed afterwards for a remake, which is why Step 1 does not offer the dropdown. After the Read.

## When a line was missed or wrong

The outline itself cannot be edited on SceneMixer, and the storyboard copies its lines word for word. Generate the storyboard, then open the segment where the line belongs and add or fix it there, with its speaker; a line given to the wrong person is fixed the same way. Editing is free. Generate the Storyboard.

| In the clip | In the read | In your episode |
| A clear spoken line | Quoted, speaker named | A line in the outline |
| Nobody speaks | Shot marked silent | No line |
| Unclear speech | Left out | Add it in the segment by hand |
| Music | Character noted | Tone for the episode’s own score, mixed under the voices |
| Ambient sound | Noted | Not carried; the storyboard writes its own |

## FAQ

### Why is a line missing?

The recogniser could not make it out. Compare with your clip and add it in the segment once the storyboard exists.

### Are the source voices used?

No. Voices come from the cast's voice settings, as in any project.

### Can the remake speak a different language than the source?

Yes; pick the dialogue language on the upload screen and the lines are translated at the read.

### Is the source music reused?

No. Its character becomes the tone of the episode’s own score, which SceneMixer plans from the episode’s storyboard — usually one to three pieces — and mixed under the voices.

### Why does a crowd shout get no line?

Group noise is treated as background sound, not dialogue.

Updated November 3, 2026


---

# Remove a Moment From the Middle of an AI Video Clip | SceneMixer

URL: https://scenemixer.com/guides/remove-a-moment-from-the-middle-of-a-clip/
Cut a stumble, a frozen beat or an odd gesture out of the middle of a finished segment and join the rest. How Trim's Remove mode works, what the join looks like.

# Remove a Moment From the Middle of a Clip

Quick answer: Open Edit this clip → Trim and switch the mode from Keep to Remove. Now the two handles mark the part to cut out: drag them around the bad moment, press Preview to watch the clip with that part skipped, and Save. What is before and after is joined with a straight cut, picture and sound together. It costs 5 credits and saves a new version; the original is kept.

Use it for: a hand that glitches for half a second, a dead pause between two lines, a moment where someone looks into the lens.

## Keep and Remove are the same track read two ways

In Keep mode the handles frame what stays, so everything outside them is cut — that is how you trim the start and the end. In Remove mode the handles frame what goes, so the outside is kept and the two outer parts are joined. If the handles are still at the two ends when you switch to Remove, they jump to frame the middle third of the clip, so you start from a real selection; if you had already moved them, they stay where they are. Check the track after switching.

## Step by step

- Select the segment. Under the player, Edit this clip is open on Trim.
- Click Remove.
- Drag the left handle to just before the moment you want gone and the right handle to just after it. The player shows the frame under the handle as you drag.
- Nudge with − and + (half a second a press) if you need to.
- Press Preview: the clip plays with the removed part skipped, so you can hear whether the join lands well.
- Press Save (5 credits). The new version appears in Version History as Trimmed.

## What the join looks like

A straight cut: the frame before the removed part is followed directly by the frame after it. When the camera and the people are still, the jump is small; when someone is moving, it shows as a jump. Cut where the motion pauses — between two lines, at the end of a gesture, at a blink — and the join is hardest to notice. The sound is cut at the same points, so a word spoken inside the removed part disappears with it.

## How much you can remove

SceneMixer lets you remove anything up to all but one second of the clip. When you drag the handles so far apart that less than a second would remain, the panel holds them back and says Keep at least 1s. If both handles sit together, nothing is removed. A cut that ends right at the start or the end of the clip is the same as trimming that end in Keep mode.

## One cut per save

Each save removes one stretch. To remove two separate moments, save the first cut, then trim the trimmed version again for the second. Each save costs 5 credits, and the second trimmed version replaces the first — the segment keeps one trimmed version. How versions are kept.

## When to fix it another way

If the glitch is in how the shot was framed or who is in it, removing it only shortens the problem. Change that shot in the storyboard and regenerate the segment. If the problem is a wrong word rather than a wrong moment, use Redub. Trim, AI visuals or Redub.

| You want to | Mode | Handles go |
| Drop the first second | Keep | Start handle one second in |
| Stop right after the last line | Keep | End handle just after the line |
| Drop a frozen beat in the middle | Remove | Either side of the beat |
| Drop two separate moments | Remove, twice | One stretch per save |

## FAQ

### Is there a crossfade at the join?

No, it is a straight cut. Choose a still moment for a clean join.

### Does removing a part change the dialogue?

Only if a line was spoken inside the removed part; that sound goes with it.

### Can I remove two parts at once?

No, one per save. Save, then trim the result again.

### Will the segment be shorter in the episode?

Yes. Composite and download use the trimmed version while it is active, at its new length.

Updated November 3, 2026


---

# Renewal Payment Failed: How to Restore Your Plan

URL: https://scenemixer.com/guides/renewal-payment-failed-what-to-do/
What you see when the automatic renewal of your plan does not go through: the Payment failed chip on the plan card, the once-per-session dialog, what Pay and restore and Choose a different plan each do, when the plan comes back, and how it differs from a failed one-off payment.

# Renewal Payment Failed: What You See on Your Plan Card, and How to Get the Plan Back

Quick answer: when the automatic renewal of your plan does not go through, the paid period runs out and the plan drops back to free. You get a small red Payment failed chip on the plan card, and once per browser session a dialog that offers to settle it. The dialog’s main button depends on how you pay: with a card it opens the payment page for the unpaid bill — where you can also use a different card — and with PayPal it takes you to PayPal to update your payment method before the next automatic retry; if PayPal has paused the subscription it points you at support instead. The second option, Choose a different plan, drops the unpaid plan without charging you and takes you back to the plan list. Nothing you already have is lost: your credits, projects and files stay exactly where they are.

This page is about the automatic renewal of a plan. A payment that fails when you are buying credits for a single order is a different story: Payment Failed? Declined Card, PayPal, Missing Credits.

## What you will see

Three things mark a failed renewal, and you may meet any one of them:

- A red Payment failed chip on the plan card. It sits on the card of the plan that failed to renew, in the membership page. Press it and the same dialog opens. It stays there until the bill is settled or the plan is dropped.
- A dialog when you sign in. Once per browser session and per account, and only after the screen is clear — it will not cover the first-run tour or another dialog, and it steps aside while something more important is open.
- A different button on that plan card. Instead of the usual subscribe button it reads Pay to restore or Fix payment, and pressing it opens the dialog rather than starting a new purchase. The same happens if you press Upgrade anywhere in the membership page or the top-bar plan menu: instead of an error you get the dialog, because the thing you need is to settle the old bill, not to buy a new plan on top of it.

While this is unresolved the membership page shows the free plan as your current one, and the paid plan’s own perks are off — depending on the plan, that is the higher number of tasks at once, the priority queue, removing the SceneMixer badge from your videos, and the discount. Your credits, your projects, your uploaded files and your finished videos are not touched. What a paid plan changes lists what each plan includes.

## What the dialog looks like

The title is Your <plan> plan is on hold, or Your membership is on hold when the plan name is not known. Below it, one line says that the last renewal did not go through and that settling it restores the membership — and that nothing is lost. Then a dark card shows the plan name, the credits that plan grants each month, two or three of its perks, and the amount still due. Under that is the main button, and under the button two smaller ones.

Nothing on this screen charges anything by itself. Pressing the main button opens a page at the company that takes the payment for your plan; you settle it there, in a new tab.

## What each button does

### Pay and restore — card payments

With a card, the button reads Pay <amount> & restore and opens the payment page for the unpaid bill in a new tab. That page belongs to the payment company, not to this site: it shows the amount and lets you pay with the card you used before or enter a different one. Either way the bill is settled. A line under the button says the payment is handled securely by the company that handles card payments here, and that you can change your card on the next page. Come back to this site afterwards; the plan comes back on its own, as described below. Subscribe, Upgrade or Cancel a Plan covers what the site can and cannot do about the card.

### Update payment in PayPal — when PayPal will try again

If you pay with PayPal and PayPal has not given up on the subscription, it collects the money itself, so there is no bill for us to hand you a payment page. The dialog says that the renewal did not go through and names the date PayPal will try again; the button reads Update payment in PayPal and opens PayPal’s automatic payment page in a new tab. Change the card or the funding source there before that date and the retry goes through on its own. If the retry runs without a change, it fails the same way.

### Contact support — when PayPal has paused the subscription

PayPal stops retrying after a certain number of failures and puts the subscription on hold itself. There is then no retry date and nothing to update; the dialog says so and the button becomes Contact support, which opens an email to us. You can also drop the paused plan with the other option below — that cancels the suspended subscription and its unpaid bill, and nothing is charged for it. In this state PayPal does not try the payment again.

### Choose a different plan

This drops the unpaid plan instead of paying for it. Press it and a short confirmation appears saying exactly what will happen: the plan that failed to renew and its unpaid invoice are cancelled, and you will not be charged. Cancel & choose does it; Back returns to the dialog. After it goes through, the page tells you to pick a new plan and scrolls you to the plan list (or takes you to the membership page, where it scrolls there). Your credits, your projects and your files are not affected by dropping a plan, and the button cannot be pressed twice into two cancellations.

### Later, and the small cross

Later and the cross in the corner do the same thing: close the dialog and change nothing. The unpaid bill stays where it is, and the Payment failed chip is still on the plan card whenever you want to pick this up again. Choose this when you would rather sort the card out in the payment company’s own account first.

## After you pay: when the plan comes back

SceneMixer restores the plan when the payment company confirms the payment to us, not when you close its page. That normally takes well under a minute. When it lands, the new period is opened, that period’s credits are granted as usual, and both the chip and the dialog disappear on their own — refresh the membership page if you are looking at it. Nothing has to be pressed again.

If you paid and the chip is still there after a while, refresh first. If the money has clearly gone and the chip is still showing after a good while, send support the billing reference from your receipt; they can check it against the payment company and put it right. How to get help says what to put in the message.

## If you do nothing at all

The payment company keeps retrying on its own schedule for a while — that is why the PayPal dialog tells you the next retry date. Two things can happen:

- A retry succeeds. Nothing more for you to do: the period is opened, the credits are granted, and the prompt stops. You never have to reopen the dialog.
- Every retry fails. What happens next depends on how you pay. With a card, the company stops retrying and ends the subscription: the chip goes with it, you are back on the free plan with all your credits and work intact, and you can subscribe again whenever you like. With PayPal, the subscription is paused rather than ended: PayPal stops trying, but the chip and the dialog stay, and the dialog’s button becomes Contact support. To clear a paused subscription yourself, use Choose a different plan — that cancels it without charging you.

One corner case worth knowing: if a retry that was already on its way goes through after you have dropped the plan, that payment is not thrown away. It is honoured — the plan you dropped is given one period and that period’s credits are granted.

## Changing the card, or the company you pay with

| You want to | Where to do it | What happens |
| Use a different card | On the payment page the dialog opens | The bill is settled with the new card and the plan returns |
| Update the card on file | On the payment page the dialog opens; at any other time, contact support — there is no card-update screen on the site itself | The plan is untouched; the next renewal uses the new card |
| Change the PayPal payment method | In PayPal, from the button in the dialog | The next automatic retry uses the new method |
| Switch to another company entirely | Choose a different plan first, then subscribe again | The unpaid plan is cancelled without a charge, so the new plan is not blocked by it |

The order matters in the last row: while the old plan is still attached with an unpaid bill, buying through a different company is refused, because the site keeps one plan per account. Dropping it first is what clears the way. Pressing Upgrade during this state does not get you a new plan either — it opens the same dialog, which is the intended route back to paying or dropping.

## Not the same as a failed one-off payment

| | Renewal of a plan | Buying credits or a one-off order |
| What failed | The automatic charge for the next period of a plan you already hold | A payment you started yourself |
| Where it shows | A Payment failed chip on the plan card, and a dialog once per session | A message in the payment window, and the order sitting under pending orders |
| What you do | Settle the bill or drop the plan, in the dialog | Try the card form again, use another card or PayPal, or wait for a clearing payment to land |
| Is anything charged meanwhile | The company retries on its own schedule; nothing is charged twice | An unpaid order expires by itself without charging anything |

Payment Failed? Declined Card, PayPal, Missing Credits covers the one-off side in full, including a payment that cleared but whose credits did not arrive.

## FAQ

### What does the Payment failed chip mean?

The automatic renewal of your plan did not go through, so the paid period has ended. Nothing is lost: your credits, projects and files stay, and settling the unpaid bill brings the plan straight back.

### Why did I get a dialog when I only signed in?

So you find out before the payment provider gives up on its own. The dialog appears once per browser session and per account, and it waits until the screen is clear rather than covering the first-run tour or another dialog.

### It popped up again after I closed it. Is that a bug?

No. The automatic prompt is once per browser session; if a retry fails later and you come back, you can be prompted again. The chip on the plan card never goes away on its own until the bill is settled, and it reopens the same dialog whenever you press it.

### What does Pay and restore do?

It opens the payment provider's own page for the unpaid bill in a new tab. Paying there, with the same card or a different one, settles the bill; the plan returns as soon as the provider's confirmation reaches us. SceneMixer charges nothing for pressing the button itself.

### I pay with PayPal. Why does my button take me to PayPal instead?

PayPal collects a subscription payment itself, so there is no bill for us to hand you a payment page for. While the subscription is still active PayPal will retry on its own; the dialog names the date it will try again, and the button takes you to PayPal to update your payment method before then.

### What does Choose a different plan do?

SceneMixer cancels the unpaid plan and its unpaid invoice without charging you, then takes you back to the plan list to pick a new one. A confirmation says exactly that before anything happens. Credits and projects are untouched.

### What happens if I do nothing at all?

The payment provider keeps retrying on its own schedule for a while, and if one of those retries succeeds the plan returns by itself. After the last retry: with a card, the subscription ends and the chip goes with it; with PayPal, the subscription is paused, so the chip and the dialog stay until you contact support or drop the plan. Until the bill is settled or the plan is dropped, the plan's own perks are off while your credits and everything you made stay as they are.

### I cancelled the plan and then a retry went through. What now?

That can happen when a retry was already on its way before you cancelled. It is honoured: the plan is given one period and that period's credits are granted.

### How is this different from a payment that fails when I buy credits?

Buying credits is a one-off order: a failed card form, a declined card or a pending order, all covered by Payment Failed? Declined Card, PayPal, Missing Credits. This page is about the automatic renewal of a plan, where an unpaid bill sits on the subscription until it is settled or the plan is dropped.

### I changed my card but the chip is still there. Now what?

Refresh the membership page. If the bill was actually paid and the chip is still showing after a good while, send support the billing reference from your receipt so they can check it against the payment provider.

Updated November 3, 2026


---

# Reuse a Character, Scene or Voice Across Projects | SceneMixer

URL: https://scenemixer.com/guides/reuse-a-character-scene-or-voice-across-projects/
How to move a character, scene or prop from one project into another from step 2: the whole entry comes over, and you can restore the original.

# Reusing a Character, Scene or Voice Across Projects

Quick answer: In step 2, open the entry you want to fill, press Edit, then Pick existing. The asset library lists every entry of the same kind — character, scene or item — that has a finished image, from all your projects, searchable by name or project. Pick one, read the preview of what it will replace, and press Replace entirely: the name, description, images, voice and state images all move over, the name is updated throughout the project, and what was there is kept so you can restore it.

Why the same sheet gives the same face is what makes reuse worth doing.

## Where reuse starts

Step 2: open a character, scene or item and press Edit. Among the image actions is Pick existing, which opens Pick from Asset Library. The editor’s Add assets picker is a different list — this project’s entries, added to one episode — and does not reach into other projects. Adding to an episode.

## Step by step

- Go to step 2 of the project you are putting the entry into, and open the card you want to fill — the character, scene or item to be replaced.
- Press Edit, then Pick existing.
- The library lists entries of the same kind, from every project you own, that already have a finished image; the newest come first. Type in the search box to filter by name or by the project it comes from, and press Load more to go past the first screenful.
- Pick one and read the bar at the bottom: the source image, the old name and the new one, the project it comes from, and what moves over — images, description and traits, voice, state images.
- Press Replace entirely. The swap happens in place: the entry keeps its position in the project and every shot that names it.
- The confirmation names the new entry, and if the name changed it adds that the old name was updated throughout the project.

## What moves over

Everything. The name, the description, the images — the main one and the face and half-body crops made from it — the voice, and every state image with the name it carries. The traits written into the entry come too: clothing, personality, lighting and the rest. A scene or an item brings no voice; a character brings one whenever it has anything on the voice row, a recorded sample or a written description.

A copied image keeps the style it was drawn in — copy within one visual style, or regenerate the entry in this project’s style, where the description is what holds the face. Regenerating a sheet.

The result is a full entry of this project: rename it, rewrite its description, regenerate its image, add states. The project it came from is untouched, a later change there does not reach here, and there is no live link between the two.

## The name travels too

Because the name moves over, the replacement renames the entry — and the new name is written everywhere the old one stood: the storyboard lines that called it by name, the episode titles and summaries, the scene lists, and dialogue that had already been parsed. Your script text is left as you wrote it. If a few places could not be updated, the confirmation says so; check the storyboards and the outlines.

## Undo: restore the original

Press Edit on the entry you replaced. Under its image it now reads Replaced from your library. with Restore the original beside it. That puts back the name, description, images, voice and state images from before the reuse, and renames the entry back throughout the project; it asks for confirmation first. While the entry’s image is generating, the link is greyed out — wait for it to finish.

Restoring takes back one replacement: it puts the entry back to what was there just before your most recent Replace entirely — replace the same card twice in a row and restore takes it back to what was there just before the second replacement. If the restore link is still shown after you have used it, there is another replacement to take back; once the link is gone, the next Replace entirely brings it back.

## Into a project in another language

If the two projects are set to different output languages, the description, the voice description, the traits and the state names are translated into the target project’s language as they move over. Names are not translated: a character called Mara is still called Mara. The translation is free and adds a few seconds to the transfer. If it does not come through, the text moves over as it is and the confirmation says the description could not be translated; edit it by hand afterwards.

## What it costs

Nothing. Picking from the library, the replacement itself and the restore are all free, and no membership is needed — any account can use them. What you do with the entry afterwards is priced as usual: regenerating its image, making a state image, generating the clips that use it. What a paid plan changes.

## When it does not go through

A replacement is refused, before anything is written, in a few situations:

- The project is busy. A storyboard being generated, a script being parsed, episodes being prepared, the entry’s image or one of its state images being made — while any of that is running, the replacement is refused and you are asked to try again when it finishes. Nothing changes and nothing is charged.
- The entry was edited moments ago. A just-made edit on the card is still syncing; until it finishes, the replacement is refused. Wait a moment and press again.
- The name is already taken. If this project already has a character, scene or item carrying the source entry’s name, the copy does not happen. Rename the one you have, or pick another source, and press again.
- The slot is an ad’s main product. The product slot of an ad cannot be replaced by reuse — its image is tied to the product photo the ad was built from.

Two messages can also appear after a replacement that did go through:

- Replaced, but the description could not be translated into this project’s language. The text came over in its original language; edit the description by hand if that matters here.
- Replaced, but the old name could not be updated everywhere. The entry and its new name are in place; look through the storyboards and the outlines for a line that still reads the old name.

A restore can be refused the same way — a project still running, or the name it is putting back now sitting on another entry — and an entry that was never replaced from the library has no restore link at all. The link appears after a replacement, and each use takes back the most recent one — while it is still shown, there is another replacement to take back.

## When to use it

A new project that continues an old story — a second season, a spin-off, the same universe in another language — is where reuse earns its keep: the lead, the house and the signature prop arrive looking the same, and the shots that name them have something to hold on to. Making the next episode · A second-language version.

Inside one project you need none of this: every episode shares the project’s cast, scenes and items as soon as a storyboard names them. Reuse is for crossing from one project to another.

## Voices without a character

Pick from library on the voice row of a character is a different list: the preset voice library with previews and samples, not your projects. A voice you uploaded travels with the character that has it — reusing the character brings the voice. Character voices.

## Uploads

Images you uploaded to the library appear in this picker under their own name once they are on a card with a finished image. Uploading is a paid-plan perk on SceneMixer; picking from the library is not. The asset library.

| You want | Do | What moves over |
| The same lead in a new drama | Step 2, open the character, Edit, Pick existing, Replace entirely | Everything: name, images, description and traits, voice, states |
| The same room or house | Step 2, open the scene, Edit, Pick existing | Everything: name, images, description and traits, states |
| A prop from another drama | Step 2, open the item, Edit, Pick existing | Everything: name, images, description and traits, states |
| A guest in one episode only | Editor, Add assets | Into this episode; this project’s entries only |

## FAQ

### Can I use a character from another project?

Yes. In step 2, open the card, press Edit, then Pick existing: every entry of the same kind with a finished image, from all your projects, is listed. It is free and needs no plan.

### Does the voice come too?

Yes. A character’s voice always comes with it; there is no checkbox. A recorded sample moves over, or the written voice description if there is no sample.

### If I edit the copy, does the original change?

No. After the copy they are separate entries in separate projects.

### Why is my character missing from the list?

It has no finished image yet, or it is a different kind from the card you opened: characters, scenes and items each have their own list.

### Is reuse free?

Yes. Picking from the library, replacing and restoring all cost nothing on any SceneMixer plan.

### Can I undo a replacement?

Yes, if the entry was replaced from the library: press Edit and choose Restore the original. Each use takes back one replacement; if the link is still shown after that, there is another one to take back.

### What happens to the old name in the storyboard?

The new name is written everywhere the old one stood: storyboard lines, episode titles and summaries, the scene lists and already-parsed dialogue. Your script text is not touched.

### The two projects use different languages. Will the description be translated?

Yes. The description, the voice description, the traits and the state names are translated into the target project’s language; the name is not. If the translation fails, the text moves over as it is and you are told.

### What if the project already has an entry with that name?

The replacement is refused and nothing changes; rename one of the two entries and try again.

### Can I replace an entry while the project is generating?

No. Wait for the run to finish, then press again. A refused replacement costs nothing.

Updated November 3, 2026


---

# Revise the Whole Episode: One Request, Every Segment

URL: https://scenemixer.com/guides/revise-the-whole-episode-with-one-request/
One request applied across an episode: pick segments, review each revision, save what you keep. What it can change, what it costs, and how to resume later.

# Revise the Whole Episode: One Request, Every Segment

Quick answer: Write one request — “make every line colder and shorter”, “add rain sound under the scene” — and press Apply to the whole episode… in the AI edit panel. Pick the segments, press Start, and the same request runs on each of them. The results come back as a list: open a row to see the change, then save them all at once or one by one. Nothing reaches your script until you save. Segments that need no change, and requests the storyboard text cannot fulfil, are not charged — so the number on the button is the most a run can cost.

This is the same engine as AI edit on a single segment, run across many segments at once: same request box, same checks, same price per segment, same History. What changes is how many segments ride on one request, and how you review the outcome. Reach for the whole-episode run when the note applies from the first segment to the last — a way of speaking, a word someone keeps using, a sound that should sit under a whole scene, a pace that drags. Reach for single-segment AI edit when one segment is off and the rest are fine.

## What it does

You give it one request. It revises each selected segment against that segment’s own text — the same request, applied wherever you ticked it. Every segment comes back as one of three things: revised, needing no change, or failed; the list tells you which, and you decide what to keep. It never rewrites the episode as one block, and it never touches a segment you did not pick.

## Where to find it

- Open the project, go to step 3 Episode Videos, and press Start Editing on the episode’s card. The whole-episode run lives in the editor, not in step 3 itself.

- In the editor, every segment’s card carries an AI edit button (the wand). It opens the panel where a single segment is revised.

- Type your request in the box. When the episode has more than one segment to revise, an Apply to the whole episode… button appears beside Revise. It stays grey until the box has something in it.

- On a phone, the AI edit button on the Script tab opens the same panel as a sheet over the script.

The whole-episode button only appears for revising. If the segment on screen is blank, or written as plain prose instead of shots, the panel is in AI write mode and there is no whole-episode button — writing a segment from nothing is a one-segment job. To revise the episode instead, open any segment that already has shots.

## Step by step

- Open the episode in the editor. Project → step 3 → Start Editing on the episode’s card.

- Open AI edit on any segment’s card and write the request. The small chips above the box (lines, action, mood, time and weather, camera, keep someone out, voice-over, another scene) are sentence starters, not a menu.

- Press Apply to the whole episode…. A dialog opens with your request quoted at the top.

- Pick the segments. Everything it can revise is listed and ticked to begin with; the count sits above the list, with Select all and Clear. Untick whatever should stay as it is. The Start button carries Up to ✦ N — the most this run can cost.

- Press Start. A progress line counts the segments as they are revised (Revising 3 of 12…), and each row shows its state as it arrives: waiting, revising, Changes: 2, No change needed, Couldn’t revise.

- Stop whenever you like. The segments already revised stay in the list; the ones that had not started are dropped, and nothing is charged for them.

- Review each row. View opens the line-by-line diff — removed lines struck through, added lines in green. If the request changed the segment’s shots or seconds, the row also says the length is going from Xs to Ys and whether regenerating that segment’s video costs more than before.

- Save what you keep. Save all (N) writes every revised segment into the script as a normal edit, or press Save on a single row to keep only that one. Discard results throws the unsaved ones away — and warns that the credits already charged for them are not refunded. Close leaves them for later.

Saving changes the storyboard text only. The segment still plays the old take until you regenerate it — the same as any edit to the text. The summary line under the list keeps the count honest as you go: revised, no change needed, could not revise, and what the run has been charged.

## If you refresh, close the page, or come back later

Revised results that are not saved yet are kept on this device — this browser, on this computer — until you save or discard them, and for a day. Reopen any segment’s AI edit panel in that episode and there is a line: You have unsaved results from an earlier run. Press View to open the list again and save from there. The panel says it plainly: results are kept on this device until you save or discard them.

Each result is tied to the exact text it was made from. If a segment’s text changed in the meantime — a hand edit, a restored version — that segment’s old result is dropped from the list, so an earlier revision never lands on newer text. If every revised segment has changed, there is nothing left to resume. Another browser or another computer will not see the results at all.

One run at a time: starting a new run replaces the list. Save or discard what you want to keep before running another request.

## What it can change

Whatever single-segment AI edit can do, it can do to each picked segment:

- Lines — reword a line, cut it short, change its tone, stop one character calling another “boss”.

- Action and mood — what the characters do in a shot, how a scene feels.

- Time, weather, camera — the hour of day, the rain, what a shot is showing.

- Sound — the ambient line under a shot, an added voice-over line.

- The shots themselves — add, remove or merge shots and change their seconds. A segment can only hold so much: its total length stays within the segment’s limit, so a very long addition comes back shorter, or fails with a reason. When a change moves the timing, the row shows the new length before you save.

### Examples that work

| Request | What comes back |

| “Make every line colder and shorter.” | Each picked segment comes back with tighter, colder lines; shot counts and timing stay as they were. |

| “Add rain sound under the whole scene.” | Each segment’s sound line carries the rain; regenerate a segment and the rain is in the clip’s audio. |

| “Split the argument into more shots — closer angles.” | Shots are added inside the segments you picked; each row shows the segment’s new length, and whether its regeneration price changed. |

### Examples that come back as “No change needed”

| Request | Why, and where it does live |

| “Give the villain a black coat.” | A character’s face, hair, build and outfit come from the character image and its states, not from the storyboard text. The note points at the character; change the image or add a state. |

| “Use a different voice for her.” | Voices are picked on the character, not written in the segment. |

| “Turn the background music off.” | Music is not part of the storyboard text; the Music switch by the player turns it off. |

Two more that read the same way: the spoken language of a project is a project setting, not a line in a segment; and anything that appeared in a finished clip that the storyboard never wrote — an extra passer-by, a stray line — came out of generation, so a revision has nothing to remove. Regenerating that segment is what usually clears it.

## What it costs

Every segment that comes back revised is charged once, at the same per-segment price as a single AI edit (the price is on the pricing page). The number on the Start button — Up to ✦ N — is the ceiling for the segments you ticked, not a bill: segments that need no change are free, failed revisions cost nothing, and a segment is charged at most once per run.

Two things stop a run part-way. In both cases everything already revised stays in the list and can still be saved, and the segments that never ran are not charged:

- The balance runs out. The run stops at the segment where the credits ended, and the top-up window opens. The segments that never ran drop out of the list, uncharged.

- Too many revisions in one hour. AI edit has an hourly ceiling per account. Single-segment edits and whole-episode runs count against the same ceiling; the run stops where it is, and you can pick it up again later.

Stopping the run yourself, discarding results, or leaving them unsaved does not refund revisions that were already charged — the charge happens when a segment is revised, not when it is saved.

## When a row does not come back changed

No change needed means the segment was read and left as it was. The note under the row usually says why: either there was nothing worth changing, or the request is aimed at something the text does not hold (the table above). Nothing is charged.

Couldn’t revise means the revision did not come home: a hiccup on the way, a reply that could not be used, or a request that would push this segment past what it can hold. The reason sits under the row, and nothing is charged. To retry, save or discard the results you have, then run the request again over just those segments.

## Writing a request that lands

- Say what to change, where, and to what: “in shot 2, make her line colder” lands; “improve the dialogue” wanders.

- One kind of change per run. Asking for a new shot and turning the music off in the same sentence usually gets half the job done.

- When a row comes back with a note, read the note — it is usually telling you the request belongs somewhere else.

- If a revision fails because the segment would grow too long, ask for less in that one segment and run the rest as they are.

## Whole episode, or one segment?

Same engine, same price per segment either way. The difference is how many segments ride on one request, and how you review them.

| You want | Use |

| One segment is off; the rest are fine | AI edit on that segment: preview the change, then Revise again, Edit by hand, or save |

| The same note applies from the first segment to the last | The whole-episode run: tick the segments, review the list, save in one pass |

| To test a wording before rolling it out | Run it on one segment first; when it lands, run the same request across the episode |

| A blank or hand-written segment needs writing | AI write on that segment — writing from scratch is not part of the whole-episode run |

## FAQ

### Does the whole-episode run rewrite my script at once?

No. Each segment is revised on its own text and listed for review; only the rows you save are written into the script.

### Which segments appear in the list?

The ones whose text is a storyboard with recognizable shots. Blank segments and segments written as plain prose are left out; open one of those and use AI write.

### Can I leave some segments alone?

Yes. Every segment it can revise starts ticked; untick any of them, or press Clear and pick. Segments you leave out are not revised and not charged.

### Am I charged for segments that need no change?

No. SceneMixer charges only for segments that come back revised, and a failed revision costs nothing. The number on the button is the maximum for the segments you picked, so a run usually costs less.

### What happens if I refresh or close the page mid-run?

SceneMixer keeps results that were already revised on this device for a day. Reopen the AI edit panel on any segment of the episode and a line offers to open them. Segments whose text changed in the meantime are dropped from the list.

### Can I stop a run?

Yes, press Stop. The segments already revised stay in the list and can be saved; the ones that had not started are dropped and are not charged.

### Can it add or remove shots, or change a segment’s seconds?

Yes, within what a segment can hold. When a request changes the timing, the row shows the length before and after, and whether regenerating that segment’s video costs more.

### Does saving regenerate the videos?

No. Saving changes the storyboard text only. Regenerate a segment when you want a clip made from the new text.

### Can I undo a saved revision?

Yes, one segment at a time: open that segment’s History and restore the version from before the change.

### Does it work on a phone?

Yes. The same dialog opens as a sheet from the AI edit button on the Script tab.

Updated November 3, 2026


---

# Faceless YouTube Channel with AI: Run It as a Drama Series

URL: https://scenemixer.com/guides/run-a-faceless-youtube-channel-with-an-ai-drama-series/
A faceless YouTube channel with AI that is not template slop: one recurring cast, episodes and Shorts from the same project, a weekly rhythm, and the labels to know.

# How to Run a Faceless YouTube Channel with an AI Drama Series

Quick answer: Most faceless AI channels post unconnected clips, and that is exactly what platforms now push down as mass-produced content. A channel built as a series — the same characters, a continuing story, a new episode on a fixed day — gives viewers a reason to subscribe and is plainly original work. In SceneMixer one project is one show: the cast is drawn once and stays the same in every episode, each episode is storyboarded, rendered and joined into one video, and a 9:16 cut of the best moment becomes the Short that brings people to the full episode. Nobody has to appear on camera, and the voices are the characters’ own.

Brief for AI Create: A weekly YouTube drama about a night-shift nurse in a city hospital who can hear patients’ last thoughts; each episode one patient, one secret, and a thread that runs across the season.

## Why a series, not a pile of clips

- Subscribers follow characters. A viewer who liked episode 3 comes back for episode 4; nobody comes back for “another AI video”.
- Platforms reward original work. Template-shaped, mass-produced uploads are what monetisation policies pull money from; a written series with its own cast is the opposite. Can you monetize AI videos on YouTube and TikTok?
- Each episode is cheaper than the last. The cast, the locations and the voices are made once and reused.

## Pick a channel concept that can run for 50 episodes

| Format | Why it keeps going | Guide |
| One case per episode (mystery, hospital, lawyer) | Endless cases, the same detectives | Murder mystery |
| History told by a witness | Every event is an episode | History drama |
| Bible or myth retellings | A library of known stories | Bible stories |
| Bedtime stories for kids | One story a night, the same friends | Bedtime stories |
| Serialised romance or revenge | Cliffhangers drive binge-watching | Revenge drama |
| Animated comedy | Recurring characters, one situation each | Sitcom |

## Long episodes and Shorts from the same story

SceneMixer sets the aspect ratio per project in step 1, so make the main show in 16:9 for the channel and a 9:16 Portrait project for Shorts from the same script and cast. Characters, scenes and voices can be reused across projects, so the Short looks like the show. Vertical or landscape · Reuse a character across projects.

- Episode: three to eight minutes, one story, a hook in the first ten seconds, a cliffhanger at the end.
- Short: 30–60 seconds, the most dramatic beat of the episode, ending on “full episode on the channel”.
Hooks and cliffhangers · Longer videos.

## Voices instead of a face

A faceless channel still needs a voice people recognise. Give every character a voice description on its card, and if the channel has a host or narrator, add it as an Off-screen character with its own voice so every episode opens the same way. Lines are performed in the video in the project’s dialogue language, one of 15 on SceneMixer, so the same show can run in Spanish or Portuguese as a second channel. Character voices · A second language version.

## A weekly rhythm you can keep

- Write a season first: eight to twelve one-line episode summaries, so the story has somewhere to go.
- Make episodes in the same project so the cast and sets carry over. Make the next episode.
- Draft cheap, finish the keepers: generate on the default tier, regenerate the shots that matter at a higher tier. Draft at 480P, finish at 1080P.
- Fix one shot, not the episode: edit that shot’s text and regenerate only it. Edit one shot.
- Post on a fixed day, episode and Short together.

## Labels and marks

YouTube asks you to declare realistic synthetic content when you upload, and some laws now expect AI content to be marked. Videos from this site also carry a corner mark unless a paid plan switches it off. What each of those means, and how they differ, is in do AI-generated videos need a label or watermark? Real people and known characters are refused by the content checks, which also keeps the channel clear of likeness complaints.

## Make it in SceneMixer, start to finish

- Script: episode 1 on Upload Script, or a brief in AI Create with the show’s premise and the season’s first arc.
- Step 1: Type TV Series; 16:9 for the channel; the visual style that suits the show; the dialogue language. Read the script.
- Step 2: make each recurring character distinctive; add the narrator or host. The first 15 setup images on an account are free.
- Storyboard: episode 1 first (free on your first project).
- Video: generate on the default tier at 10 credits a second; composite; download and upload.
Post to YouTube, TikTok and Reels · Can you make money with AI videos? · for a vertical sample, watch The Violin on Mercer St.

## FAQ

### How do I start a faceless YouTube channel with AI?

Pick a show concept that can run for many episodes, write a season outline, make episode 1 in SceneMixer with a recurring cast, and post episodes and Shorts on a fixed day.

### Can AI drama channels be monetised?

Platforms judge what an upload is, not what made it: original series do better than mass-produced clips. See the answers page on monetising AI videos for the current rules.

### Will the characters look the same in every episode?

Yes. Each character is drawn once as a reference sheet and every shot is matched to it, episode after episode.

### Do I need to show my face or use my voice?

No. The characters speak their own lines, and a narrator or host can be an off-screen character with its own voice.

### Can I make Shorts from the same show?

Yes. Make a 9:16 project from the same script and reuse the cast, scenes and voices.

### How long should an episode be?

Three to eight minutes for the channel, 30 to 60 seconds for the Short.

### Do I have to label AI videos on YouTube?

YouTube asks for a declaration on realistic synthetic content; the answers page on labels explains what applies.

### Can I run the same show in another language?

Yes. Make a second language version from the finished project and post it on a second channel.

Updated November 3, 2026


---

# Run a Kids Cartoon YouTube Channel with AI: Characters and Series

URL: https://scenemixer.com/guides/run-a-kids-cartoon-youtube-channel-with-ai/
Run a kids cartoon channel with AI: Family 3D Animation, 3D Toy Cartoon and Watercolor Storybook styles, a recurring cast, gentle stories, and an original world.

# How to run a kids cartoon YouTube channel with AI

Quick answer: A kids cartoon channel is a small cast children recognise, a gentle problem each episode and a regular upload day. In SceneMixer, pick a child-friendly look in Step 1 (Family 3D Animation, 3D Toy Cartoon, Watercolor Storybook or Claymation), draw the cast once (animals, kids, a robot) so they look the same every episode, and write short episodes with clear voices. Keep the world original: known cartoon characters are refused.

Brief for AI Create: A kids cartoon series: a small hedgehog who runs the forest's lost-and-found returns one lost thing each episode and learns who it belongs to; gentle, funny, no villains; 10 episodes of about two minutes.

## Pick a kid-friendly look

| Look | Visual Style | Suits |
| Feature-film family 3D | Family 3D Animation | Adventures, family comedy |
| Toy-like 3D | 3D Toy Cartoon | Preschool, talking toys and animals |
| Picture book | Watercolor Storybook | Bedtime stories, fables |
| Handmade | Claymation | Gentle comedy, shorts |
3D cartoon films · Claymation shorts · Bedtime story series.

## A cast children recognise

- Animals and toys are characters. Say what they are and whether they have hands and clothes. Non-human characters.
- One clear mark each: a red scarf, a missing button. Same face.
- Voices from SceneMixer’s preset library, the same every episode. Voices.
- Short, clear lines. Children's episodes work with simple sentences, one idea per shot. Dialogue.

## Channel basics

- Format: 16:9 Landscape for full episodes; a 9:16 cut for Shorts.
- Series, not clips. Platforms push down mass-produced unconnected clips; a continuing series with a cast is original work. Faceless channels.
- Labels and rules: follow the platform's rules for children's content and AI disclosure. Do AI videos need a label? · Can you monetise AI videos on YouTube and TikTok?.
- Subtitles help early readers; they can be added to episodes on accounts with a purchase. Subtitles.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday; Visual Style Family 3D Animation, 3D Toy Cartoon; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: a small cast of animals or children, each with one mark and a preset voice; the forest or the town as recurring locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | The lost thing | Who it belongs to |
| 1 | A red mitten. | A shy rabbit. |
| 2 | A key on a string. | The old owl who forgot his door. |
| 3 | A drawing. | A fox cub who wants to be friends. |
| 4 | A tiny boot. | A mouse on the way to a wedding. |
| 5 | The hedgehog's own hat. | Everyone helps him look. |

## FAQ

### How do I make a kids cartoon channel with AI?

Pick Family 3D Animation, 3D Toy Cartoon, Watercolor Storybook or Claymation in SceneMixer's Step 1, draw a small cast once, and publish a continuing series on a regular day.

### Will the characters look the same every episode?

Yes. Each character is drawn once in SceneMixer’s Step 2 and that sheet is used in every shot.

### Can animals talk?

Yes. Animals are characters with voices from the preset library.

### Can I use famous cartoon characters?

No. SceneMixer refuses known characters at the IP check; create your own.

### Which style is best for preschool?

3D Toy Cartoon or Watercolor Storybook, with short, clear lines.

### Do I need to label AI content?

Follow each platform's rules; see our answer on AI labels.

### Can I add subtitles?

Yes, on SceneMixer accounts that have made a purchase.

### How long should episodes be?

Two to five minutes for full episodes; under a minute for Shorts.

Updated November 3, 2026


---

# Run an AI Horror Story Channel: Format, Schedule and Costs

URL: https://scenemixer.com/guides/run-an-ai-horror-story-channel/
Run an AI horror story channel: a recurring host voice, a four-minute episode format, 12 story seeds, a monthly schedule, cost math, covers and what to keep original.

# How to run an AI horror story channel

Quick answer: A horror story channel is narrated stories with moving pictures, posted on a schedule. In SceneMixer one project is one show: a recurring host (an off-screen narrator with the same voice every episode), a house or town you draw once, and a new story of about four minutes, which is exactly the Short Drama limit of 240 seconds. Write your own stories, or adapt ones you have permission to use. The content rules block gore, so the fear has to come from a rule, a sound and a cut. Plan a month ahead, make episodes in batches and fix single shots rather than whole episodes.

Brief for AI Create: A four-minute horror story for a channel, told by a host: a night-shift lift technician is called to a tower whose lift stops, every night, at a floor that is not on the panel; one building, one lift, three escalations, the last shot is the new button.

## The four-minute episode

A repeating shape is what turns stories into a channel. This one fits one Short Drama episode:

| Part | Seconds | What happens |
| Cold open | 0 to 15 | The strangest image of the story, no explanation |
| Host opening | 15 to 25 | The same line every week, in the host's voice |
| Setup | 25 to 90 | The place, the person, the rule |
| First escalation | 90 to 140 | The rule is tested; a sound before a sight |
| Second escalation | 140 to 190 | The rule is broken; the almost-reveal |
| The shock | 190 to 220 | One real shock, then a cut to black |
| Host sign-off | 220 to 240 | One quiet line that leaves the question open |
What makes a short scary, without gore · Short drama, series or film.

## Stories you can use

Use your own stories, or ones whose author has given you permission, and change names and details so no real person can be identified. Stories you find online usually have an author: ask first and keep the reply. Never depict the real people in a story. Turning an online story into a drama · Content rules.

A rule is the easiest engine for a weekly horror story. Twelve seeds, each a place with one rule that gets broken:

| The rule | The story |
| The lift stops at a floor that is not on the panel | A technician is sent to fix it, and the floor starts to answer |
| The tenant in 4B pays his rent in jars of salt | The landlord counts them, and one night there is one jar too many |
| The night bus stops at a stop not on the map | The driver always lets one passenger off there |
| The baby monitor picks up a lullaby | The house next door has been empty for a year |
| The lighthouse log has a new entry every night | It is in handwriting the keeper recognises |
| The school photograph has one child too many | Nobody can say which one |
| The car park counts its levels down | Below level 1, the numbers keep going |
| The hotel skips rooms 13 and 14 | A guest asks for 14, and the desk clerk goes pale |
| A contact called Home is saved in a phone | It is not the phone owner's home |
| A village sets one extra place at the table on harvest night | A visitor asks why |
| A voicemail arrives from the listener | It is dated tomorrow |
| The radio host reads the weather | It describes the studio, from the inside |

## Make the host a character

The host is what makes a channel feel like a channel. Add the host in SceneMixer as an Off-screen character: a voice and no drawing. Play the library and pick one voice, then keep it for every episode, and use the same opening and closing lines. How character voices work · Voice-only characters.

Two ways to organise the channel: one project per season, with the host and any recurring house made once, or one project per story with the host and house brought across using Pick existing. Reuse a character, scene or voice across projects · Add the next episode.

## A month on one page

| Week | Do | Result |
| 1 | Write four one-page stories; draw the host's voice and the recurring place | Four scripts and the shared cast |
| 2 | Read the storyboards for episodes 1 and 2 (the first is free on your first drama); render them | Two episodes |
| 3 | Render episodes 3 and 4; fix single shots instead of redoing scenes | Four episodes |
| 4 | Make covers; schedule one episode a week | A month of posts |
A four-minute episode is 240 seconds of video at 10 credits a second on the default tier, plus 1 credit a second to join; four a month is 960 seconds. A segment usually takes ten to twenty minutes to come back and you can close the page while it runs. Drafting at the lowest resolution and finishing the keepers at a higher one keeps the bill down. Draft at 480p, finish at 1080p · Spending fewer credits.

## Covers, titles and labels

Open the Cover art tab in Step 2, set the genre mood to Horror, type a title of 12 characters or fewer and press Generate 3 options. The title is painted into three posters, and a landscape 16:9 version comes with the one you pick, which is the shape of a video thumbnail. Your first project on SceneMixer gets its first set of three free. Make a promo cover poster.

Say plainly in the description that the video is AI-made and who it is for, and follow each platform's rules on AI disclosure and children's content. Do AI videos need a label? · Can you monetise AI videos?.

## Keep the channel from looking mass-produced

- A series, not a pile of clips. A host, a recurring place and a fixed day. A faceless YouTube channel.
- A new rule every week. The shock changes; the shape does not.
- No gore. It is blocked, and the best endings do not need it.
- No real people. Villains and victims are invented.
- Fix the dark. Name a light source in every dark shot, or the frame comes out muddy. A haunted house drama · AI horror movie generator · Posting to TikTok, YouTube Shorts and Reels.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Corporate Cold, Gothic Dark; Costume Style Professional Uniforms, Soft Everyday; Visual Style Digital HD, Black & White Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Give the host a voice and no drawing; draw the technician, the tower lobby, the lift and the corridor of the floor that is not on the panel. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode length | Seconds | Segments of 15 s | Type to choose |
| 3 minutes | 180 | 12 | Short Drama |
| 4 minutes | 240 | 16 | Short Drama (its limit) |
| 5 minutes | 300 | 20 | TV Series (its limit) |
| 8 minutes | 480 | 32 | Film |

## FAQ

### Can I make horror story videos with AI?

Yes. Write the story, draw the cast and a place once, add an off-screen narrator and render the segments. Gore is blocked on SceneMixer, so the fear comes from a rule, a sound and a cut.

### How long should a horror story video be?

Three to five minutes is a good channel length. On SceneMixer, four minutes is the limit of a Short Drama episode; five minutes is the limit of a TV Series episode.

### Can I use creepypasta or stories from the internet?

Use your own, or ones whose author has given you permission. Stories online usually have an author, so ask and keep the reply.

### Can the same narrator voice every episode?

Yes. The host on SceneMixer is an off-screen character with a voice you pick once and reuse.

### Can I use my own voice for the host?

Members can upload a voice recording to SceneMixer for a character. Only upload a voice you have the right to use.

### How much does one episode cost?

A four-minute episode is 240 seconds of video at 10 credits a second on the default tier plus 1 credit a second to join.

### Can I make a cover for the channel?

Yes. SceneMixer's Cover art tab paints a title into three poster options and gives you a landscape 16:9 version of the one you pick.

### Does it need an AI label?

Follow the platform's rules on AI disclosure. Videos on a free SceneMixer account carry a small AI label; a paid membership turns it off by default.

Updated November 3, 2026


---

# Consistent Character AI Video: Same Face, Every Shot | SceneMixer

URL: https://scenemixer.com/guides/same-face-in-every-shot/
How SceneMixer keeps a character's face, build and outfit identical across every shot and episode, what you can change on purpose, and what to do if a shot drifts.

# The Same Face in Every Shot

Quick answer: Each character in a SceneMixer project is one entry with one approved image. Every shot of every episode that shows her is produced from that entry, so the face in a close-up is the face in a wide shot is the face in episode nine. Outfit changes are explicit states you choose, not accidents. If a single shot drifts, you regenerate that shot — the character herself does not move.

Open a finished episode and click through the segments. Watch the lead across a close-up, a two-shot and a wide shot.

## Where the face comes from

When your script is read, SceneMixer writes a description for every character: face shape, hair, build, age, what they wear. In step 2 each description becomes a character image. That image is the character. You look at it once, and if it is not who you had in mind you change the description and generate again, or upload your own photo instead.

Leads and supporting cast work the same way. So does a character who only appears in episode 6.

## What “consistent” covers

Face, hair, skin, build, age. The clothes in the entry. Glasses, a scar, a necklace, if the description has them. And the voice: each character has one voice, chosen once, used for every line she speaks. How voices are chosen.

It holds at every distance and every angle. A close-up, an over-the-shoulder shot, a full shot from the other side of the room — the same person.

## Across shots, segments and episodes

This is not one clip that happens to hold together. A segment is several shots; an episode is several segments; a series is many episodes. The same entry feeds all of them. Generate episode 3 today and episode 4 next month, and the lead is unchanged.

Because the source is the entry and not the previous clip, a bad shot does not poison the next one. Regenerate it, or leave it and move on; the shot after it is made from the entry either way.

## Changing a character on purpose

A different look for the whole series: open the character in step 2, edit the description, generate a new image. Every shot from then on uses it. Shots you already generated keep the old look until you regenerate them.

Your own face or artwork: upload an image on the character. It replaces the generated one.

A costume change, an injury, a transformation: add a state to the character and assign it to the segments where it applies. States, step by step.

One episode only: in the editor, Edit (this episode) on a character swaps her image for that episode and leaves the rest of the series alone.

## When one shot drifts

It still happens: a shot comes back with someone who is not quite her. Almost always the storyboard for that shot did not say who was in it. A line that reads “a woman turns to the window” gives the shot nothing to hold on to; a line that reads @Mara turns to the window does. Open the shot, make sure the character is named with the @ chip, and regenerate that one segment. SceneMixer keeps every version, so you lose nothing by trying. Naming who is in frame.

| What you see | Usual cause | Fix |
| A stranger’s face in one shot | The shot text does not name the character | Add the @ chip for her, regenerate that segment |
| Right face, wrong outfit | The story changed her clothes but the entry did not | Add a state for the new outfit and assign it to those segments |
| Hair or age changed after you edited the description | Old shots were generated before the edit | Regenerate the shots you want updated |
| Same face on two characters | Two entries with near-identical descriptions | Give each a distinct anchor in the description and regenerate their images |

## FAQ

### Do I need to upload a photo to keep a character consistent?

No. The generated character image is enough. Uploading is an option when you want a specific face or artwork.

### Will the character look the same in a later episode?

Yes. Every episode is generated from the same entry. Nothing you do in episode 3 changes the entry unless you edit it.

### What if I change the description after some shots are already made?

New shots follow the new description. Old shots keep what they had until you regenerate them.

### Can two characters share a face by accident?

If their descriptions are nearly the same, their images can end up similar. Give each a clear distinguishing feature in the description and regenerate.

### Does consistency cost extra?

No. It is how the project works. SceneMixer charges for generating shots, not for keeping them consistent.

Updated November 3, 2026


---

# Prop Consistency in AI Video: Same Prop, Every Shot | SceneMixer

URL: https://scenemixer.com/guides/same-prop-in-every-hand/
Each prop is one entry with one image, and every shot that names it gets that object. How to name props in a shot, and what to do when one changes.

# The Same Prop in Every Hand

Quick answer: Props on SceneMixer are treated like cast. When your script is read, the objects that matter are listed as props, each gets its own image in step 2, and any shot that names a prop is generated with that object — not a look-alike. The letter she reads in episode 2 is the letter he burns in episode 5. Name the prop in the shot with @ and it follows; leave it unnamed and the shot has to guess.

Newer storyboards. In storyboards written or regenerated since late October 2026 there is no Props line in the segment header: a prop appears when a shot names it with @ (Props tab), and stays in the hand it is in until the text says otherwise. What the system keeps track of between segments.

Open a finished episode, switch the left panel to Items, and look for the same object across two segments.

## Props are entries, like characters

Reading the script produces three lists: characters, scenes, props. The props list holds the objects the story depends on — a weapon, a document, a keepsake, a device. Each entry has a description and, after step 2, an image. That image is the prop.

Everyday objects that nobody would notice — a random cup, a chair — do not become props. They do not need to be the same cup twice.

## Telling a shot which prop you mean

In the storyboard, a prop appears as an @ chip, the same way a character does. Type @, open the Props tab, pick the prop, and the chip drops into the text. A shot that says @Sealed letter is handed over gets that letter: same paper, same seal, same size in the hand.

A shot that says “a letter” with no chip is asking for any letter. It may come back close. It may not. The chip is what makes it the same one. How to write props into a shot.

## What holds and what changes

What holds: the object’s shape, colour, material and markings, across shots, segments and episodes. A blade is the same blade drawn or sheathed — but drawn and sheathed are two different pictures.

That is what states are for. A prop can have states: the letter before and after it is burned, the sword in its scabbard and out of it, the phone whole and cracked. Each state is its own image, and you choose which segments use it. States, step by step.

## Adding a prop the script did not list

If the reading missed an object you care about, add it. In step 2, open the Items tab and add a new entry with a description; generate its image. From the editor, Add assets on the left panel brings an existing prop into this episode. Then name it in the shots with @.

## When a prop comes out wrong

Check three things. Is it in the props list at all? Did the shot name it with the chip? Is the shot asking for too many objects at once? A shot crowded with five named props tends to lose one. Split the action across two shots, name the prop in each, and regenerate the segment. SceneMixer keeps every version.

| Situation | What to do |
| The same object must appear in several episodes | Make sure it is a prop entry; name it with @ in each shot |
| The object changes (breaks, opens, burns) | Add a state for the changed version and assign it to those segments |
| The object was not picked up from the script | Add it under Items in step 2, generate its image, then name it |
| A shot shows a different-looking object | Add the @ chip for the prop in that shot and regenerate the segment |

## FAQ

### Does every object in the story become a prop?

No. Only objects the story depends on. Incidental things stay incidental.

### How do I make a shot use a specific prop?

Name it in the storyboard with the @ chip from the Props tab. Unnamed objects are not guaranteed to match.

### Can a prop change during the series?

Yes, through states: one image per version of the object, assigned to the segments where that version applies.

### Can I upload my own image of a prop?

Yes. Open the prop in SceneMixer's step 2 and upload an image; it replaces the generated one.

### Will a prop look the same in a character's hand and lying on a table?

Yes. It is the same object at any size and angle; the shot text says where it is and who holds it.

Updated November 3, 2026


---

# Consistent AI Video Backgrounds: Same Room Every Time | SceneMixer

URL: https://scenemixer.com/guides/same-room-in-every-episode/
A location is generated once and every scene set there uses it, from episode 1 to episode 20. How locations are made and changed, and how time of day and weather work.

# The Same Room in Every Episode

Quick answer: A location on SceneMixer is an entry, like a character: the script reading lists the places the story happens in, step 2 gives each one an image, and every scene set in that place — in any episode — is generated from that image. The kitchen has the same window, the same table, the same tiles every time you come back to it. Time of day and weather are written per segment and change on top of the room; the room itself does not drift.

Open a finished episode, switch the left panel to Scenes, then find two segments set in the same place.

## A location is made once

When the script is read, the places it happens in become the Scenes list: a kitchen, a courtyard, the front seat of a car. Each gets a description — what is in the room, what it is made of, where the light comes from — and in step 2 an image. Scene images are always widescreen, whatever aspect ratio your videos use, because they describe a place rather than a frame.

That image is the place. Every segment whose header says the scene is this kitchen is generated from it, in this episode and every later one.

## What stays and what changes on top

Stays: the architecture, the furniture, the materials, the colours, where the window is. Changes per segment: time of day, weather, and what the characters are doing in it. You write those in the segment header — Time, Weather — and they are applied on top of the same room. Night in the kitchen is the same kitchen, dark. The segment header, field by field.

A room that is physically different later in the story — wrecked, redecorated, flooded — is a state of the scene: its own image, assigned to the segments where it applies. States, step by step.

## One place, many angles

You are not limited to the angle of the scene image. Shots in that room can look toward the window, away from it, down the table, from the doorway. They are all the same room because SceneMixer holds the layout for you. How positions hold across cuts.

## Changing a location

Open the scene in step 2, edit its description, regenerate. Or upload your own photo of a place. From then on, new shots use the new image; shots already generated keep the old one until you regenerate them.

To change it for one episode only, use Edit (this episode) on the scene in the editor. The rest of the series keeps the original.

## Adding a place the script did not name

Add it in step 2 under Scenes, with a description, and generate the image. Then set it as the location in the segment header of the shots that happen there. From the editor, Add assets brings an existing scene into the current episode.

| You want | Where to do it |
| The same room in every episode | Nothing — that is the default |
| The same room at night, or in rain | Time and Weather in the segment header |
| The room after it was wrecked | A state of the scene, assigned to those segments |
| A different look for the whole series | Edit the scene in step 2 and regenerate |
| A different look in one episode | Edit (this episode) in the editor |

## FAQ

### Why are scene images widescreen when my videos are vertical?

A scene image describes a place, not a frame. SceneMixer always makes it widescreen so it captures the whole room; your video shots are framed from it in whatever ratio your project uses.

### Will the same location look the same in a later episode?

Yes. Every episode is generated from the same scene entry.

### How do I get night, rain or dusk?

Write them in the segment header's Time and Weather fields. The room stays the same underneath.

### Can I use a photo of a real place?

Yes. Upload it on the scene entry in step 2; it replaces the generated image.

### Can a location change during the story?

Yes, through states: one image per version of the place, assigned to the segments where it applies.

Updated November 3, 2026


---

# Say What Is There, Not What Is Not: Why 'No Crowd' Draws a Crowd

URL: https://scenemixer.com/guides/say-what-is-there-not-what-is-not/
Generation models treat every noun as a candidate for the picture, even after 'no' or 'without'. How to turn each negative into a positive description that holds.

# Say What Is in the Picture, Not What Is Not: Why ‘No Crowd’ Gives You a Crowd

Quick answer: To a generation model, a sentence is a list of things to consider drawing. ‘No crowd in the background’ puts crowd on that list; ‘without rain’ puts rain on it. The word no carries much less weight than the noun, so negatives often produce the very thing they name. The fix is always the same: say what is there. ‘An empty street, wet asphalt reflecting the shop signs’ instead of ‘no people, no rain’. This site’s own fixed wording in the clip instructions is written this way for the same reason; your storyboard and sheet text should be too.

Find a segment whose text says “no” or “without” something and generate it twice: the named thing tends to appear. Rewrite as a positive and compare.

## Why a negative sentence backfires

The model does not reason about your sentence; it weighs the words. Nouns that name visible things weigh a lot; function words weigh little. So ‘a quiet square with no vendors’ is, to the model, mostly square, vendors. Lists of forbidden things are worse: ‘no text, no logos, no extra people, no blur’ is a list of four things to consider. The general rule about causes and results: Describe the Cause, Not the Result.

## Turning negatives into positives

Ask what you would see if the negative were true, and write that. No people → an empty square, chairs stacked against the café wall. No rain → dry pavement, hard afternoon sun. Not smiling → mouth closed, jaw set. No text on screen → bare walls, blank shop fronts. Not looking at camera → eyes on the door at frame left. Each of these names only things that should be there.

## Where this matters most

Background people: the storyboard has a dedicated line for whether a scene has people in the background; use it rather than writing ‘nobody else’. Background People in a Scene. Character sheets: ‘no jewellery’ invites jewellery; describe the bare neck and hands. Scene plates: ‘no furniture’ invites furniture; describe the empty floor and what is on the walls. Expression: ‘not angry’ is meaningless; write the expression you want.

## The one negative that works

‘No dialogue’ on a shot is understood by the video models as a control word, and the site keeps it. Everything else about what the audio contains is written positively (ambient sound only, footsteps on gravel). Do not extend the exception to pictures.

## Checking your text

Search a segment for no, not, without, avoid, don’t. Each hit is a noun you are inviting. Rewrite; keep the sentence about what the lens sees.

You do not have to search alone: after you save a storyboard edit, SceneMixer shows an amber note above the text when a sentence reads like “do not show…” about a person, warning that the person named tends to end up in the frame. It is only a note; it does not change your text or cost anything. Rewrite the sentence as who is in the frame and save again. The amber notes after you save.

| Negative | What the model hears | Positive |
| No crowd in the background | crowd, background | Empty stands, rows of bare seats |
| Without rain | rain | Dry pavement, hard sun |
| Not looking at the camera | camera, looking | Eyes on the door at frame left |
| No text or logos | text, logos | Bare brick walls, plain glass fronts |
| Not smiling | smiling | Mouth closed, jaw set |

## FAQ

### Does this apply to the script itself, before parsing?

Less; the parse rewrites the script into descriptions. It matters in descriptions and storyboards, which go to the models as written.

### What about 'No dialogue'?

Kept on purpose; the video models treat it as a control word.

### Can I list things to avoid in a sheet description?

Better not; describe what is there instead. Lists of avoided things are lists of candidates.

### The clip added a person I never mentioned.

Check for a background-people line, a crowd noun in the ambient sound, or a negative naming people.

### Is there a 'negative prompt' field?

No. Some models ignore such fields anyway; the positive description is the reliable tool.

Updated November 3, 2026


---

# Scene Styles and Costume Styles, One by One: Every Preset Explained

URL: https://scenemixer.com/guides/scene-and-costume-style-presets-one-by-one/
Every Scene Style and Costume Style in Step 1, grouped and described in a line each: modern cities, schools, palaces, world regions, xianxia, wuxia, fantasy and sci-fi.

# Scene styles and costume styles, one by one

Quick answer: Step 1 has two lists that decide what the world looks like and what people wear. Scene Style is the art direction of the sets: 74 presets in 7 groups. Costume Style is the wardrobe of the whole film: 90 presets in 9 groups. Pick one of each, or let AI choose from your script. Neither one decides which country the story is in — your script and Target Region do that — they set materials, period and mood. This page lists every preset with one line on what it gives you. Presets marked new were added on October 25, 2026.

Two settings that work differently: how scene and costume styles are applied when the script is read and Target Region. The visual style (live action, anime, 3D) has its own list: visual style presets, one by one.

## How to pick them in Step 1

- Open Scene Style and Costume Style in Step 1 and click a card. Each card has a picture; the one you select plays a short loop of that look.
- If you pick a 3D or 2D Visual Style, the scene and costume cards switch to 3D or 2D versions of the same presets, so you see the look you will get.
- The two lists are independent: a period costume in a modern scene, or the other way round, is allowed. Choose the pair that fits the story.
- The visual, scene and costume style all need a choice before the script can be read. In a drama project, if some are still unpicked when you press Next, a prompt asks Let AI choose the remaining styles? Choose Let AI choose and it picks them from your script, marked AI pick with a short reason. You can change every one of them.
- Pick before the script is read: scene and costume styles are written into the scene and character descriptions during the read, so they are fixed once Step 1 is done. See how to change the look afterwards.

SceneMixer members can also write their own scene or costume style in the same dropdown with Custom Style; use the presets below when one is close.

## Scene styles (74)

A scene style is the world the whole film is shot in: building materials, period, how crowded the rooms are, the kind of light. The groups below follow the order in Step 1.

### Modern (19)

- Urban Opulence — Rich people’s interiors: solid stone, brass and walnut, double-height rooms and floor-to-ceiling glass.
- Street Life — Busy old-town streets in China: crowded, colourful and orderly; arcade shopfronts in the south, hutongs and courtyards in the north.
- Corporate Cold — Glass, steel and bare concrete with straight lines and symmetry: the bigger the room, the less in it.
- Neon Nights — An Asian metropolis after rain: wet streets mirroring signboards and neon, towers pressing the street into a canyon.
- Soft Romance — Cream and light-wood rooms with big windows and sheer curtains; every space looks freshly sunned.
- Campus Youth — Chinese schools and universities: terrazzo corridors, blue-and-white railings, camphor shade and plastic running tracks.
- Forensic Cold White — Police and court spaces: white tile, stainless steel, one-way glass and rows of tube lights under grid ceilings.
- Hospital Ward — A general hospital: white walls with a pale green dado, shiny floors, rails and equipment in neat rows.
- Island Resort — A tropical island resort: white sand, shallow turquoise sea, boardwalks and thatched roofs, with wind and light passing through the rooms.
- Military Base — An army camp and drill ground: right angles, white lines, everything aligned, in olive and khaki.
- American High School — A US suburban public high school: red-brick building, rows of metal lockers, a polished gym floor and a grass field.
- Japanese High School new — A Japanese school year: white concrete building, sliding classroom doors, wooden shoe lockers at the entrance, cherry blossom in spring and red leaves in autumn.
- Korean High School new — A Korean school and Seoul streets: multi-storey building, artificial-turf yard, lockers along the classroom wall, cram-school lightboxes and convenience stores at night.
- Elite Boarding School new — Old private schools and colleges in England and New England: stone quadrangles, chapel windows, long dining tables and ivy on the walls.
- Suburban Mansion — A wealthy American suburb: white colonnades, double-height halls, big windows, a trimmed lawn and a pool.
- New York Metropolis — Manhattan and Brooklyn: glass towers beside brownstones, iron fire escapes, wide pavements, the sky squeezed to a slit.
- Parisian Streets — A cream-limestone city: Haussmann balconies, zinc roofs, tree-lined boulevards, café terraces and stone bridges over the Seine.
- Nordic Minimal — White walls, pale oak, big windows and lots of empty space; few pieces of furniture and daylight as the star.
- Gulf Metropolis — A Dubai- or Doha-like city: supertall glass towers, wide waterfront avenues, marble-and-brass interiors and strong sun.

### Fantasy (14)

- Epic Scale — Ancient dynasties and wars at a vast scale, used for the East (Qin–Han, Three Kingdoms, Sui–Tang), Rome, Persia, South Asia and the Middle East; the architecture follows your script’s culture.
- Gothic Dark — A dark Gothic world from medieval Europe to the Victorian era: pointed arches, flying buttresses, stained glass and candles, tall and narrow, cold black and dark green.
- European Palace — European courts of the 17th–19th centuries: gilded Baroque stucco, marble parquet, crystal chandeliers, halls of mirrors and long galleries.
- Cyber Metropolis — A near-future megacity in layers: clean glass-and-metal towers above, crowded streets with holograms and exposed pipes below, lit by neon.
- Wasteland — The world after collapse: vast ruins, fallen overpasses and sand-covered city skeletons, with tidy survivor camps among them.
- Pastoral Healing — An idealised slow country life: log cabins, flower fields, streams and handmade things, small and warm.
- Steampunk — Victorian industrial fantasy: exposed brass pipes and gears, riveted iron plate, leather upholstery, gas lamps and steam hanging in the air.
- Starship — Spaceships and stations: curved bulkheads, flush panels, matte metal and composites, repeating hull frames and stars through the portholes.
- Elven Forest — A forest kingdom grown into giant trees: timber platforms and rope bridges in the canopy, vines and glowing fungi, light spots through mist.
- Dungeon — Rough-cut stone walls and arched passages, barred doors, chains and torch brackets, wet flagstones and a dripping ceiling, with torches as the only light.
- Magic Academy — An old school of magic: Gothic stone halls and spiral stairs, tall bookshelves, floating candles, stained glass and astronomical instruments.
- Mars Base — A Mars colony: white modules and domed greenhouses in a red desert under an orange sky, cramped metal corridors with light strips inside.
- Alien World new — An alien planet’s surface and settlements: a purple sky with two moons, glowing fungal forests taller than buildings, crystal pillars, floating rock and fluorescent lakes.
- Undersea Kingdom new — A kingdom of the sea: spiral-shell palaces and mother-of-pearl arches on coral reefs, kelp forests and giant anemones, water shading from pale teal to deep blue.

### World Regions (24)

- Medieval Castle — Medieval Europe: rough-cut stone, oak beams and iron chandeliers; small deep windows, slanting daylight, hearth and torchlight at night.
- Wild West — The American West: red earth, sandstone mesas, plank buildings and saloon fronts under noon sun or an orange evening sky.
- Southeast Asian Town — Humid, bright tropical towns such as Penang, Malacca and Bangkok: shophouses, coloured shutters, palms and bougainvillea, stalls and motorbikes.
- Latin Colonial Town — Colonial towns in Mexico, Colombia and Cuba: saturated stucco walls, iron balconies, cobbled squares and church towers.
- Japanese Machiya — Traditional Japanese townhouses and temples: wooden lattices, paper screens, tatami, verandas and gardens, with light softened by paper.
- Silk Road City — A Central Asian oasis city: mud-brick walls and arched gates, blue-green glazed domes, a crowded bazaar of carpets and brass under deep blue sky.
- Frozen Wilderness — Polar and far-north snowfields in Scandinavia, Siberia and Alaska: snow, ice lakes and conifers, with a cabin and chimney smoke the only warmth.
- Middle Eastern Souq — Old-city lanes in Cairo, Marrakech or Istanbul: narrow alleys under vaults, brass, carpets and lamps on every wall, light from skylights.
- Mediterranean Coast — Small towns in Greece, southern Italy and Spain: white walls, blue doors, stone steps, harbour and sea in strong sunlight.
- Indian Bazaar Streets — Indian city streets in Delhi, Jaipur and Mumbai: colonial facades under bright shopfronts, cloth awnings, garlands and temple spires.
- African Savanna — East African savanna in Kenya and Tanzania: golden grass, acacia, far hills and a huge sky, with sparse, sturdy buildings.
- Eastern European Old Town — Historic old towns like Prague, Kraków and Tallinn: coloured stucco facades and red tile roofs around squares, clock towers, cobbles and trams.
- Tropical Rainforest — The Amazon or Borneo: layer on layer of green, buttress-root trees, vines, streams and waterfall mist, with only spots of light under the canopy.
- Greco-Roman — Ancient Greece and Rome in white marble: colonnades, temple pediments, forums and open-air theatres in strong Mediterranean light.
- Ancient Egypt — Massive sandstone columns covered in reliefs, Nile reeds and palms, pyramid outlines in the desert, hard sun and sharp shadows.
- Joseon Korea new — Palaces and scholar houses of Hanyang: grey tile roofs with gently upturned eaves, beams painted in red, green and blue, paper-screen windows and stone terraces.
- Ottoman Palace new — Istanbul palaces and mansions of the 16th–19th centuries: blue-white and turquoise floral tiles on the walls, layered low seating and gilding.
- Indian Royal Palace new — Royal palaces and forts of northern India: red sandstone and white marble arcades, domed pavilions and mirror mosaics.
- Thai Royal Palace new — Thai royal palaces and temples: stacked red-and-green tiled roofs with upturned ends, gilded eave carvings and golden bell-shaped stupas.
- English Manor new — Regency to Edwardian high society: country houses and London townhouses with Palladian stone fronts, tall windows and white plaster mouldings.
- Viking Fjord new — A Viking-age fjord settlement: steep cliffs around deep blue water, turf-roofed timber longhouses with carved dragon heads on the ridge, wooden docks.
- Age of Sail Harbor new — A 17th–18th-century Caribbean colonial harbour: wooden decks and rigging, colourful stucco colonial houses and a turquoise sea.
- Pre-Columbian Americas new — Maya and Aztec temple cities and Inca mountain towns: stepped pyramids with shrines on top, carved stelae and stone plazas, jungle all round.
- Polynesian Island new — A South Pacific volcanic-island village: palm-thatched open halls, carved totem posts, a turquoise lagoon and black sand, green volcano behind.

### Xianxia (5)

- Celestial Colossus — Heaven and immortal sects at a giant scale: palaces floating above the clouds, white jade steps and gilded eaves.
- Dark Asura — Demon realms in black and gold: matte iron, obsidian, bone totems and dark gilding, with fire and lava as the light.
- Classical Elegance — The taste of scholar families in Song and Ming China: plain silks, raw wood, grey tile and white walls, courtyards and landscapes, sparse and refined.
- Imperial Splendor — Ming and Qing palaces: vermilion pillars, gilded caisson ceilings, yellow glazed roofs and layered hangings in strict symmetry.
- Ancient Farmstead new — Ancient Chinese villages and market towns for farming, famine and family-split stories: rammed-earth walls, grey tile, bamboo fences, terraced fields.

### Wuxia (3)

- Dark Wuxia — The grim side of the jianghu: wet flagstones in rain, black tile and dark wood, inns, alleys and gorges, with dark red only as an accent.
- Ink Jianghu — The poetic jianghu: bamboo groves, ferry landings, mist on the water and far hills, built from bamboo, wood, reed and rope.
- Frontier Vastness — Desert and border country: gobi, wind-carved rock, a lone fort and tents under a low horizon and a huge sky.

### Town & Rural (5)

- American Small Town — A Midwest or New England town: wide streets and low buildings, clapboard houses with porch swings, a main street of two-storey brick.
- Chinese Village — Today’s Chinese countryside: brick-and-tile houses next to new self-built ones, swept yards; plain courtyards and wheat in the north, villages and rice paddies among hills in the south.
- Chinese Folk Occult new — The folk-occult world of the Chinese countryside: old family halls, Taoist temples, mourning halls and ancient tombs, with ancestral tablets, oil lamps and incense smoke.
- Miao Mountain Village new — A stilt-house village in the mountains of southwest China: timber houses stepping up the slope, grey tile roofs, balcony corridors and terraces.
- Tibetan Plateau new — Monasteries, pastures and towns on the plateau: whitewashed walls under maroon and gold roofs, prayer flags, snow mountains, meadows and lakes.

### Period (4)

- Republican Era — Shanghai and the big cities of the 1910s–40s, east meeting west: lane houses, stone bank buildings on the Bund, ballroom staircases, warm brown tones.
- Retro Hong Kong — Hong Kong in the 1980s–90s: neon signs stacked out over the street, tenement buildings, tea restaurants and rooftops in warm green and magenta.
- 80s Nostalgia — Chinese towns in the 1980s: terrazzo floors, green-painted wall dados, enamelware and wooden furniture in warm yellow-brown.
- WWII Front new — The 1930s–40s at war, on the European, Pacific and Chinese fronts: muddy trenches, sandbag bunkers, barbed wire and shell-damaged buildings.

## Costume styles (90)

A costume style is the wardrobe system: cut, fabric, colour and how the clothes change with rank or occasion. The same style covers everyone in the film, so a school uniform is the same uniform for the whole class.

### Contemporary (21)

- Power Tailoring — Tailored business suits and crisp shirts in dark and neutral colours with sharp lines.
- Haute Couture — High-fashion gowns: big areas of solid colour and sharp structured cuts, with the craft hidden in fabric and finish; the more formal, the more red carpet.
- Soft Everyday — Soft everyday clothes between the flat, the café and the walk to school: cream tones, knits and fluffy fabrics in a relaxed fit.
- Professional Uniforms — Job uniforms: cut and colour are the same across one organisation, so the role reads at a glance.
- School Uniform — East Asian school uniforms, from tracksuit-style to western blazers: clean and youthful, with personal style only in the details.
- Japanese School Uniform new — Sailor or blazer uniforms and stand-collar gakuran in navy, black and white; differences between students sit only in the neck ribbon and cardigan.
- Korean School Uniform new — Slim navy or grey blazers, plaid skirts and ties with a name tag on the chest, and a long padded coat over it in winter.
- Prep School Uniform new — Uniforms of Western private and boarding schools: crested blazer, striped tie, plaid skirt or grey trousers, coloured piping on the jacket edge.
- Modern Military — The modern army of an invented country in an international western style: field dress, service dress and dress uniform with shoulder boards, medals and sashes; the insignia are always invented.
- Streetwear — Young street fashion from skate parks, courts and subway exits: oversized cuts, layering, prints and bold sneakers.
- American Roadside — Road and biker culture of mid-to-late 20th-century America: leather jackets, denim and boots, metal buckles and worn leather.
- Sportswear — Team kit for leagues, training grounds and podiums: fitted sports fabric, club badges, numbers and strong team colours.
- Wedding — A western wedding: the bride in white with a long train, the groom in fitted formalwear, guests in soft pale colours.
- Beach Resort — Holiday clothes for the pool, the sand and dinner at the hotel: light, breathable linen and prints that move in the wind, in sunny colours.
- Loungewear — Home clothes: soft knit and cotton in loose shapes and pale colours, a little tidier when guests come.
- Smart Casual — The commute layer between suit and weekend: well-fitted pieces in neutral colours and good fabric.
- Y2K Millennium — Fashion around the year 2000: low waists, cropped tops, metallic and candy colours, shiny and sheer fabric.
- Minimal Luxury — A wardrobe with the decoration stripped out: neutral colours and large plain surfaces, won by fabric quality and proportion.
- Outdoor Gear — Technical clothing for hills and the commute: layered, matte windproof fabric, colour blocking and taped seams.
- Rock & Punk — The rock and punk scene from the 1970s: black leather, plaid and metal studs, with tears and wear that are on purpose.
- Arab Everyday Wear new — Everyday modest dress across the Arab world: long-sleeved ground-length abayas with headscarves, ankle-length thobes with head cloths, in draping fabric and plain colours.

### Western Period (10)

- Plate Armor — Full plate armour of the 15th–16th centuries: polished curved steel with fluting and overlapping plates, a surcoat with arms over it.
- Medieval Nobility — Feudal lords and ladies of the 11th–14th centuries: long robes and overgowns in velvet and brocade, fur trim and heraldic colours.
- Victorian — Britain’s middle and upper classes in the mid-to-late 19th century: corseted long dresses with high collars, three-piece suits and top hats in dark and jewel colours.
- Vintage Hollywood — Studio-era stars of the 1930s–50s: bias-cut satin gowns, set waves and red lips, wide-shoulder suits, with satin, sequins and fur.
- Western Cowboy — The American West of the later 19th century: wide-brim hats, boots, vests and neckerchiefs, leather fringe, dusty earth tones.
- Jazz Age — The 1920s: straight low-waist dresses with fringe and beads that swing as they dance, three-piece suits and slicked-back hair.
- Sixties Mod — The 1960s mod look: A-line minis, straight cuts, bold colour blocking and pop patterns.
- Regency — England from 1811 to 1820: high-waisted straight gowns in fine cotton and silk, fitted tailcoats with high cravats; light and clean.
- American ’80s new — The American 1980s: neon-contrast nylon windbreakers, high-waist washed jeans, big-shoulder blazers and big hair.
- WWII Era new — Soldiers and civilians of the 1930s–40s on any front: olive and grey-green uniforms, belts and puttees, wide-shoulder civilian suits and nipped-waist tea dresses; the insignia are always invented.

### World Traditional (17)

- Hanbok — Traditional Korean dress in the Joseon cut: a short jacket over a high-waist full skirt, wide robes for men, soft bright colours.
- Kimono — Traditional Japanese dress: straight-cut robes with a wide obi knotted at the back, patterns that follow the season.
- Samurai Armor new — Samurai of the Warring States to Edo periods: black-lacquered plates laced with coloured silk cord and tall helmet crests; plain haori and hakama for everyday.
- Silk Road Robes — Sogdian and Western Regions dress of the 6th–10th centuries, as in Dunhuang murals: lapel robes with narrow sleeves, pearl-roundel brocade and tall boots in rich colour.
- Middle Eastern Robes — Traditional robes of the Gulf Arab world in two lines, white and black: wide, flowing, with gold-thread edging.
- South Asian Sari — Hindu saris for daily, festival and wedding wear: one long cloth wrapped and draped, with embroidery and gold borders, bangles and forehead jewellery.
- Latin American Folk — Village festival dress from Mexico and Central America: large floral embroidery, ruffled wide skirts and bright shawls in vivid contrast.
- African Prints — Contemporary African city wear built on wax-print cotton: vivid geometry and big flowers in modern cuts.
- Thai Court — Court dress of Thailand’s Chakri dynasty: gold-woven silk wraps and a sash across the chest, layers of gold jewellery and a pointed crown.
- Peranakan Nyonya new — Straits Chinese of the early 20th century in Malacca, Penang and Singapore: embroidered kebaya fastened with brooches over a batik sarong; men in white stand-collar shirts and jackets.
- Tibetan — Traditional Tibetan dress: broad robes with a sash tied across, fur trim, and coral, turquoise and silver all over, in strong colour.
- Miao Silver — Miao festival dress from southeast Guizhou: silver crowns, collars and plates sewn onto embroidery and batik; the grander the occasion, the more silver.
- Eastern European Folk — Festival wardrobe of a Slavic farming village: red-and-black embroidery on white linen, embroidered vests and flowered skirts, flower crowns and ribbons.
- Ottoman Court — The court at Topkapi in the 16th–18th centuries: robe over robe in gold brocade with big flower patterns, tall turbans and jewels.
- West African Regal — Festival dress of Yoruba, Asante and Hausa peoples: wide silhouettes in batik and woven cloth, big colour blocks and geometric patterns, tall headwraps that add height.
- Indian Royal — Mughal and Rajput royal dress from the 16th–19th centuries: layers of jewel-coloured silk and gold brocade, gold embroidery and jewels from head to foot.
- Persian Court — The Safavid court at Isfahan in the 17th century, as recorded in miniature paintings: layered gold-woven robes, fine flower patterns, tall turbans and jewels.

### Dynastic (5)

- Imperial Court Regalia — Ming-dynasty court dress: wide robes with crossed or round collars, gold weave and embroidery at the collar and chest; the grander the occasion, the truer the colour and the more gold.
- Tang Opulence — Tang-dynasty Chang’an: high-waisted gowns with wide sheer sleeves and shawls draped over the arms, crisp round-collar robes for men; saturated colour and plenty of gold.
- Song Restraint — Song-dynasty scholar dress: slim straight-lined long shirts and beizi vests, pale soft colours, patterns hidden as tone-on-tone weave.
- Qing Palace Robes — Manchu banner dress of the Qing court: straight robes with stand collars and many rows of edging, tall headdresses, richness in embroidery and trim.
- Court Rank Robes — Ming officials’ dress by rank: round-collar wide-sleeved robes, black gauze hats and rank badges, so rank reads at a glance.

### Fantasy & Sci-Fi (16)

- Near-Future Techwear — Technical clothing for a near-future city: layered technical fabric, taped seams and reflective details in calm neutral colours.
- Cyberpunk Augment — The low streets of a cyberpunk city: black techwear layered with leather, exposed implants and glowing cables, cold neon contrast.
- Wasteland Scavenge — Survivors’ gear cobbled together from salvage: leather, canvas, felt and metal parts in layers, pads riveted from old tyres, plate and leather, colours worn by wind and sand.
- Gothic Court — Victorian Gothic court dress: tight waists and long lines, velvet and lace catching the light on black and deep wine red.
- European Court Gown — Court dress of the European monarchies from the 17th to the 19th century: corsets, wide skirt supports and lace frills; embroidered coats for men with gold and silver thread and gems.
- Starship Uniform — Uniforms of a future star fleet: one-piece fitted cut, high stand collars and asymmetric fronts, with department colours and rank marks to tell people apart.
- Steampunk Attire — Victorian tailoring with brass and gears: nipped waists, long coats and layered leather, buckles; old-fashioned and intricate.
- Fantasy Adventurer — The look of an adventurers’ guild: leather, linen and wool in layers with clean cloaks and belts; higher ranks have finer work and more gold or silver trim.
- Elven Attire — Woodland elves: slim lines and soft draping fabric, leaf and vine patterns, silver with green.
- Dwarven Forge — Mountain dwarves of the forges and mines: heavy leather with metal trim, braids and gem pieces; sturdy and finely worked.
- Orc Tribal — Orc tribes of the wastes: rawhide, fur and metal pieces joined together with tusk and bone ornaments and tribal markings; rough and strong.
- Pirate — Caribbean pirates of the 17th–18th centuries: loose shirts, long coats, wide sashes and boots, bandanas and tricorn hats, hems lifting in the wind.
- Academy Robes — The uniform of a boarding school of magic: robes over shirt and tie, house badges and clear house colours; tidy with a touch of mystery.
- Space Explorer — Near-future crewed spaceflight: white pressure suits with reinforced patches and mission badges, and fitted in-cabin work suits.
- Alien Tribal — Non-human alien tribes whose clothing looks grown from the body: streamlined shapes, shell-like and translucent materials, glowing lines along the body’s contours.
- Merfolk new — Merfolk of the undersea kingdom: a top of pearls and shells or scale armour, a scaled fish tail with translucent fins from the waist down, a crown of pearls and conch.

### Folk & Tribal (9)

- Nordic Viking — Scandinavia in the 8th–11th centuries: wool, linen and fur in layers, braided trim, metal brooches and arm rings.
- Celtic Tribal — Iron Age tribes of Britain and Ireland before Rome: plaid wool cloaks, twisted gold neck rings and spiral designs.
- Steppe Nomad — Riding peoples of the Eurasian steppe: thick long coats, wide sashes and tall boots with fur trim, silver and coral.
- Native American — Peoples of the Plains and woodlands: buckskin with beadwork and fringe, earth tones with bright beads and feathers.
- Andean Highlands — Quechua and Aymara highlands: layered bright handwoven shawls and full skirts, geometric woven bands.
- Polynesian — Pacific islanders of Hawaii, Tahiti, Samoa and Aotearoa: bark-cloth patterns, flower garlands and shell ornaments, light and bright.
- Maasai — East African cattle herders of Kenya and Tanzania: red plaid wraps with bead collars and ear ornaments, slim and clean.
- Inuit Arctic — Arctic dress from Alaska to Greenland: fur-hooded parkas and long boots with contrast trim and beads.
- Aztec — Aztec nobles and warriors of central Mexico: feathered headdresses and cloaks, gold and turquoise, blazing colour.

### Antiquity (3)

- Greco-Roman Attire — The classical Mediterranean from the 5th century BC to the 2nd century AD: whole cloths draped into natural folds, white and pale colours with gold touches.
- Ancient Egyptian Attire — The New Kingdom along the Nile: finely pleated white linen, wide collars and gold jewellery, heavy black eye makeup.
- Byzantine — The Byzantine court of Constantinople: heavy brocade robes inlaid with pearls and gems, gold woven with purple and crimson.

### Jianghu & Xianxia (6)

- Jianghu Travelwear — Martial heroes’ outfits: close-cut belted tunics under a light robe or cloak, hems and hair ribbons lifting as they move; plain, good cloth.
- Chinese Lamellar Armor — The gilded lamellar armour of generals and palace guards in historical dramas: polished plates in overlapping layers, gilded edges and beast-head fittings; the same cut for men and women.
- Immortal Sect Robes — Robes of the righteous sects: layered sheer fabric, wide sleeves, long light sashes and ribbons that lift in the wind even when the wearer stands still.
- Dark Path Robes — The same cut as the sect robes but sunk into black and dark red, with bone ornaments and dark gilding.
- Taoist & Monk Robes new — Taoist priests, monks, feng shui masters and folk exorcists: crossed-collar wide-sleeved robes, kasaya, and satin ritual robes embroidered with trigrams and cranes.
- Divine Regalia — Ceremonial dress of an ancient royal court, black and gold: wide robes and a tasselled crown, gold patterns laid over black.

### Period (3)

- Republican Modern — Shanghai and Beiping in the 1910s–40s: fitted cheongsam and western coats, long and slim lines.
- Retro Hong Kong — Hong Kong in the 1990s, from Central offices to the night market: wide-shoulder suits, high-waist jeans and shiny pieces in bold colour.
- Eighties Work Unit — China’s state-factory compounds in the 1980s: work clothes, polyester shirts and sweaters in mild, clean colours.

## Pairs that go together

The two lists are separate, but many stories want a matching pair. These pairings are a starting point, not a rule.

| Story | Scene Style | Costume Style |
| Joseon-era palace drama | Joseon Korea | Hanbok |
| Korean school drama | Korean High School | Korean School Uniform |
| Japanese school drama | Japanese High School | Japanese School Uniform |
| Boarding-school mystery or romance | Elite Boarding School | Prep School Uniform |
| Regency romance | English Manor | Regency |
| Ottoman palace intrigue | Ottoman Palace | Ottoman Court |
| Mughal-era royal court | Indian Royal Palace | Indian Royal |
| Thai court story | Thai Royal Palace | Thai Court |
| Viking saga | Viking Fjord | Nordic Viking |
| Pirate adventure | Age of Sail Harbor | Pirate |
| Aztec or Andean epic | Pre-Columbian Americas | Aztec or Andean Highlands |
| Mountain village in southwest China | Miao Mountain Village | Miao Silver |
| Tibetan highlands | Tibetan Plateau | Tibetan |
| Pacific island | Polynesian Island | Polynesian |
| Peranakan family saga | Southeast Asian Town | Peranakan Nyonya |
| Samurai story | Japanese Machiya | Samurai Armor |
| Alien planet | Alien World | Alien Tribal |
| Mermaid story | Undersea Kingdom | Merfolk |
| Chinese folk horror | Chinese Folk Occult | Taoist & Monk Robes |
| Second World War | WWII Front | WWII Era |
| Xianxia sect story | Celestial Colossus | Immortal Sect Robes |
| Xianxia demon realm | Dark Asura | Dark Path Robes |

Genre guides that use these pairs: Joseon palace drama · Ottoman palace drama · Regency romance · high school drama · pirate or Viking adventure · mermaid fantasy · folk horror.

## What a style does not decide

- The country and the faces. Target Region sets character ethnicity, names and scene styling; a style only supplies materials, period and mood. When your script clearly states a place and an era, the read follows the script.
- Single places. There is no preset for a courtroom, a restaurant kitchen or a backstage. Those come from your script: the read describes every place the script names, in the scene style you chose.
- Who the characters are. Names, ages and looks come from the script and the character sheets. A costume style sets how the clothes are cut, not who wears them.
- Changes in the story. A torn coat or a wedding dress for one scene is a state of a character, not a new costume style. Costume changes and scene states.

## When nothing fits exactly

- Pick the closest preset for the world and the era. The presets are broad on purpose: Epic Scale carries ancient empires of any culture, and Pastoral Healing covers cosy country stories anywhere.
- After the read, open the scenes and characters in Step 2 and edit any description that is off. Descriptions are plain text, and the next sheets and clips follow what you wrote.
- If you are a member and want a look that no preset has, write a custom style in the same dropdown. How to write one.

## FAQ

### How many scene styles and costume styles are there?

SceneMixer has 74 scene styles in 7 groups and 90 costume styles in 9 groups, as of October 26, 2026. Twenty-nine of them were added on October 25, 2026.

### Is there a style for a Korean, Japanese or American school?

Yes. Scene styles: Korean High School, Japanese High School, American High School and Elite Boarding School. Costume styles: Korean School Uniform, Japanese School Uniform and Prep School Uniform.

### Which styles give a Joseon-era palace drama?

Joseon Korea for the scene and Hanbok for the costume.

### Can I use a period costume in a modern setting?

Yes. Scene style and costume style are chosen separately and both go into the read.

### Do the styles change who the characters are?

No. They set the look of sets and clothes. Ethnicity, names and ages come from the script and Target Region.

### Why is there no preset for a courtroom or a kitchen?

A preset is a whole world, not one room. If your script has a courtroom or a kitchen, the read describes it in the scene style you picked.

### Can I change a style after the script is read?

Not the setting. Edit the scene and character descriptions in Step 2 and regenerate those sheets.

### Why do the cards look different when I choose an anime or 3D visual style?

The scene and costume cards switch to a 3D or 2D version of each preset to match the visual style, so the picture shows what you will get.

Updated November 3, 2026


---

# Scene Image Doesn't Match the Description: How to Write a Location

URL: https://scenemixer.com/guides/scene-image-does-not-match-the-description/
Scene images are empty location plates drawn from the description. Why a threshold or a shop keeps coming out wrong, what to put first, and what never belongs in it.

# When the Scene Image Does Not Match the Description: How to Write a Location the Model Can Draw

Quick answer: A scene image is a reference plate of the location with nobody in it: three layers of depth, the materials, and where the light comes from. The model draws exactly the description, reads the first part most carefully, and ignores what it cannot place. A scene that keeps coming out wrong is almost always described by its meaning (“the threshold where they part”) rather than by its geometry, or has the important thing buried at the end, or asks for people, events or mood that a plate cannot hold. Write the room as a set builder would: what is in front, what is in the middle, what is at the back, what the walls and floor are, where the window or lamp is. First things first.

Open a scene card and read its description: if the first sentence is about the story rather than the space, that is the fix.

## What a scene image is for

Every shot in that location uses the plate as a reference, so it has to be neutral: an empty set, evenly lit, with its layout readable. That is why the site draws scenes at 16:9 with no people and strips mood words, events and light effects before generating. A scene plate is not a still from the film; it is the room the film will be shot in. Same Room in Every Episode.

## Why the picture disagrees with the text

Meaning instead of geometry: “the doorway where she says goodbye” tells the model nothing about the doorway. The key element last: long descriptions are read front-to-back with fading attention; a threshold mentioned in the fourth sentence is often missing. Things a plate cannot hold: people, actions, weather effects, “dramatic” light; SceneMixer removes some of these, and the rest confuse the drawing. Too little: two words (“small kitchen”) leave everything to chance.

## How to write it

Front, middle, back: foreground: a low wooden threshold with a worn brass strip; middle: a narrow entry hall, coats on hooks left, shoes on a rack right; back: a half-open door to a lit kitchen. Then surfaces: oak floor, cream plaster walls, a ceiling lamp with a paper shade. Then the light source: daylight from a window at the far end, lamp off. Put the element the story needs (the threshold) in the first sentence. Nothing about who is there or what happens. The general rule: Describe the Cause, Not the Result.

## Regenerating and checking

Edit the description on the card, regenerate, and compare with the description line by line: is the threshold there, is the door where you said, is the light from the right side? One or two rounds usually settle it. If the image keeps adding people or drama, look for a word in the text that invites it. Regenerate One Sheet.

## Interiors that look like exteriors, and the reverse

Say it explicitly in the first words: interior, windowless, exterior, street. A scene style set in Step 1 can carry outdoor architecture into indoor rooms if the description does not pin the space down; the fix is the same first sentence. Scene style applies at parse.

| Description says | The plate shows | Rewrite |
| The threshold where they part | A random doorway | Foreground: a low wooden threshold with a worn brass strip… |
| A gloomy, oppressive corridor | A dark smudge | Hospital corridor, green tiles to waist height, fluorescent tubes every three metres |
| Kitchen; she cooks while he argues | People, a pan, chaos | Interior, galley kitchen, steel counters left, stove right, window at the end |
| Small shop | Any shop | Corner grocery, shelves both sides, counter at the back, glass door front |

## FAQ

### Why are there no people in scene images?

The plate is a reference for every shot in that room; people in it would appear in every shot.

### The description is long and detailed but still wrong.

Move the element that matters to the first sentence; the tail of a long description is read loosely.

### Does the visual style change the scene image?

Yes, how it is drawn; not what is in it. The layout comes from the description.

### Can I upload a photo of a real place?

Yes, as the scene's image, if you have the rights; it then serves as the plate.

### Does regenerating a scene image affect existing clips?

No. New clips use the new plate.

Updated November 3, 2026


---

# Scene Style and Costume Style Apply at Parse Time, Not Later

URL: https://scenemixer.com/guides/scene-style-and-costume-style-apply-at-parse/
Visual style is applied at every generation; scene style and costume style are written into the descriptions during the read. All four are fixed once the script is read.

# Why the Scene Style and the Costume Style Are Chosen Before the Script Is Read

Quick answer: Step 1 has four style settings, all chosen before the script is read and fixed for the project once the read finishes. They work in two different ways. Visual style (live action, anime, 3D and so on) is applied each time a sheet or a clip is generated. Scene style (the art direction of the sets) and costume style (what people wear) are used while the script is read: they are written into the scene and character descriptions, and the descriptions are what generation reads. So to change the look of sets or costumes you already have, edit their descriptions in Step 2 and regenerate those sheets.

Open a scene in Step 2 of a parsed project and read its description: the scene style is already written into it, in words you can edit.

## The four style settings and when each applies

Visual style: how things are drawn — medium and finish. Applied at generation, every time. Scene style: what the sets are — materials, period, density of dressing. Written by the read into every scene description. Costume style: what people wear — cut, fabric, period. Written by the read into every character’s clothing field; if a clothing field is left empty, the costume style fills it when the sheet is generated. Content type: pacing and structure; also used by the read. All four are set in Step 1 before the script is read and cannot be changed in the project afterwards. Choosing them well: How to Choose Visual Style Settings.

## Why the two are handled differently

A visual style is a way of drawing and can be attached to any description. Scene and costume styles are content: a Song-dynasty courtyard is a different room from a Brooklyn loft, and a character’s clothes are part of who they are. If those were applied at generation on top of descriptions written for another setting, the two would fight — a modern apartment description with a period style pasted on gives you red pillars in a kitchen. So they go in once, during the read, where the descriptions are written in the first place.

## Changing the look after the read

Edit the descriptions yourself in Step 2 — the scene descriptions for a new art direction, the clothing fields for new costumes — and regenerate those sheets; storyboards and clips generated afterwards read the edited descriptions. The visual style itself cannot be changed in the project; for a different one, start a new project with the same script. What you can still change after the read.

## Custom styles

SceneMixer members can enter a custom scene or costume style in their own words. It is applied the same way — during the read — and follows the same writing rule as any description: materials, cuts and periods, not adjectives. Describe the Cause, Not the Result.

## Existing clips

Editing a description changes nothing already generated. Regenerate the sheets and segments you want in the new look. Segments keep every earlier version; a regenerated sheet replaces the old one.

| Setting | Applied | To change it after the read |
| Visual style | At every generation | Not possible in the project; start a new project |
| Scene style | During the read, into scene descriptions | Edit scene descriptions in Step 2, then regenerate |
| Costume style | During the read, into clothing fields | Edit clothing fields in Step 2, then regenerate |
| Content type | During the read | Not possible in the project; start a new project |

## FAQ

### Can I change the scene style after the script is read?

Not the setting. Edit the scene descriptions in Step 2, where the style was written, then regenerate those sheets.

### Can the script be read again with a new style?

Not in the same project. Start a new project with the same script and choose the styles before it is read.

### Can I mix a period costume style with a modern scene style?

Yes; they are independent axes and both go into the read.

### Can I change the visual style later?

No. SceneMixer fixes it once you continue from Step 1, because it shapes every sheet and clip in the project.

### Where do I write a custom style?

In the same Step 1 dropdown, as a member; describe materials and cuts, not moods.

Updated November 3, 2026


---

# AI Video Scenes Inside a Car, Bus, Train or Plane: The Location Image

URL: https://scenemixer.com/guides/scenes-inside-a-car-bus-train-or-plane/
Vehicle interiors are drawn from the front of the cabin looking back, every row of seats in view, so back-seat passengers stay in the back. What changed, what to do.

# Scenes Inside a Car, a Bus, a Train or a Plane: How the Location Image Is Drawn

Quick answer: When a location is the inside of a vehicle — a car, a van, a bus, a train carriage, a plane cabin, a truck cab — its location image is drawn from the front of the cabin looking toward the back, with every row of seats facing the camera and the rear window furthest away. Videos follow the location image, so a passenger written into the back seat now stays in the back seat. Vehicle images made before this change are not redrawn automatically; regenerate the location image once if back-seat scenes come out wrong.

Open the location card of a car interior in Step 2. If the image looks from the back seats toward the dashboard, regenerate it.

## Why the direction matters

A video of a scene is shot the way its location image shows the space. Most car interiors used to be drawn from the back seats looking toward the dashboard — the natural photo — which leaves the back seats out of the picture. A scene with someone in the back then had nowhere to put them, and they ended up in the front. Drawing the cabin from the front, facing back, puts every row in view, so each person has a seat the video can use. Same room in every episode.

## Which vehicles

- Cars and vans — the front seats nearest, the back seats (and third row) behind them, the rear window at the far end.
- Buses, train carriages and plane cabins — the rows of seats running away from the camera, facing it.
- Truck cabs — one row: the driver’s and passenger’s seats, with the sleeper or the back wall behind.

Rooms, streets and every other kind of location are drawn as before. Recognising a vehicle interior is automatic; there is nothing to switch on.

## Writing the location

Describe the vehicle itself: what it is, how old, how clean, the seats, the lighting. Do not write where the camera is (“seen from the back seat”): for vehicle interiors the viewpoint is set by the site and such phrases are left out. Who sits where belongs in the storyboard, not in the location description. How to write a location.

## Older vehicle images

Location images made before this change keep their old angle; they are yours and are not redrawn without you. If scenes with back-seat passengers come out wrong, open the location in Step 2 or in the episode’s asset library and generate it again (7 credits on SceneMixer's default image tier, or free if you still have free sheets). Then regenerate the affected segments. Regenerate one sheet.

## When you use your own photo

A location you upload, or have redrawn from your own photo, keeps your photo’s angle. If back-seat scenes matter, use a photo taken from the front of the cabin looking back. Use your own photo.

| Vehicle | The image shows |
| Car, van | Front seats nearest, back seats behind, rear window far |
| Bus, train, plane | Rows of seats facing the camera, running back |
| Truck cab | Driver and passenger seats, sleeper or back wall behind |
| Your own uploaded photo | Your photo's own angle |

## FAQ

### Do I have to do anything for new projects?

No. New vehicle interiors are drawn this way automatically.

### My back-seat passenger keeps appearing in the front.

If the car's image was made earlier, regenerate it; then check the storyboard says where each person sits, and regenerate the segment.

### Can I choose a different angle for the location image?

For vehicle interiors SceneMixer sets it so every row is visible. Individual shots in the video still use their own angles.

Updated November 3, 2026


---

# Screenplay Template for Short Films and AI Video, With an Example

URL: https://scenemixer.com/guides/screenplay-template/
A screenplay template for short films: scene headings, action lines, character cues and dialogue, with an example SceneMixer reads into locations, a cast and shots.

# Screenplay template for short films and AI video

Quick answer: A screenplay has four building blocks: a scene heading that says where and when, action lines that say what we see, a character cue, and the dialogue under it. Keep the same four blocks and a script is also what SceneMixer reads best: scene headings become locations, character cues become a cast with one voice each, dialogue is spoken in the video and action lines become the shots. Below is a blank template and a one-page example.

Copy the example, change the names and the place, and paste it into a new project to see the cast and storyboard it produces.

## The four blocks

| Block | Write | Example |
| Scene heading | INT. or EXT., place, time | INT. LAUNDROMAT - NIGHT |
| Action line | What we see, in the present tense | Rain on the window. MARA, 30s, sits among the dryers. |
| Character cue | The speaker's name in capitals | MARA |
| Dialogue | What they say | I thought you'd left already. |

Using scene headings and character cues · Writing dialogue so it is picked up.

## A one-page example

INT. LAUNDROMAT - NIGHT

Rain on the window. MARA, 30s, sits among the dryers with a folded suitcase beside her. JON, 30s, comes in shaking out an umbrella.

JON
You said the 9 o'clock train.

MARA
I thought you'd left already.

He sits two machines away. The dryers turn. Neither of them speaks for a long moment.

MARA
It's the last load.

## How the example becomes a project

| In the example | In SceneMixer |
| INT. LAUNDROMAT - NIGHT | One location, drawn once |
| MARA and JON | Two characters, each with a sheet and a voice |
| The two lines of dialogue | Spoken in the video by the right person |
| The action lines | Numbered shots in the storyboard |

Script to film · Script to animation · Upload a script and read it.

## Tips for a script that renders well

- One place per scene; move scenes, not characters.
- One action per action line.
- Short lines; the voice carries them best.
- No inner thoughts: say them aloud or leave them out.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw Mara and Jon and the laundromat; they are the whole cast and the whole set. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Format | Reads best as |
| .txt | Plain text with the four blocks |
| .docx | Unpacked to plain text in your browser |
| .md | Plain text |
| PDF | Not accepted; copy the text out |

## FAQ

### What is a screenplay template?

A layout of scene headings, action lines, character cues and dialogue.

### Does SceneMixer need screenplay format?

No. Prose works too, but the four blocks are read most reliably.

### Which files can I upload?

.txt, .docx and .md, or pasted text. SceneMixer refuses PDFs.

### What becomes a location?

Each distinct place in a scene heading.

### What becomes a voice?

Each character cue is one character with one voice in SceneMixer.

### How long can a script be?

SceneMixer takes up to 300,000 characters of English per project.

### Is reading it free?

Reading on SceneMixer costs 4 credits per 1,000 characters, English counted lighter; new accounts start with 50 credits.

Updated November 3, 2026


---

# Script Length Caps and How Characters Are Counted (3,333 = 1,000)

URL: https://scenemixer.com/guides/script-length-caps-and-how-characters-count/
Every length on the site is counted in Chinese-equivalent characters: alphabetic text counts 3,333 characters as 1,000. The caps, the maths, and what fails.

# Script Length Caps, and Why 3,333 English Characters Count as 1,000

Quick answer: Every length on the site — the parse cap, the parse fee, the storyboard fee, the AI Create target — is counted in Chinese-equivalent characters. Chinese, Japanese and Korean text counts one character as one; alphabetic text (English, German, Spanish, Russian, Arabic and the rest) counts 3,333 characters as 1,000, because a thousand Chinese characters carry about as much story as that many letters. Scripts under 10,000 equivalent characters are read in one pass, longer ones in two passes up to 100,000; the counter on the home page already does the conversion for you.

Paste an English script on the home page: the counter shows both the raw character count and the cap, already converted.

## Where the cap shows

Under the script box on the home page a counter shows characters against the cap. It reads the language of what you pasted: mostly alphabetic text gets the lighter count, mostly Chinese text the full count. Mixed text is judged by which script dominates. The same number decides the parse price shown on the button.

## The factor, and why it exists

An English sentence needs roughly three times as many characters as the same sentence in Chinese. Charging and capping by raw characters would make an English script three times as expensive as its Chinese twin for the same amount of story, so alphabetic text is scaled: 3,333 characters count as 1,000. The factor applies to every alphabetic script — Latin, Cyrillic, Greek, Arabic, Hebrew, Armenian, Georgian — and not to Chinese characters, kana or hangul, which stay one-to-one. SceneMixer’s cap, price tag and actual charge all use the same conversion, so the number you see is the number you pay.

## The three caps

Scripts under 10,000 equivalent characters are read in one pass. Longer scripts are read in two passes up to 100,000 equivalent characters (for alphabetic text that is 300,000 raw characters). Anything beyond is refused before you pay. How the two-pass read works: Long Scripts Are Read in Two Passes.

## What the count prices

Parsing costs 4 credits per 1,000 equivalent characters of the whole script. Each episode’s storyboard costs 1 credit per 100 equivalent characters of that episode’s share of the script. AI Create prices its target length the same way, in the language you write in. Where your credits go in one episode.

## AI Create and the cap

AI Create writes to a target length you choose; targets that would not fit the parse cap are hidden from the menu. If a script still comes out over the cap, SceneMixer fails the run, the credits are refunded and no project is created — pick a shorter target and run again. Choosing the Script Length and What It Costs.

| Text | How it counts | Example |
| Chinese, Japanese, Korean | 1 character = 1 | 10,000 characters = 10,000 |
| English, German, Spanish, Russian, Arabic… | 3,333 characters = 1,000 | 33,330 characters = 10,000 |
| Mixed | By the dominant script | A Chinese script with English names counts as Chinese |

## FAQ

### Is the cap in words or characters?

Characters, converted to Chinese-equivalent. For English, divide the character count by about 3.3.

### Why is my Japanese script priced like Chinese?

Kana and kanji carry as much per character as Chinese, so SceneMixer counts them one-to-one.

### Does the factor change the price after parsing?

No. The cap, displayed price and charge on SceneMixer use the same conversion.

### My novel is 400,000 characters. Can I parse it?

Not as one SceneMixer project. Split it into parts under the two-pass cap and parse each part.

### Where is the count for AI Create?

In the length menu on SceneMixer's AI Create tab; targets over the cap are not offered.

Updated November 3, 2026


---

# Seedance 2.0 Price per Second: Tiers, Limits and Drafts | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-0/
The Seedance 2.0 family on this site: what its three tiers cost per second, what a short generation is good for, and when a tighter cut beats a long take.

# Seedance 2.0: the family for tight cuts, priced per second

Quick answer: On this site Seedance 2.0 is the family whose takes top out at 15 seconds. Seedance 2.0 renders at 1080P for 63 credits per second (about $0.39 at the lowest credit price); Seedance 2.0 Fast at 720P for 25 (about $0.16); and Seedance 2.0 · 480P for 13 (about $0.08). Shorter generations suit tighter cutting, and they keep the same dialogue and ambient sound.

## When a short segment is the right cut

A 15-second generation is one breath: an exchange, a reveal, a turn. For a lot of short-drama beats that is exactly the cut, and it keeps every render moving quickly. When a scene needs to stay in one shot longer — a slow walk, a long speech, a chase held in a single frame — the site has families with a higher ceiling; the trade-offs are laid out in segments, seconds and how they cut.

Length is per generation, not per scene: a scene that needs more screen time is simply more than one segment, and the storyboard can be split that way from the start.

## The three tiers and what they cost

One ceiling, three resolutions, prices per second of rendered picture:

| Tier | Credits per second | About, at the lowest credit price | One generation, up to | Resolution |
| Seedance 2.0 | 63 | $0.39 | 15 seconds | 1080P |
| Seedance 2.0 Fast | 25 | $0.16 | 15 seconds | 720P |
| Seedance 2.0 · 480P | 13 | $0.08 | 15 seconds | 480P |

Line the two families up at the same resolution and the numbers sit side by side: 63 credits per second on Seedance 2.0 against 75 on Seedance 2.5 Pro, 25 on Seedance 2.0 Fast against 38 on Seedance 2.5 Lite, and 13 on Seedance 2.0 · 480P against 19 on Seedance 2.5 · 480P — and any segment can move up later without redoing the storyboard.

## What you can make with it

The still above is The Last Envelope, an episode made end-to-end on this site with Wan 3.0 · 720P — storyboard, renders, voices, composite; the tier is one setting, so the same project flow runs on any Seedance 2.0 tier instead.

So the natural fit is anything cut short and often: a first pass over a whole episode, a quick look at a beat before committing to it, alternate readings of the same line, vertical versions of a finished scene. A finished episode composites from the active version of every segment and can be shared straight from the editor.

## Drafts, keepers and versions

Nothing here forces a decision up front. Render a segment at the entry tier, watch it, then re-render it on the tier the final cut deserves; both versions stay listed on the segment, and the composite uses whichever one is active. Put the whole episode through the editor before you finish anything: the storyboard, the voices and the timing are all cheaper to change than pixels.

And when a take only needs a tweak — a trimmed opening, a swapped line — the editor’s segment tools handle that without a full re-render. The mechanics are in editing a shot and regenerating. The path from a pasted script to a rendered segment, sketched:

## How to start

Before drafting, set the step in Select Model, the button in the editor’s top bar: Seedance 2.0 · 480P is where a first pass belongs — 13 credits a second — and a short beat is what this family drafts best. The rest of the flow is unchanged: a project holds the whole run, its episodes each a run of short segments. SceneMixer opens the account with 50 welcome credits to cover the starting fees; the script analysis is billed by length, sheets beyond the 15 free ones per account are billed each, and every delivered second of video is billed.

Then let the episode decide its own rate — draft cheap, finish where it counts is the route the site’s guides describe, and editing a shot without a full re-render keeps a near-miss cheap to fix.

## FAQ

### What does Seedance 2.0 cost per second?

On SceneMixer: 63 credits per second on Seedance 2.0, 25 on Seedance 2.0 Fast, and 13 on Seedance 2.0 · 480P — about $0.39, $0.16 and $0.08 at the lowest credit price. Charged when a segment delivers; a failed render is refunded or held for a check, not charged.

### What fits in one generation on these tiers?

One scene beat: an exchange, a reveal, a move — a single generation on SceneMixer tops out at 15 seconds. A scene with more beats is several segments, and the storyboard can plan it that way — the segment-length guide walks through the cutting rhythm.

### Can I mix Seedance 2.0 segments with other tiers in one episode?

Yes. The tier is one setting — the editor’s tier button, or Settings → Model Preferences → Video Model — that applies to the segments you render next, and every segment records the tier it was made with, so an episode can blend drafts and finals. The composite stitches whatever the active versions are. Only the ceiling is fixed: on the 2.0 tiers a single generation never runs past 15 seconds.

### Which of the three 2.0 tiers should I use?

The same pattern as everywhere on the site: draft on the entry tier, finish on the one you will publish. The three map to larger screens, phone screens and first passes — each has its own page: the standard tier, Fast, and the entry tier.

### Is Seedance 2.0 still worth it next to 2.5?

If your cuts are short, yes: 63 credits per second against 75 on the two top tiers, 25 against 38 on the middle ones, and 13 against 19 on the entry ones — and the lower ceiling only rules out scenes that want to run long. The 2.5 page prices its own table if you want to line the two up side by side.

Updated November 3, 2026


---

# Seedance 2.0 · 480P Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-0-480p-tier/
What the entry tier of the Seedance 2.0 family costs per second on SceneMixer, how long one take runs and when to move up.

# Seedance 2.0 · 480P: price per second, limits and when to use it

Quick answer: : Seedance 2.0 · 480P costs 13 credits per second (about $0.08), the lowest rate in its family: takes up to 15 seconds at 480P with dialogue and ambient sound. It is where a first pass over a short beat belongs.

The still above is from Rocket Pup, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

15 or 30-second segments explains the segment length setting this tier lives under.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

## What a segment on this tier can be

Length: up to 15 seconds in one generation — none of the Seedance 2.0 tiers renders 30-second takes. The segment length is chosen when an episode’s storyboard is generated; if a segment is longer than 15 seconds, this tier is greyed out for it in the editor. Resolution: : 480P in the project’s aspect ratio. Sound: lines and ambient sound generated with the picture. Shots: several per segment, cut where the storyboard says.

## References it accepts

Per generation:

## What it costs

SceneMixer charges 13 credits per second. Of the other Seedance 2.0 tiers, Fast (720P) is 25 and the standard tier (1080P) is 63 per second. A 15-second segment is 15 × the rate. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## What it is for

Cheap iteration on 15-second segments: getting the blocking, the cut points and the delivery of the lines right before spending Fast or standard rates. Fine detail is soft at 480P by design; that is a job for the higher tiers, not a reason to touch the storyboard.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When to switch tier

Up to Fast for shareable 720P; up to standard for 1080P keepers. Across to a tier with 30-second takes (Seedance 2.5 · 480P renders up to 30 seconds) when a scene needs one unbroken take longer than 15 seconds.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Why does this tier stop at 15 seconds?

That is the segment length all three Seedance 2.0 tiers render on SceneMixer. The 30-second long-take mode needs a Seedance 2.5 or Wan tier.

### Is it the cheapest video tier on the site?

It is the cheapest of the Seedance 2.0 tiers. Compare with Wan 3.0 · 480P on the pricing page; the per-second figures differ between the two sites.

### Can I mix this tier with 30-second tiers in one episode?

Yes, as long as the segments themselves fit each tier’s cap; each segment remembers its tier.

### Does a rejected generation cost anything?

Normally no: SceneMixer returns the credits. The content-check section above explains the daily limit on refunded rejections.

### Is sound included?

Yes. Dialogue and ambient sound are generated with the picture.

Updated November 3, 2026


---

# Seedance 2.0 Fast Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-0-fast-tier/
What Seedance 2.0 Fast costs per second on SceneMixer, the middle step of its family, how long one take runs and when it is enough.

# Seedance 2.0 Fast: price per second, limits and when to use it

Quick answer: : Seedance 2.0 Fast costs 25 credits per second (about $0.16): takes up to 15 seconds at 720P, between Seedance 2.0 · 480P and Seedance 2.0. It reads the same references as the other two, and for many scenes it is the step where the picture is good enough to cut with.

The still above is from Neon Debt, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

The Seedance 2.0 480P tier is for drafting; this page is the step up.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

## What changes against 480P

The inputs do not; the picture does. Both faces in a two-shot read, hair and cloth keep their grain. On a phone at full screen this is where resolution stops being what you notice.

The Seedance 2.0 cap stays: up to 15 seconds per generation, the project’s aspect ratio, sound in the same render, every version kept.

## References it accepts

Same slots as the other Seedance 2.0 tiers:

## What it costs

SceneMixer charges 25 credits per second, against 13 on the 480P tier and 63 on the standard tier. Draft at 480P, regenerate keepers here; the versions panel keeps both and the composite uses the active one. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## When Fast is the right stop

Sharing a segment link, posting a vertical clip, sending a cut for notes, assembling an episode watched on a phone or tablet. The standard tier is for large screens and longer edits.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When to switch tier

Down to 480P while the storyboard is changing; up to standard for a 1080P keeper; across to a tier with 30-second takes when a scene needs an unbroken take longer than 15 seconds.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Why is it called Fast?

It is the 720P one of the three Seedance 2.0 tiers; on this site the name marks the middle of the three, not a different workflow.

### Does Fast render 30-second takes?

No. The Seedance 2.0 tiers on SceneMixer render segments up to 15 seconds; use a Seedance 2.5 or Wan tier for 30-second takes.

### Can I mix Fast and 480P segments in one episode?

Yes. Each segment remembers its tier; the composite stitches each segment’s active version.

### Is 720P enough for a vertical short?

For phone playback, usually yes.

### Is a Fast render an upscale of my 480P render?

No. Every generation is a fresh render from the same inputs.

Updated November 3, 2026


---

# Seedance 2.0 Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-0-standard-tier/
What the standard Seedance 2.0 tier costs per second on SceneMixer, how long one take runs, and when the highest step of the family is worth its rate.

# Seedance 2.0: price per second, limits and when to use it

Quick answer: : Seedance 2.0 costs 63 credits per second (about $0.39), the top rate of its family: takes up to 15 seconds at 1080P with dialogue and ambient sound. Take length is the ceiling to plan around: scenes that need longer are built from several segments.

The still above is from Crimson Vow, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

Original vs play version: which file to download after a 1080P render.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

It appears in the tier list as plain Seedance 2.0, without a resolution suffix.

## What the standard tier buys

Room to look closely: background faces in a wide shot, a slow push-in that holds, texture in skin, cloth and rain. Most valuable on a television, a projector, a desktop player, and in edits you will crop or reframe. On a phone the gap to Fast is small.

Family cap unchanged: up to 15 seconds per generation, the project’s aspect ratio, sound in the same render, every version kept.

## References it accepts

Same slots as the other Seedance 2.0 tiers:

## What it costs

SceneMixer charges 63 credits per second, against 13 on 480P and 25 on Fast. A 15-second segment is 15 × 63. Draft low, keep what works, pay this rate once per keeper. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## Downloads and playback at this tier

A finished segment has an original file (what you download; what the composite is built from) and a play version (a lighter copy for the in-page player and share links, same resolution, compressed harder). Judge the render from the original.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When it is not worth it

While the storyboard is moving; for a clip only watched on phones; for a segment whose problem is blocking or dialogue. Use 480P to iterate and Fast to share; come here for the final version of a segment that already works. For a take longer than 15 seconds, no Seedance 2.0 tier helps — switch to a Seedance 2.5 or Wan tier.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Is this the same as Seedance 2.5 Pro?

No. Both are 1080P, but the Seedance 2.0 tiers on SceneMixer render 15-second segments and Seedance 2.5 renders up to 30; prices differ too.

### Does it take longer than 480P?

Usually somewhat longer for the same seconds.

### Do I have to regenerate the whole episode here?

No. Regenerate only the keepers; the composite uses each segment’s active version.

### Can I download without the AI badge?

On a paid SceneMixer plan, new renders come without the badge by default, and you can choose to keep it.

### Can a segment longer than 15 seconds use this tier?

No. In the editor this tier is greyed out for such a segment; pick a Seedance 2.5 or Wan tier.

Updated November 3, 2026


---

# Seedance 2.5 Price per Second: What Each Tier Costs | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-5/
The Seedance 2.5 family on this site: what each tier costs per second, how the three differ, how long one take can run, and the reference limits they share.

# Seedance 2.5: three tiers, priced per second

Quick answer: On this site Seedance 2.5 comes in three tiers. Seedance 2.5 Pro renders at 1080P for 75 credits per second (about $0.47 at the lowest credit price); Seedance 2.5 Lite at 720P for 38 (about $0.24); and Seedance 2.5 · 480P for 19 (about $0.12). One generation runs up to 30 seconds with dialogue and ambient sound, and every tier takes the same references — pick by the picture you need and the budget you have.

## Where the tiers sit on this site

Seedance 2.5 is one family of video tiers on SceneMixer, and picking one is a single setting: the tier button in the episode editor, or Settings → Model Preferences → Video Model. The choice applies to the segments you render from then on; segments already made keep the tier they were rendered with, and each version records it. The price tag on a segment’s generate button always shows the tier you are about to use, before you spend anything.

The three tiers cover the three resolutions a finished episode usually needs: drafts, phone screens and larger displays. Each has its own page if you want the detail behind it: Pro, Lite, and the entry tier.

## What each tier costs per second

Three tiers, one price scale. Everything SceneMixer charges for a video segment is per second of rendered picture, at the tier you picked:

| Tier | Credits per second | About, at the lowest credit price | One generation, up to | Resolution |
| Seedance 2.5 Pro | 75 | $0.47 | 30 seconds | 1080P |
| Seedance 2.5 Lite | 38 | $0.24 | 30 seconds | 720P |
| Seedance 2.5 · 480P | 19 | $0.12 | 30 seconds | 480P |

The step diagram below shows the order the tiers sit in, with the live numbers in its legend:

## What you can make with it

The still above is The Eagle’s Vow, an episode made end-to-end on this site with Wan 3.0 · 720P — the same project flow this page describes, from the pasted script to the composite; the tier is a single setting, so the same storyboard renders on any Seedance 2.5 tier instead.

Everything the site produces can be made on these tiers: multi-episode series cut into segments, one-scene tests, dialogue scenes with voices, and vertical versions of the same story. The family does not change what a project can be — it changes how each second looks and what it costs. A project holds its episodes; each episode is a run of segments; a finished episode can be composited from the active version of every segment and shared.

Dialogue and ambient sound come with the picture, so a talking scene needs no separate audio step. And because every render is kept as a version, a segment can be re-rendered on a higher tier as the keeper while the cheap draft stays in the same list.

## How to start

Open Select Model in the editor’s top bar before the first render and set the drafting rate: Seedance 2.5 · 480P at 19 credits a second is the one to try a segment on, and then only the keepers move to the two steps above. Nothing earlier is included, mind: SceneMixer reads the script for a fee set by its length, and while it grants 15 sheets free to every account, a larger cast is billed image by image. Renders are charged per second delivered.

Two guides go with this: the draft-then-finish pattern in full, and what fits in one take before the shot list is planned.

## FAQ

### What does Seedance 2.5 cost per second?

SceneMixer charges 75 credits per second on Seedance 2.5 Pro, 38 on Seedance 2.5 Lite, and 19 on Seedance 2.5 · 480P — about $0.47, $0.24 and $0.12 at the lowest credit price. Credits are taken when a segment delivers.

### How long can one Seedance 2.5 take run?

SceneMixer allows up to 30 seconds in a single generation, on all three tiers. Longer scenes are built as several segments; the site splits a storyboard into segments for you.

### Which of the three tiers should I pick?

Draft on the entry tier and finish on the one viewers will see — the draft-then-finish pattern. Regenerating a segment on SceneMixer is a new charge either way, and every version is kept, so nothing is lost by drafting low.

### How is this different from Seedance 2.0 on this site?

2.0 is the family with 15-second takes; 2.5 doubles the ceiling to 30 and SceneMixer prices the three resolutions differently, so the choice often comes down to whether your cuts want more air.

### Seedance 2.5 or Wan 3.0?

Compare the two families at the resolution you need. On SceneMixer, Wan 3.0 · 1080P renders at 38 credits per second where Seedance 2.5 Pro asks 75; Wan 3.0 · 720P at 19 where Seedance 2.5 Lite asks 38; and Wan 3.0 · 480P at 10 where Seedance 2.5 · 480P asks 19. A single generation runs up to 30 seconds on the Seedance 2.5 tiers and up to 30 seconds on the Wan 3.0 ones, with up to ten images and five voice clips as references. So the price at the resolution you need is the thing to weigh; the side-by-side comparison lays both columns out on one page.

Updated November 3, 2026


---

# Seedance 2.5 · 480P Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-5-480p-tier/
What the entry tier of the Seedance 2.5 family costs per second on SceneMixer, how long one take runs, and what it is good for before you move up.

# Seedance 2.5 · 480P: price per second, limits and when to use it

Quick answer: : Seedance 2.5 · 480P costs 19 credits per second (about $0.12), the entry rate of its family: takes up to 30 seconds at 480P with dialogue and ambient sound. Use it to prove a segment — framing, timing, performance — before moving the keepers to Seedance 2.5 Lite or Seedance 2.5 Pro.

The still above is from Starfall Command, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

The site default, Wan 3.0 · 480P, is the other 480P long-take tier; this page is the Seedance 2.5 side.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

## What a segment on this tier can be

Length: up to 30 seconds in one generation (15 seconds when the episode’s storyboard was generated with 15-second segments). Resolution: : 480P in the project’s aspect ratio. Sound: SceneMixer generates lines and ambient sound with the picture, each character in their own voice. Shots: a segment can hold several shots; the storyboard sets the cuts.

## References it accepts

Per generation, shared by all three Seedance 2.5 tiers:

## What it costs

19 credits per second. In the same family, Lite is 38 and Pro is 75 per second. Against the site default, Wan 3.0 · 480P is 10 per second. Pick this tier when you want to stay inside the Seedance 2.5 family for the segments you will later finish on Pro. Credits are taken on delivery only.

## What it is for

Drafting inside the family: blocking a 30-second take, hearing the lines, confirming who is in frame and where the cuts fall, before spending Lite or Pro rates. Because SceneMixer’s three tiers share inputs and rules, nothing you learn here has to be redone on the way up.

At 480P, fine detail is soft by design; judge blocking, timing and dialogue here, judge texture on Pro.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When to switch tier

Up to Lite (720P) for a segment you will share; up to Pro (1080P) for the version you publish. Across to Wan 3.0 · 480P if you only need a cheaper 480P draft and do not care which family finishes it. Each segment on SceneMixer remembers its tier; mixing is normal.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Is this the same as the default 480P tier?

No. It is the 480P entry of the Seedance 2.5 family, priced and labelled separately from SceneMixer's default, Wan 3.0 · 480P. Both render 30-second segments with sound.

### Can a segment drafted here be finished on Pro without changes?

Yes. The three Seedance 2.5 tiers on SceneMixer share references, voices and segment length; only the tier changes.

### How long can a segment be?

SceneMixer allows up to 30 seconds when the episode’s storyboard was generated with 30-second segments, 15 when it uses 15-second segments. The length is chosen when you generate the storyboard.

### Does a rejected generation cost anything?

Normally no: a rejection returns the credits and shows in your SceneMixer history. The content-check section above explains the daily limit on refunded rejections.

### Is sound included?

Yes. Dialogue and ambient sound on SceneMixer come out of the same generation as the picture.

Updated November 3, 2026


---

# Seedance 2.5 Lite Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-5-lite-tier/
What Seedance 2.5 Lite costs per second on SceneMixer, the middle step of its family, how long one take runs and when it is the right place to stop.

# Seedance 2.5 Lite: price per second, limits and when to use it

Quick answer: : Seedance 2.5 Lite costs 38 credits per second (about $0.24): takes up to 30 seconds at 720P, the middle step of its family. It reads the same references as the other two tiers, which makes it the natural place to render keepers before deciding whether a segment deserves Seedance 2.5 Pro at 75.

The still above is from The Last Envelope, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

The 480P entry tier is the family’s drafting tier; this page is the step up.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

## What changes against the 480P tier

Inputs do not change; the picture does. Two faces in a shot both read; hair, cloth and rain keep their grain. On a phone at full screen this is the tier at which resolution stops being the thing you notice.

Rules stay the family’s: up to 30 seconds per generation, the project’s aspect ratio, sound in the same render, every version kept.

## References it accepts

Same slots as 480P and Pro:

## What it costs

SceneMixer charges 38 credits per second, against 19 on the 480P tier and 75 on Pro. The usual pattern is to draft at 480P and regenerate keepers here; the versions panel keeps every render, and the composite uses the active one. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## When Lite is the right stop

Sharing a segment link, posting a vertical clip, sending a cut for notes, or assembling an episode that will be watched on a phone or tablet. Also the sensible tier for a 30-second take you have already proved at 480P: same take, cleaner picture, without the Pro rate on seconds that will be watched small.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When to switch tier

Down to 480P while the storyboard is still changing. Up to Pro for a version headed for a large screen or a longer edit. Each segment keeps the tier it was made with.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Is Lite a lower-quality model or a lower resolution?

On SceneMixer, it is the 720P tier of the family; the storyboard, references and voices are the same as on the other two tiers.

### Can I mix Lite and 480P segments in one episode?

Yes. Each segment remembers its tier; the composite stitches the active version of each.

### Is 720P enough for a vertical short?

For phone playback, usually yes. Pro is for large screens and longer edits.

### Is the segment length the same on Lite?

Yes: SceneMixer allows up to 30 seconds per segment on every Seedance 2.5 tier, when the storyboard uses 30-second segments.

### Is a Lite render a sharpened copy of my 480P render?

No. Every generation is a fresh render from the same inputs; expect the same scene with small differences in motion.

Updated November 3, 2026


---

# Seedance 2.5 Pro Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-5-pro-tier/
What Seedance 2.5 Pro costs per second on SceneMixer, how long one take runs, what it reads as references, and when the finishing tier is worth the rate.

# Seedance 2.5 Pro: price per second, limits and when to use it

Quick answer: : Seedance 2.5 Pro costs 75 credits per second (about $0.47 at the lowest credit price), and for that a take runs up to 30 seconds at 1080P with dialogue and ambient sound in the render. It reads the same references as Seedance 2.5 · 480P and Seedance 2.5 Lite, so it is the finishing pass: prove the segment on a cheaper tier, then render the version you publish here.

The still above is from The Veil at Dawn, an episode this site took from script to composite on Wan 3.0 · 480P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

Original vs play version: which file to download after a Pro render.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

## What Pro buys

Room to look closely. Wide shots keep background faces readable, a slow push-in holds together, skin, cloth, rain and dust keep their texture. It matters on a television, a projector or a desktop player, and in any edit you will crop or reframe. On a phone held at arm’s length the gap to Lite is small, which is why Pro is a finishing tier rather than a default.

Family rules unchanged: up to 30 seconds per generation, the project’s aspect ratio, sound in the same render, every version kept.

## References it accepts

Same slots as the rest of the family:

## What it costs

SceneMixer charges 75 credits per second, against 19 on the 480P tier and 38 on Lite. A 30-second take is 30 × 75. Draft low, keep what works, pay this rate once per keeper. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## Downloads and playback at this tier

A finished Pro segment has two files. The original is what you download and what the episode composite is built from. The play version is a lighter copy used by the in-page player and share links so phones on slow connections do not stall: same resolution, compressed harder. Judge a Pro render from the original, not from the browser preview.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When Pro is not worth it

While the storyboard is still moving; for a clip only ever watched on phones; for a segment whose problem is blocking or dialogue rather than detail. The 480P tier shows you the same thing for a fraction of the credits, and Lite is the sharing tier. Come to Pro for the final version of a segment that already works.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Is Pro the highest resolution on the site?

It is the 1080P top of its family. MiniMax H3 offers a 2K render with different reference rules and 15-second segments.

### Does Pro take longer?

Usually somewhat longer than 480P for the same seconds.

### Do I need to regenerate the whole episode on Pro?

No. Regenerate only the keepers; the composite uses each segment’s active version whatever its tier.

### Can I download without the AI badge?

On a paid SceneMixer plan, new renders come without the badge by default, and you can choose to keep it. Download the original file.

### Is the segment length the same on Pro?

Yes: up to 30 seconds per segment.

Updated November 3, 2026


---

# Seedance 2.5 Prompt Guide: Writing the Long Take | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-5-prompt-guide/
How to write shots for Seedance 2.5 on this site: when a long segment should be one held shot, how many lines fit, and how to end a take so the next one joins.

# Seedance 2.5 prompt guide: writing the long take

Quick answer: On Seedance 2.5 Pro, Seedance 2.5 Lite and Seedance 2.5 · 480P one generation runs up to 30 seconds, against 15 on the 2.0 family. So the writing question is not how to fit a scene into a short clip but how much of the ceiling a scene deserves: write a single held shot when the action is continuous, a short chain when it is not, and leave air at the end for the join.

## What the longer ceiling is for

Nothing about how you write changes between tiers of the family: the same shot grammar — marker and seconds, framing and camera, who is in the picture, what happens, the line, the ambience — feeds all three, and the picture is the only thing that moves with the tier. What the family adds is headroom. A take can be a conversation that plays out while two people walk, a push-in that has time to arrive, a reveal that is allowed to land before the cut. If a scene never needed that, a family with a shorter ceiling does the same job; if it did, this is the reason to be here.

## One shot or several

The still above is from The Eagle’s Vow, an episode this site took from script to composite on Wan 3.0 · 720P. Its segments were written the way a Seedance 2.5 Pro segment would be, and the long-take storyboard setting that sets a segment to the longer ceiling is chosen when the storyboard is generated.

Decide first whether the beat is one thing or several. If one person crosses a room, says what she came to say and reaches the door, that is one continuous action: write one shot, one framing, one camera move, and give it the seconds it needs.

[Shot 1] 14s: Tracking shot alongside, medium shot. @Mara walks from the desk to the window, @Jules a step behind her. @Mara (quiet, not turning): I already know what it says. (voice: 🔊Mara) @Jules (low): Then why keep it? (voice: 🔊Jules) Ambient: rain on the glass.

If the beat changes framing, it is several shots, and the seconds of each add up to the segment. Writing them as a chain is not a downgrade: a held shot is right when the action is continuous, and cuts are right when the scene turns.

## Pace the lines to the seconds

A long take tempts people to put more in it. The measure that keeps it honest is speech: about three English words to the second, so a line needs roughly a third as many seconds as it has words, and the action around it needs its own time on top. Fill the ceiling with dialogue and there is no room for anyone to move. A take that lets two people trade a line each and then breathe reads as a scene; a take that crams the whole argument reads as a recording.

Seconds are tuned in the storyboard, before anything is generated, and changing them costs nothing. Render the draft on Seedance 2.5 · 480P at 19 credits a second, read the pacing, and only then spend on the tier the viewers will see.

## End the take where the next one can start

Each segment opens where the previous one ended: whoever was seated, wherever the hands were, that is the starting state for the next call. So write the last shot of a long take to land on a position you would be happy to continue from — a settled pose, a clear framing, nobody caught mid-step. A take that ends in the middle of a gesture hands the next segment a problem it has to solve in its first second. The longer the take, the more this matters, because there are fewer joins to hide in; which segments the composite joins shows where the joins fall.

## References and what to attach

On SceneMixer, the family takes up to ten images and five voice clips per generation, with about fourteen seconds of voice in total, the same as the Wan steps. A long take that carries several speakers needs its voices attached, and a crowd does not need a sheet for every face. Spend the images on the people the camera will find, and name only them in the shot.

## When a shorter take is better

Not every scene wants the ceiling. A reaction needs a couple of seconds, an insert even fewer, and padding them to fill a segment adds seconds nobody asked for. Write the length the beat has, not the length the tier has. The price page and the family page have the per-second numbers; the comparison with the 2.0 family is the one to read when the choice is the ceiling.

## FAQ

### How long can a Seedance 2.5 segment be?

SceneMixer allows up to 30 seconds in one generation on all three tiers, with dialogue and ambient sound. Longer scenes are built from several segments.

### Should a long take be one shot or several?

One shot when the action is continuous — a walk and talk, a push-in that has to arrive. Several when the framing changes. The seconds of the shots always add up to the segment.

### How many lines fit in one long take?

As many as the seconds allow, at about three English words to the second, with room left for the action. A take full of dialogue leaves nobody time to move.

### Does the writing change between the three Seedance 2.5 tiers?

No. The shots are written the same way for Seedance 2.5 Pro, Seedance 2.5 Lite and Seedance 2.5 · 480P; what changes is the resolution and the price per second, so a take written once renders on any of them.

### What should the last shot of a take look like?

A settled frame: a clear framing and a position the next segment can start from. The next segment opens where this one ended.

Updated November 3, 2026


---

# Seedance 2.5 vs MiniMax H3: Price, Length and Limits | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-5-vs-minimax-h3/
Seedance 2.5 and MiniMax H3 side by side on SceneMixer: price per second, resolution, the longest take each allows and the references each takes, with when each suits.

# Seedance 2.5 vs MiniMax H3: price, length and limits

Quick answer: Both families use the same shot grammar; they differ in limits. Seedance 2.5 Pro renders takes up to 30 seconds with up to ten images and five voice clips, from 19 to 75 credits a second; MiniMax H3 renders up to 15 seconds at 2K with nine images and three voice clips, from 16 to 25. Choose by the length of the take and the size of the cast first, and by price second.

## What the two have in common

You write the same thing for both. The storyboard shots — marker and seconds, framing and camera, who is in the picture, one thing that happens, the line, the ambience — are the input, and dialogue and ambient sound come out with every render on both families. Nothing in a shot list has to be rewritten to move between them; the setting that decides is the tier button in the editor’s top bar, and it applies to what you render next. Segments already made keep their tier, and each version records it, so a project can use both without losing track.

## Side by side

The numbers on the site today, from the same table the pricing page prints:

| | Seedance 2.5 | MiniMax H3 |
| Tiers | Seedance 2.5 Pro, Seedance 2.5 Lite, Seedance 2.5 · 480P | MiniMax H3, MiniMax H3 Fast |
| Resolution | 1080P, 720P, 480P | 2K, 768P |
| Credits per second | 75 / 38 / 19 | 25 / 16 |
| About, at the lowest credit price | $0.47 / $0.24 / $0.12 | $0.16 / $0.10 |
| Longest take | 30 seconds | 15 seconds |
| Images per generation | Up to ten | Up to nine |
| Voice clips per generation | Up to five | Up to three, and an image is needed alongside |
| Sound | Dialogue and ambient sound with every generation | Dialogue and ambient sound with every generation |

Read the table as limits, not as quality. The families draw the picture differently and the tier pages describe what each is for; what the table settles is how long a take can run, how many references a segment can carry, and what a second costs at the resolution you pick.

## Where each one fits

The still above is from Rust City Knockout, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard would render on either family here; what would change is the take length and the references available to each segment.

When the take decides. A beat that must run in one unbroken shot for longer than 15 seconds has only one home of the two: Seedance 2.5 Pro, Seedance 2.5 Lite and Seedance 2.5 · 480P go up to 30. A beat that fits comfortably inside 15 — an exchange and its reaction, a reveal — fits either, and MiniMax H3 is where the cuts inside a generation are the point.

When the cast decides. A segment with several speakers wants its voices attached, and the two families cap them differently: five clips on Seedance 2.5 Pro against three on MiniMax H3, with an image needed alongside on the latter. A scene with a crowd of named faces is easier to carry with the larger budget.

When the picture decides. : MiniMax H3 renders at 2K and MiniMax H3 Fast at 768P; the Seedance 2.5 tiers render at 1080P, 720P and 480P. If the largest picture is what the finish needs, compare the price per second at the resolution you want in the table, not the names.

## Drafting across the two

Nothing stops a project from drafting on one family and finishing on the other. Render the storyboard on the entry tier of either, read the pacing, then re-render only the keepers on the tier the viewers will see; both versions stay listed on the segment and the composite uses the active one. The only thing to check first is the cast: if a segment relies on more than three voices or more than nine images, it has to be trimmed before it moves to MiniMax H3. The draft-then-finish pattern works through the mechanics, and each family has its own page: Seedance 2.5 and MiniMax H3.

## FAQ

### What is the difference between Seedance 2.5 and MiniMax H3 on SceneMixer?

The limits. Seedance 2.5 Pro takes run up to 30 seconds with up to ten images and five voice clips; MiniMax H3 takes run up to 15 seconds with up to nine images and three voice clips, at 2K. Both render dialogue and ambient sound with every generation.

### Which one is cheaper per second?

Compare at the resolution you need. SceneMixer's table lists 19, 38 and 75 credits per second for the three Seedance 2.5 tiers and 16 and 25 for the two MiniMax H3 tiers, with the resolutions beside them.

### Can I use both in one project?

Yes. The tier you set applies to the segments you render next, and every version remembers the tier it was made with, so segments from both families can sit in the same episode.

### Do I have to rewrite my shots to switch?

No. Both families take the same shot text. The one thing to check is the cast: a segment that carries more than three voice clips or nine images has to be trimmed before it moves to MiniMax H3.

### Which should I draft on?

Whichever family has the entry tier you want: Seedance 2.5 · 480P at 19 credits a second, or MiniMax H3 Fast at 16. Draft there and re-render only the keepers on the finishing tier.

Updated November 3, 2026


---

# Seedance 2.5 vs Seedance 2.0: What Changes on This Site | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-5-vs-seedance-2-0/
Seedance 2.5 against Seedance 2.0 on SceneMixer, tier by tier: the price per second at each resolution, the longest take, and how that changes where a scene gets cut.

# Seedance 2.5 vs Seedance 2.0: what changes on this site

Quick answer: The two families come in the same three resolutions, and the main thing that changes is the length of a take: Seedance 2.5 Pro runs up to 30 seconds in one generation, Seedance 2.0 up to 15. The price per second differs tier by tier — 13 to 63 credits on 2.0, 19 to 75 on 2.5 — so the choice is the ceiling you need against the rate you pay.

## The pairs, tier by tier

Each Seedance 2.0 tier has a Seedance 2.5 tier at the same resolution, so you can compare them the way you would pick a step: look across a row.

| Resolution | Seedance 2.5 | Per second | Seedance 2.0 | Per second |
| 1080P | Seedance 2.5 Pro | 75 credits ($0.47) | Seedance 2.0 | 63 credits ($0.39) |
| 720P | Seedance 2.5 Lite | 38 credits ($0.24) | Seedance 2.0 Fast | 25 credits ($0.16) |
| 480P | Seedance 2.5 · 480P | 19 credits ($0.12) | Seedance 2.0 · 480P | 13 credits ($0.08) |

The rows show prices and nothing else. The ceiling is the other column that matters: 30 seconds a generation on the 2.5 tiers and 15 on the 2.0 tiers. Prices are read from the same live table the pricing page prints, so a change on the site changes this page.

## Where the length changes the cut

The still above is from Starfall Command, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard can be grouped into fewer, longer segments or more, shorter ones, depending on the ceiling of the family it is rendered on.

On a family with the shorter ceiling, a scene that wants about a minute of continuous action is several segments, and the joins fall between them. On the longer one the same scene may be one or two. Neither is wrong: every segment opens where the previous one ended, and a join is invisible when the ending is a settled frame. What changes is how many joins you plan for, and how many renders you pay for when one beat needs a retake.

## What does not change

The writing. The shots go into the storyboard in the same shape for both families — marker and seconds, framing and camera, who is in the picture, one thing that happens, the line, the ambience — and the references are the same on SceneMixer: up to ten images and five voice clips per generation. Dialogue and ambient sound come out of every render on both. You can change the tier between renders without touching the text, and every version remembers which one it was made on.

## How to choose

Start from the take. If most of your beats fit comfortably in a short segment, Seedance 2.0 and its siblings are the family for tight cuts; if some beats need a single held shot, the longer ceiling is the reason to be on Seedance 2.5 Pro, Seedance 2.5 Lite or Seedance 2.5 · 480P. Then look across the row at the resolution the finish needs. Drafting is the same on both: render on the entry tier, read the pacing, re-render only the keepers. The two family pages cover the rest: Seedance 2.0 and Seedance 2.5. If the comparison you want is with the Wan steps, that one has its own page.

## FAQ

### What is the difference between Seedance 2.5 and Seedance 2.0 on SceneMixer?

The length of a take and the price at each resolution. A Seedance 2.5 Pro generation runs up to 30 seconds and a Seedance 2.0 generation up to 15; both families have a tier at each of the three resolutions, and the table above lists the price per second beside each.

### Do I need to rewrite my storyboard for 2.5?

No. Both families take the same shot text and the same references. If a scene was split to fit a shorter ceiling, you can leave it as it is or regroup it into longer segments.

### Can one project use both?

Yes. The tier applies to the segments you render next, and every version records the tier it was made with.

### Which tier pairs with which?

Seedance 2.0 pairs with Seedance 2.5 Pro, Seedance 2.0 Fast with Seedance 2.5 Lite, and Seedance 2.0 · 480P with Seedance 2.5 · 480P: the same resolution in each pair.

Updated November 3, 2026


---

# Seedance 2.0 for Short Drama: Workflow Guide | SceneMixer

URL: https://scenemixer.com/guides/seedance-2-short-drama/
How to use Seedance 2.0 inside a series pipeline: tiers, prompting, reference images for consistency, shot length and cost per second.

Guide

# Seedance 2.0 in a Real Short-Drama Pipeline

Updated November 3, 2026

Quick answer: Seedance 2.0 (ByteDance's video model family) is one of the strongest engines for vertical short drama in 2026 — but the model is only half the job. A watchable series also needs storyboard prompts written for the model's strengths, reference images that lock character identity across shots, and per-segment economics. SceneMixer builds exactly that pipeline, with Wan 3.0 as its default video engine: it writes model-tuned prompts for every 4–15 second segment, injects character reference art automatically, and lets you pick the model tier per episode.

## What Seedance 2.0 is good at (and how a pipeline exploits it)

- Multimodal references — Seedance 2.0 accepts image references alongside text. SceneMixer auto-attaches each character's reference sheet (full-body art plus face crops) to every segment that character appears in; that, not prompt wording, is what keeps faces consistent across a series.

- Native audio — dialogue and ambience generate with the picture, which suits dialogue-driven vertical drama.

- 4–15 second clips — matching the natural shot length of micro drama. One segment = one clear action beat; SceneMixer's storyboard writer paces exactly that.

- Physics and close-ups — two-person dialogue close-ups, emotional beats, rain/wind/light interactions render reliably; large crowds and complex choreography don't. The storyboard generator is constrained to favor the former.

## Prompting principles the pipeline bakes in

Community prompt guides for Seedance 2.0 converge on a few rules: describe one action per clip, state camera movement explicitly, define lighting and mood, keep prompts tight. SceneMixer's segment scripts are generated in that shape automatically — start state, shot-by-shot beats with camera language, timing that matches the segment duration — and stay fully editable when you want manual control.

## Standard vs Fast tier

| Tier | When to use | Cost behavior |

| Seedance 2.0 (standard) | Hero episodes, final takes, maximum quality | Higher per-second rate |

| Seedance 2.0 Fast | Drafting, iteration, volume production | Lower per-second rate, suited to drafting and volume |

In SceneMixer the tier is a per-episode dropdown; billing is per second of output, so a 90-second episode is a predictable cost either way, and regenerating one 8-second segment never re-bills the episode.

## What it costs

SceneMixer is credit-based — you pay for what you generate, not a flat seat license. Credit packages start at $1.49; membership tiers from $7.99/month add monthly credits, higher concurrency, and member-only features such as custom style prompts, uploads, and AI image fine-tuning. Script analysis is priced by text length (4 credits per 3,333 characters, reference images billed separately, with 15 free per account across characters, scenes and props); storyboard scripts are billed per episode at 1 credit per 333 characters, and regenerating one costs the same; video generation starts at 10 credits per second (≈ $0.06/s at the lowest credit price) of output (the rate depends on the model tier you pick), and episode compositing is 1 credit per second — see the full pricing table.

## Frequently asked questions

### Do I need my own Seedance API key to use SceneMixer?

No. SceneMixer runs its video models on the backend (Wan 3.0 by default) — you pick the model tier from a dropdown and pay in credits; there is no API setup.

### Does prompt wording alone keep characters consistent in Seedance 2.0?

Not reliably. Identity lock comes from image references: SceneMixer generates reference art per character at analysis time and attaches the right references to every segment automatically.

### Can I write my own prompts for a segment?

Yes — every segment's storyboard script on SceneMixer is editable for free, and you can regenerate just that segment with your edits, keeping version history.

### Which aspect ratios does the pipeline generate?

16:9 landscape (SceneMixer's default) and 9:16 vertical (the usual short-drama frame), chosen per project.


---

# Seedream 5.0 Lite: 2K Sheets Near the Default Price | SceneMixer

URL: https://scenemixer.com/guides/seeddream-5-lite-tier/
Seedream 5.0 Lite (listed as SeedDream 5.0 Lite) on SceneMixer: 2K sheets close to the default price, a description read end to end, and how it compares with Pro.

# SeedDream 5.0 Lite

Quick answer: : SeedDream 5.0 Lite gives you the SeedDream family’s 2K output and end-to-end reading of a description at 8 credits per sheet (about $0.05 at the lowest credit price) — close to the default tier’s price, well under Pro’s. It is the tier for 2K locations and props, and for long descriptions on a budget; Pro adds finer skin on faces.

SeedDream 5.0 Pro is the sibling for faces; this page is the budget 2K tier.

## Where you pick it

Settings → Model Preferences → Image Model. The choice applies to every sheet you generate from then on — character, location and prop images in step 2, regenerations, state images — until you change it. Sheets already made keep their pixels; nothing is re-rendered by switching.

## What a sheet on this tier looks like

2K, like Pro: a 9:16 character sheet or a 16:9 location image with roughly twice the default tier’s pixel count. Rooms, streets, props and costumes come out with the detail the video tiers can use; faces are good, with Pro keeping the edge in skin texture under close-up.

## How it reads your description

End to end, the family trait. A long location description with a stated light source, or a prop with several materials, lands as written. It also follows literal phrasing literally: describe causes, not results.

## What it costs

Every account has 15 free sheets in total (across projects, any type) on a project whose script has been read, drawn on the default image model, Qwen-Image 3.0, only. On this tier, SceneMixer charges each sheet from the first one: 8 credits (about $0.05 at the lowest credit price). A rejected generation is not charged, and most failures are refunded automatically; a few kinds of failure are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

Against the default tier at 7 per sheet and Pro at 25.

## Content checks

Description checked before rendering; a flagged description is not charged. See content rules and checks.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the character or scene description tripped the check | Edit the description and generate again |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Generate again in a few minutes |
| A sheet that ignores part of the description | Too many details competing in one description | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |

## When to switch tier

To Pro for the faces the series hangs on. To the default tier when 2K is not needed and every credit counts. Existing sheets are untouched.

| | This tier |
| Role | 2K locations, props and long descriptions on a budget |
| Price per sheet | 8 credits |
| Long descriptions | Read end to end |
| Faces | Good; Pro keeps the edge in close-up |
| State images | Edited from the base sheet |
| Free sheets | 15 per account, default tier only |

## FAQ

### Is Lite lower resolution than Pro?

No; both render at 2K. Pro renders finer skin texture on faces.

### Is Lite a good default for locations?

Yes, when you want 2K plates; the default tier is cheaper if 2K is not needed.

### Does it read long descriptions fully?

Yes, like Pro.

### Can I use Lite for scenes and Pro for characters?

Yes, by switching the preference between the two groups.

### Does switching affect existing sheets?

No. Regenerate the ones you want on the new tier.

Updated November 3, 2026


---

# Seedream 5.0 Pro on SceneMixer: 2K Sheets and Price per Sheet | SceneMixer

URL: https://scenemixer.com/guides/seeddream-5-pro-tier/
Seedream 5.0 Pro (listed as SeedDream 5.0 Pro) on SceneMixer: 2K sheets, a long description read end to end, the most photographic skin, and its price.

# SeedDream 5.0 Pro

Quick answer: : SeedDream 5.0 Pro renders sheets at 2K and reads the whole of a description — the end as carefully as the start — which is why it was the site’s default before the cheaper Qwen tier took over. It costs 25 credits per sheet (about $0.16 at the lowest credit price). Pick it for photographic skin in close-up and for detailed descriptions where every clause has to land.

Qwen-Image 3.0 Pro is the recommended tier; this page is the alternative for detail-heavy work.

## Where you pick it

Settings → Model Preferences → Image Model. The choice applies to every sheet you generate from then on — character, location and prop images in step 2, regenerations, state images — until you change it. Sheets already made keep their pixels; nothing is re-rendered by switching.

## What a sheet on this tier looks like

2K output: a 9:16 character sheet has roughly twice the pixel count of the default tier’s, and a 16:9 location image the same. Skin is the tier’s strength — pores, uneven tone, the small asymmetries that make a face read as a person rather than a render. It also draws exactly what the description says, including result-style phrases, so a description that says red at the nostrils gets a red nose.

## How it reads your description

End to end. Where the Qwen tiers weight the opening, this tier acts on the last paragraph as readily as the first, which makes it the tier for long, layered descriptions — period costume with many pieces, a room with a specified light source, a character defined by a dozen small features. The flip side: it follows literal phrasing literally. Describe the cause of what you want to see, not the look you expect.

## What it costs

Every account has 15 free sheets in total (across projects, any type) on a project whose script has been read, drawn on the default image model, Qwen-Image 3.0, only. On this tier, SceneMixer charges each sheet from the first one: 25 credits (about $0.16 at the lowest credit price). A rejected generation is not charged, and most failures are refunded automatically; a few kinds of failure are held until support checks them. Regenerating a sheet is a new charge; the old image stays until you replace it.

Against the default tier at 7 per sheet and SeedDream 5.0 Lite at 8.

## Content checks

The description is checked before rendering; a flagged description is not charged, so edit the wording rather than retrying. See content rules and checks.

## Common errors

What you may meet on this tier:

| What you see | Usual cause | What to do |
| “Blocked by content safety rules — edit the script or description and retry.” | A word or phrase in the character or scene description tripped the check | Edit the description and generate again |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once | Generate again in a few minutes |
| A sheet that ignores part of the description | Too many details competing in one description | Put identity details first; trim decoration |
| Generate All stops partway | Balance ran out; the remaining sheets were not charged | Top up, then generate the rest |

## When to switch tier

To SeedDream 5.0 Lite for locations and props at 2K without the Pro price. Back to a Qwen tier when you are iterating on a description and cost per attempt matters more than the last few percent of texture. Existing sheets are untouched by switching.

| | This tier |
| Role | 2K sheets; photographic skin; long descriptions |
| Price per sheet | 25 credits |
| Long descriptions | Read end to end |
| Literal phrasing | Followed literally; describe causes |
| State images | Edited from the base sheet |
| Free sheets | 15 per account, default tier only |

## FAQ

### Is this the highest-resolution image tier?

It is one of several 2K tiers, with SeedDream 5.0 Lite, the Nano Banana, GPT Image 2.5 and Grok Imagine tiers; the Qwen tiers render at lower pixel counts.

### Why was it replaced as the default?

Cost. Sheets are references for the video tiers, and the cheaper tier gives them what they need for most projects.

### Will it draw exactly what I write?

More literally than the other tiers. Write causes (a lamp to the upper right) rather than results (light on her forehead).

### Does it handle very long descriptions?

Yes; it reads to the end. That is its main advantage over the default tier.

### Can I mix tiers within one cast?

Yes. Switch the preference before regenerating the sheets you want on this tier.

Updated November 3, 2026


---

# Seedream 5.0 Pro and Lite on SceneMixer: 2K Sheets, Price, Use

URL: https://scenemixer.com/guides/seedream-5/
Seedream 5.0 Pro and Lite on SceneMixer (listed as SeedDream): 2K reference sheets, a description read end to end, the price per sheet, and which one to pick.

# Seedream 5.0 on SceneMixer: Pro and Lite for 2K character, location and prop sheets

Quick answer: SceneMixer offers both Seedream 5.0 models as image tiers, listed in Settings as SeedDream 5.0 Pro (25 credits per sheet, about $0.16) and SeedDream 5.0 Lite (8 credits, about $0.05). Both render the reference sheets of an AI series at 2K and read a description end to end — the last clause as carefully as the first. Pro adds the most photographic skin of any tier here, which is why it suits faces in close-up; Lite is 2K at close to the default tier’s price, which makes it the tier for locations and props. Every account’s first 15 sheets are free, but only on the default image model; on Seedream every sheet is charged.

Each tier has its own page: Seedream 5.0 Pro and Seedream 5.0 Lite.

## Seedream, SeedDream: the same model

Seedream is ByteDance’s image model family. This site’s settings panel lists the two tiers as SeedDream 5.0 Pro and SeedDream 5.0 Lite; they are the Seedream 5.0 models. If you came here searching for Seedream 5.0 Pro or Lite, these are the tiers.

## What they do on SceneMixer

Image tiers here draw sheets, the references every shot of a series is matched to: each character full length on white, each location as an empty plate with its light source readable, each prop alone on white, and a state image for every second look. The video tiers then render shots against those sheets.

The Seedream tiers have two habits that set them apart. They draw at 2K, roughly twice the pixel count of the default tier’s sheets. And they read the whole of a long description, where the Qwen tiers weight the opening. The flip side of the second habit: they follow literal phrasing literally.

## Where you pick them

Settings → Model Preferences → Image Model. The choice applies to every sheet you generate from then on — character, location and prop images in step 2, regenerations, state images — until you change it. Sheets already made keep their pixels; nothing is re-rendered by switching.

## Pro or Lite

| | Seedream 5.0 Pro | Seedream 5.0 Lite |
| Shown in Settings as | SeedDream 5.0 Pro | SeedDream 5.0 Lite |
| Credits per sheet | 25 | 8 |
| Output | 2K | 2K |
| Long descriptions | Read end to end | Read end to end |
| Skin in close-up | The most photographic of any tier here | Good; Pro keeps the edge |
| Best for | The leads’ faces, detailed costume | Locations, props, a large cast at 2K |

A common split: Lite for every location and prop, Pro for the two or three faces the series hangs on, the default tier for the rest of the cast.

## How they compare with the other image tiers

| Image tier | Credits per sheet | Best at |
| Seedream 5.0 Pro | 25 | 2K, photographic skin, long descriptions |
| Seedream 5.0 Lite | 8 | 2K locations and props on a budget |
| Qwen-Image 3.0 Pro | 18 | Recommended tier; faces in close-up |
| Qwen-Image 3.0 | 7 | The default; a whole cast cheaply |

Both Seedream tiers on SceneMixer read Chinese descriptions well, so they suit projects whose descriptions are in Chinese as much as English ones. Which image tier for your sheets has the full decision table.

## Writing for Seedream

Because Seedream reads to the end and draws what it reads, it rewards long, layered descriptions and punishes careless ones. Write causes, not results: a lamp to the upper right of the bed, not warm light on her forehead; a burn scar on the back of the hand, not a hint of red. Leave mood out of sheets. Worked examples for characters, locations and props are in the Seedream 5.0 prompt guide.

## Price and free sheets

Every account has 15 free sheets in total, drawn on the default image model, Qwen-Image 3.0, only. On Seedream every sheet is charged: Seedream 5.0 Pro is 25 credits per sheet and Lite 8. A description blocked by the content check is not charged; regenerating a sheet is a new charge.

## FAQ

### Can I use Seedream 5.0 on SceneMixer?

Yes. Both Seedream 5.0 Pro and Lite are image tiers, listed in Settings, Model Preferences, Image Model as SeedDream 5.0 Pro and SeedDream 5.0 Lite.

### Is SeedDream the same as Seedream?

Yes. The settings panel on SceneMixer spells it SeedDream; the models are ByteDance's Seedream 5.0 Pro and Lite.

### How much does Seedream 5.0 Pro cost here?

SceneMixer charges 25 credits per sheet; Lite is 8. The first 15 sheets on every account are free only on the default image model, so every Seedream sheet is charged.

### Is Seedream 5.0 free?

On SceneMixer, your first 15 sheets are free only on the default image model, so Seedream sheets are always charged.

### What resolution are Seedream sheets?

2K on SceneMixer for both Pro and Lite, about twice the pixel count of the default tier's sheets.

### Seedream 5.0 Pro or Lite?

Pro for faces in close-up and detailed costume; Lite for locations, props and a large cast at 2K.

### Does Seedream read Chinese descriptions?

Yes, well. SceneMixer recommends it for descriptions written in Chinese.

### Can Seedream make video?

No. It draws the reference sheets; the video tiers on SceneMixer render the shots.

Updated November 3, 2026


---

# Seedream 5.0 Prompt Guide: Characters, Locations and Props

URL: https://scenemixer.com/guides/seedream-5-prompt-guide/
Writing descriptions for Seedream 5.0 Pro and Lite on SceneMixer: it reads every word and draws it literally. The order, three examples, and what to leave out.

# Seedream 5.0 prompt guide: writing character, location and prop sheets

Quick answer: Seedream 5.0 reads a description to the end and draws what it reads, literally. That makes it the best tier here for long, layered descriptions — and the least forgiving of careless ones. On SceneMixer you write only the subject; the site frames every sheet the same way (a character full length on white, a location as an empty plate, a prop alone on white). So: describe what can be drawn, in a sensible order, write causes rather than results, and leave light, mood and effect words out. Below: the order for each kind of sheet, three worked examples and the fixes for the usual failures.

What the tiers are and cost: Seedream 5.0 on SceneMixer.

## The one habit that changes how you write

Most image models weight the start of a prompt and skim the rest. Seedream does not: the last sentence counts as much as the first. Two consequences follow. A long description is fine, as long as every clause is something you want drawn. And a phrase that describes a result is drawn as that result: a hint of red at the nostrils becomes a red nose; light falls softly on her face becomes a lit face on every shot the sheet is used in.

## Character sheets

| Order | What to write |
| 1. Face | Shape, brows, eyes, nose, mouth, skin as a person has it, hair line and cut, the age you see |
| 2. Anchor | One feature nobody else in the cast has |
| 3. Build | Height, frame, how they stand |
| 4. One outfit | Named piece by piece; a second outfit is a state, not a second clause |

A man in his early twenties with a narrow face, sharp cheekbones and hooded dark eyes; straight black brows, a slightly crooked nose that was once broken; light brown skin with a few acne marks on the jaw; short black hair shaved at the sides. A small silver ring through the top of the left ear. Medium height, wiry, hands often in his pockets. A faded black bomber jacket over a grey hoodie, black jeans, worn white high-top sneakers.

Every word is drawable; the once-broken nose and the ear ring are the anchors. Notice what is missing: no handsome, no moody lighting, no anime style (the style is a project setting). More on faces in describe a face so it is not a generic face.

## Location sheets

A location sheet is a plate to shoot in: the place with nobody in it, three layers of depth, the materials, and where the light comes from. Write it as a set builder would — front, middle, back, walls and floor, the light source — not by what happens there.

A narrow ground-floor noodle shop at night, seen from the doorway. In front: two small square steel tables with red plastic stools. In the middle: a long counter on the left with a glass case of raw noodles and a row of steel pots. At the back: a swinging kitchen door with a round window. White tiled walls, a grey concrete floor, a single fluorescent tube on the ceiling above the counter, a red lantern hanging in the doorway.

No people, no event (where they first met cannot be drawn), no weather effects that belong to one shot. The fluorescent tube and the lantern are sources the video can light from. If a location keeps coming out wrong, see scene image does not match the description.

## Prop sheets

A prop sheet is the object alone on white, lit to show its material. Describe shape, size against a hand, material, wear and one identifying mark, in the state it is usually carried. A structurally different form — a cracked glass, a drawn blade — is a separate state with its own image.

A brass pocket watch about the size of a palm, round, with a hinged lid engraved with a crane in flight. The brass is worn pale at the edges; a short broken chain hangs from the crown. Lid closed.

The engraved crane is the identifying mark; lid closed is the carried state, and the open watch would be a state of its own. One kind of prop never gets a reference: a photo or portrait of a character. Props: how to describe them has the rules.

## What to leave out

- Light and mood on any sheet. They are carried into every shot; write them per shot in the storyboard.
- Effect words. Masterpiece, 8K, ultra-detailed, cinematic draw nothing; Seedream already renders at 2K.
- Negatives. No hat, not smiling: say what is there instead (bare-headed, mouth closed). See say what is there, not what is not.
- Results. Write the scar, not the redness; the lamp, not the glow.
- Real people, brands and known characters. They trip the IP check on every tier.

## Pro or Lite for each kind of sheet

Both read the same way and draw at 2K. Use Pro (25 credits per sheet) for faces in close-up, where its photographic skin shows; use Lite (8) for locations and props, where the difference is small. Every account’s first 15 sheets on SceneMixer are free, but only on the default image model: get the wording right there, then regenerate the keepers on Seedream.

## Fixing a sheet that came out wrong

- Something you did not ask for appears: a result phrase was drawn literally. Rewrite it as a cause.
- A location has people or fog: the description mentions an event or weather. Remove it.
- The costume swallows the figure: the outfit is described at length. Cut it to the main pieces.
- Two outfits blended: move the second to a state.
- Blocked by the text check: rephrase in neutral visual terms; a blocked sheet is not charged.

To change one detail of a sheet you like, write only the change: write the change, not the whole picture.

## FAQ

### What is the best prompt structure for Seedream 5.0?

On SceneMixer: for a character, face, anchor, build, one outfit; for a location, front, middle, back, materials, light source; for a prop, shape, size, material, wear, one mark.

### Does Seedream read long prompts?

Yes, to the end. That is its main difference from the default tier, and why every clause should be something you want drawn.

### Should I write the prompt in English or Chinese?

Either. Seedream reads both well.

### Do quality words like 8K or masterpiece help?

No. They describe nothing drawable, and the tier already renders at 2K.

### Why does my location image have people in it?

Usually the description names an event or a crowd. A location sheet is an empty plate; people come from the storyboard.

### Why did it draw a red nose?

Seedream draws result phrases literally. Write the cause in the storyboard (a cold, crying) and keep the sheet neutral.

### Can I add lighting to a location?

Name the source (a window, a lamp, a fluorescent tube) so the video can light from it. Leave the mood of each shot to the storyboard.

### Do the same rules apply to Seedream Lite?

Yes. Pro and Lite read descriptions the same way; Pro differs in skin texture on faces.

Updated November 3, 2026


---

# “Segment May Not Have Been Read Correctly”: What to Fix

URL: https://scenemixer.com/guides/segment-may-not-have-been-read-correctly/
The check that runs when you generate a segment you edited by hand: the four things it can flag (no shots, unrecognised lines, lines no shot uses) and how to fix each.

# “This segment may not have been read correctly”: what the prompt means and how to fix it

Quick answer: If you edited a segment’s storyboard text by hand and press generate, the web editor first checks that the text can still be read as numbered shots and spoken lines. When something cannot, a box opens: Segment N may not have been read correctly by the system, with the reasons listed, then The video may not come out as you expect. Please check this segment first. Generate anyway? Cancel and nothing is charged — fix the text first. Confirm and the segment is generated at its normal price. The check only looks at segments you inserted or edited, and it never rewrites your text.

Wrote the dialogue differently from the standard shape? How to write a line, and what Save tidies up.

## When the prompt appears

Only on the website’s episode editor, and only when you press Generate on a segment that is yours: one you inserted, or one whose text has been edited since it was written (by hand or with AI edit, so it has a text history). A segment as first written by the site is not checked, and neither is the free preview.

The box has two choices: Cancel keeps you on the segment so you can fix it, Generate anyway sends it as it is. SceneMixer charges nothing until you confirm, and a charge, if any, is the normal price on the button. The check is a warning, not a lock: if you meant the text to look that way, confirm it.

In Generate all, a segment that would trigger the box is not asked about one by one — it is skipped, and the list says Segment N was skipped: it may not have been read correctly. Fix it and generate that segment on its own. Generate all segments at once.

## The four reasons it can list

| What the box says | What it means | Fix |
| It isn’t split into shots in the [Shot 1] 3s: format, so the system can’t find the shots and will send the whole text to the video model as it is. | The segment has no numbered shots with seconds, so there is nothing to time, cast or voice. | Put every shot on its own line starting [Shot 1] 3s:, [Shot 2] 4s: and so on, or restore the earlier version from History. |
| In shot N, a line with a voice tag wasn’t recognized as dialogue, so it may not be spoken. | The voice chip is there but the line around it is not in a shape the reading accepts, for example the colon is missing or the speaker chip is not a character chip. | Rewrite it as @Name (how): words. (voice: Name) and press Save, which tidies the common casual ways of writing a line. |
| In shot N, text written as “Name: line” wasn’t recognized as dialogue (no voice tag), so it may not be spoken, or may be spoken by the wrong person. | It looks like a line, but without a voice chip it is only text in the picture description. | Add the voice chip after the words: type @, open the Voice tab, pick the same character. |
| The line “…” isn’t used by any shot, so it won’t be spoken in the video. | The segment still lists the line, but no shot carries it any more, usually because you cut the shot or moved the words. | Put the line back inside the shot where it is spoken, or delete it if you no longer want it said. |

More than one reason can appear in the same box. The shot numbers in the box are the numbers you see in the text, so you can go straight to them.

## Fixing it, step by step

- Press Cancel in the box. You are back on the segment.
- Press Edit and read the text from the top. Every shot starts with its number and seconds; every spoken line sits inside its shot.
- Fix what the box listed. For lines, use the @ button: Cast tab for who speaks, Voice tab after the words.
- Press Save. Read the message that appears — Save also tidies casual dialogue into the standard shape and tells you which shots it touched.
- Press Generate again. If the box comes back with fewer reasons, work through the rest; if it does not come back, the segment reads cleanly.

Editing the text is free; only the video is charged. If you pasted a whole screenplay scene into a segment, split it into shots first — how to write one shot · what you can change in the storyboard.

## Is it safe to press Generate anyway?

Sometimes. If the first reason (no shots) is the only one and you wrote the segment as one flowing paragraph on purpose, the video model gets your paragraph as written and may still do something good. But the segment will not be timed shot by shot, and no line in it will be given a voice by the site. If the reason is about dialogue, the likely result is a silent character or the wrong voice, and the clip is charged either way.

A cheap way to decide: fix the text and generate at a low-cost tier first. Checklist before you generate videos · Spend fewer credits.

## Not the same as the amber note after Save

After you save a hand edit, an amber note can appear under the text about links or references that went missing in a copy and paste. That is about what you saved. The prompt on this page is about whether the text can be read at the moment you generate. They are separate checks. Amber notes after you save · When the video says a line you never wrote.

## FAQ

### Why did SceneMixer ask me to check my segment before generating?

You edited the segment's text, and the check found something it cannot read as shots or spoken lines. The box lists what and where; it is a warning, not a block.

### Am I charged if I press Cancel?

No. Nothing is charged until you confirm Generate anyway or generate again after fixing.

### Does the check change my text?

No. It only reads it. Save is the step that tidies dialogue into the standard shape.

### Why was a segment skipped in Generate all?

Generate all skips a segment that would show this box, and says so. Fix it and generate that segment on its own.

### What does a line with a voice tag not recognized mean?

The voice chip is present but the line is not in a shape the reading accepts, often a missing colon or a speaker that is plain text rather than a character chip. Rewrite it and press Save.

### Do I need a voice chip on every line?

Yes. A line without one is not treated as dialogue, so it may not be spoken or may be spoken by the wrong person.

### Will the check appear for segments the site wrote?

Only after they are changed by hand or with AI edit, or if you inserted the segment yourself. A segment as first written by SceneMixer is not checked.

### Can I turn the check off?

No. You can always confirm Generate anyway.

Updated November 3, 2026


---

# Segment Versions: Switch Back, Callback Clips, Which One Downloads

URL: https://scenemixer.com/guides/segment-versions-switch-back-reuse-download/
Every generation of a segment is kept in Version History. How to switch the active one, how callback segments reuse a clip, and which version downloads and composites.

# Segment Versions: Switching Back to an Earlier Take, Callback Clips, and Which Version Downloads and Composites

Quick answer: A segment is not overwritten when you regenerate it: each take is stored as a version, and one of them is active. The active version is what plays in the preview, what downloads, what the share link shows and what the composite joins. Switch between versions from the segment’s version control, at no cost. A segment the storyboard marks as a callback to an earlier shot reuses that clip automatically, also free. Deleting a segment removes all of its versions.

Regenerate any segment once, then open its version control: both takes are there with their tiers and dates; click the older one to make it active.

## Where versions live

On the timeline card and in the preview pane there is a version control once a segment has more than one take. It lists every generation with its tier, its length and when it was made. Click one to make it active; the thumbnail, the preview and the price history all follow. Nothing is generated by switching. The card’s other icons: Timeline Cards: Every Icon.

## What the active version decides

Preview: what plays when you select the segment. Download: the original and play files of the active take. Share: the link plays the active take; switching versions changes what viewers see. Composite: the episode is joined from each segment’s active version at the time you press composite. So you can keep a 480P draft active on one segment while others are at 1080P, and composite later once you have upgraded it. Composite and Share an Episode.

## Callback segments reuse a clip

When an episode ends by returning to its opening shot, the storyboard can mark that final segment as a callback to the earlier one. Pressing generate on it copies the earlier segment’s clip instead of generating: it becomes this segment’s active version, and SceneMixer charges nothing for it. Generate the earlier segment first; until it has a clip, the callback is refused with a message saying so. The callback segment’s own text does not change what the clip shows. There is no picker for reusing a clip on any other segment; a repeated shot is generated like any segment.

## Versions and regenerated storyboards

Regenerating an episode’s storyboard rebuilds its segments, and versions belong to segments; when the segments are replaced the old clips are no longer attached. Composite or download what you want to keep before regenerating a storyboard. Editing a single segment’s text keeps its versions. Re-parse an Episode: What You Lose and What You Keep.

## Sample projects

In sample projects visitors see one version per segment, the one chosen for the sample; the version control is not shown there. Your own projects always show every version to you.

| Action | Cost | Effect |
| Regenerate | Per second of the tier | New version, becomes active |
| Switch active version | Free | Preview, download, share and composite follow |
| Generate a callback segment | Free | The earlier segment's clip becomes this segment's active version |
| Delete the segment | Free | All versions removed |

## FAQ

### Is there a limit on versions?

No count limit. SceneMixer removes versions that are not active and are older than 90 days; the active one stays until the segment is deleted.

### Which version does the composite use?

The one active on each segment when you press composite.

### Can I download an older version?

Make it active, then download; or use the version control's own download where shown.

### Does switching versions change the share link?

The link stays; what it plays follows the active version.

### Do versions survive a storyboard regeneration?

No. Segments are rebuilt and old clips are detached. Composite first.

Updated November 3, 2026


---

# Set 30-Second Segments but Got Short Ones? Make a Segment Longer

URL: https://scenemixer.com/guides/set-30-seconds-but-segments-are-short/
Long take 30s is the most a segment can hold, not a length every segment is filled to. Why segments come out short, and how to lengthen one without re-parsing.

# You Chose Long Take 30s, but Some Segments Are Short: Why, and How to Lengthen One

Quick answer: Long take 30s sets the ceiling, not the length. The storyboard ends a segment where the scene turns, so a segment with two short shots may run only a few seconds while the next runs close to thirty. A segment’s length is the sum of its shots’ seconds. To lengthen one, edit it: give its shots more seconds, or add a shot or a line, up to the ceiling. Editing is free. Do not re-parse the episode to get longer segments: re-parsing deletes its storyboard and videos, charges the storyboard again, and does not promise longer segments.

Open a short segment, change one shot from 3 seconds to 6, and watch the total on the card change before you generate anything.

## Why a segment can be short

An episode is cut into segments by the story: a segment holds one stretch of the scene, and a new one starts where the action or the place changes. Standard 15s and Long take 30s, chosen at the top of step 3 before you parse an episode’s storyboard, set how long a segment may become, not how long it must be. Short dialogue exchanges and quick inserts make short segments; long scenes with speeches make long ones.

## How long a segment is

Each shot line starts with its number and its seconds, and the segment is the sum of them. The editor shows the total, turns it red when it goes over the ceiling, and updates the price on the card as you edit. Shot seconds and timing.

## Lengthening one segment

- More seconds on the shots that need them. A spoken line needs time to be said; a slow reveal needs time to land. Raise those, not every shot.
- One more shot. Add a reaction, an insert of a hand or an object, or a line of dialogue, with its own seconds.
- Keep to the ceiling. In a 30-second episode SceneMixer caps the total at thirty; leave a second or two of room.

Then generate that segment. Its old clip, if it had one, stays in Version History. Changing segment duration.

## Fewer, longer segments for a whole episode

If you want the whole episode in long takes, choose Long take 30s before you parse its storyboard, and generate it on a tier that can render thirty seconds; the model selector greys out the ones that cannot. Switching an episode that is already parsed means Re-parse, which replaces its storyboard and deletes its videos. 15-second or 30-second segments.

## Two short segments into one

There is no merge button. Copy the shots of the second segment into the first, adjust the seconds so the total stays under the ceiling, save, and delete the second segment. Insert or delete a segment.

| You see | Why | What to do |
| A 4-second segment in a 30-second episode | That stretch of the story is short | Add seconds or a shot in the editor |
| The total turned red | The shots add up to more than the ceiling | Take seconds off a shot that can spare them |
| A 30-second segment cannot be generated | The selected tier stops at fifteen | Pick a tier that renders thirty seconds |
| Whole episode too choppy | It was parsed in Standard 15s | Re-parse in Long take 30s (videos are deleted) |

## FAQ

### Will re-parsing the storyboard make segments longer?

Not reliably, and it deletes the episode's videos and SceneMixer charges for the storyboard again. Edit the short segment instead.

### Why does a shot with dialogue get more seconds?

A line needs time to be said. If you shorten that shot, the line may be rushed or cut.

### Is a longer segment more expensive?

Video is priced per second, so a longer segment costs more to generate. SceneMixer shows the price on the card before you press.

### Can I join two segments?

There is no merge button. Move the shots into one segment and delete the other.

Updated November 3, 2026


---

# Settings, Item by Item: Models, Badge, Language, Tour | SceneMixer

URL: https://scenemixer.com/guides/settings-item-by-item/
Every row in the Settings window and what it changes: the AIGC badge switch, the text, image and video model defaults and their prices, the language, the tour, Google.

# Settings, Item by Item

Quick answer: The ⚙ at the bottom of the sidebar opens Settings in two sections. Generation: Remove AIGC Watermark (paid plans; it decides whether future videos carry the badge), then Text Model, Image Model and Video Model — the default tier for each kind of generation, each option with its price and a Recommended tag; some say Available after upgrade. Preferences: Language for the interface, Quick tour → Replay to see the on-page tips again, Google Account → Link, and the Terms of Service. Nothing here costs credits by itself; the model rows change what the next generation costs.

The badge switch and which video tier are the two rows people ask about.

## Remove AIGC Watermark

A switch for paid plans. On: videos generated from now on carry no corner badge. Off: they do. It applies to videos you generate after changing it, not to existing ones. Below it, the note that you assume the compliance responsibility for badge-free videos. Free accounts see the row with Available after upgrade. The badge.

## Text Model

The tier used for reading scripts and writing storyboards. Options show their price per 1,000 characters; one carries Recommended. A higher tier costs more per read and per storyboard; the default is what the prices on the buttons assume. Some tiers are Available after upgrade.

## Image Model

The tier used for cast, set and prop sheets, with the price per image. The default is what the free allowance and the sheet prices are counted in. A higher tier changes the look of new sheets; existing ones stay. Regenerating a sheet.

## Video Model

The default tier for new segments, with the price per second; the tiers tagged First one free are the ones the free preview can use. This is a default: the model button in the editor still picks a tier per segment, and the price on Generate follows that choice. 480P, 720P or 1080P.

## Language

The interface language on SceneMixer is one of sixteen. It also sets the language a project’s text is shown to you in, and the language AI Create writes in. It does not change what the cast says in the videos; that is the dialogue language in step 1. Working in your language.

## Quick tour

Replay shows the on-page tips again on every page, as on your first visit.

## Google Account

Link connects a Google account to this one so either route signs you in. Shows Linked once done. Signing in.

| Row | Changes | Costs credits? |
| Remove AIGC Watermark | Whether future videos carry the badge (paid plans) | No |
| Text Model | Default tier for reads and storyboards | Changes their price |
| Image Model | Default tier for sheets | Changes their price |
| Video Model | Default tier for new segments | Changes their price |
| Language | Interface and shown text | No |
| Quick tour / Google Account | Tips again / a second sign-in route | No |

## FAQ

### Where is Settings?

The ⚙ at the bottom of the left sidebar.

### Does changing the video model change existing videos?

No. It sets the default for new segments; the editor can still pick per segment.

### Why does a model say Available after upgrade?

That tier is part of a SceneMixer paid plan.

### Does the language setting change what the cast says?

No. That is the dialogue language in step 1.

### Where is the deep-thinking switch?

There is none any more; the read and the storyboard use the tier you pick.

Updated November 3, 2026


---

# Share Links: Who Can See Them, Expiry, Play Limits, How to Revoke

URL: https://scenemixer.com/guides/share-links-who-can-see-expiry-limits/
A share link opens a public page that plays one clip or one episode. Who can open it, why it never expires by itself, the daily play limit, and the only way to revoke it.

# Share Links: Who Can See Them, Whether They Expire, the Daily Play Limit, and How to Take One Back

Quick answer: A share link opens a public page for one segment or one composite: anyone with the link can watch, without an account. The link does not expire on its own and cannot be revoked separately; it stops working when the segment, episode or project it points to is deleted. Each link has a daily play limit: past it, the page still opens but shows the poster instead of playing until the next day. The page is in the project’s display language and carries the AI badge that the clip carries.

Open the share menu on any finished segment, copy the link, and open it in a private window: no login is asked.

## What a share link is

Pressing Share on a finished segment or on a composite creates a link to a page on this site that plays that clip, with a title, the episode context and a way for viewers to start their own project. The page plays the light play version, not the original file; downloads are for you, from the editor. Share menu and platform buttons: Composite and Share an Episode.

## Who can see it

Anyone who has the link, logged in or not. Only the project’s owner can create links; a link cannot be made for someone else’s project. Search engines may index share pages, so treat a link as public once it exists. Sample projects can be shared by anyone who can open them.

## Expiry, and taking a link back

Links on SceneMixer have no expiry date and there is no revoke button: the link is tied to the clip, not stored as a separate object. It stops working the moment the segment, the episode or the whole project is deleted; single versions and single composites cannot be deleted. A segment link plays the segment’s active version, so regenerating the segment or switching versions changes what the link shows. To take a shared clip down, delete the segment — Segment Versions.

## The daily play limit

Each link can be played a set number of times per day, and the site as a whole has a daily budget for shared playback. Past either limit the page still loads and shows the poster frame with a note; playback resumes the next day. The limits exist because shared clips are served to people who are not paying for them; for larger audiences, download the file and publish it on your own channel.

## What viewers see

The clip, the episode and project titles in the project’s display language, and the AI badge if the clip carries one. Viewers cannot see your storyboard, your other segments or your account. The badge: The AI Badge Switch.

| Question | Answer |
| Who can open it | Anyone with the link |
| Does it expire | No |
| Can I revoke it | Only by deleting the segment, the episode or the project |
| Play limit | Per link per day, plus a site-wide daily budget |
| What plays | The play version, with the AI badge if the clip has one |

## Reporting a shared clip

Every share page carries a Report infringement link under the clip. It opens the site’s Copyright & Takedown page, /dmca/, which explains how a rights holder sends a notice and what the site does with it.

## FAQ

### Can I password-protect a share link?

No. Links are public; download the file if you need controlled distribution.

### Does sharing cost credits?

No.

### The page shows a poster and no video.

The link or the site has reached its daily play limit; it plays again the next day.

### Can I share a project, not just a clip?

One segment or one composite per link. Composite the episode and share that.

### Do share links work after I delete the project?

No. Deleting the project ends every link in it.

Updated November 3, 2026


---

# Share Your First AI Video, Get Free Credits: How the Reward Works

URL: https://scenemixer.com/guides/share-your-first-video-get-credits/
After your first finished video, SceneMixer offers credits for sharing it once. Which shares count, how long the offer lasts, who qualifies, and how to decline it.

# Share Your First Video and Get Credits: How the First-Video Reward Works

Quick answer: When your first real video clip is ready, a window offers 100 credits for sharing it. Post it to a social platform or send it through your phone’s share sheet, and the credits are added right away. It is one time per account and lasts 14 days from your first video. Chat apps and email do not count. Closing the window with × only postpones it; Don’t remind me again gives the reward up for good.

When the window appears, the video playing in it is the one you will share. Pick a social platform and post it; the credits land before you are back.

## When it appears

After the first video clip of your account is finished — a generated segment, not the free few-second preview and not a trim or an AI edit. The window can appear when that clip lands, and when you come back to the site, for 14 days after that first clip. It waits while a guided tour is on screen. For a one-click ad, it waits until the whole ad has been put together and offers the finished ad. If you have deleted that first clip, it offers your most recent one.

## Which shares count

- Social platforms in the window: X, Facebook, LinkedIn, Reddit, Pinterest, Threads, Tumblr and VK.
- Your phone’s share sheet, once you actually pick where to send it.

WhatsApp, Telegram, LINE and email are not in the reward window and do not count; trying them from elsewhere says That destination does not count for this reward. Share it on a social platform. Your chance is not used up by that.

## Getting the credits

Share once, and SceneMixer adds 100 credits to your balance on the spot; a Reward received window confirms it. In your credit history the row is called First video share reward. Nothing else is asked of you: no minimum views, no need to keep the post up. Reading your credit history.

If the share went out but the credits did not arrive, the window says so; press any share button once more. If the amount was changed while the window was open, it shows the new amount and asks you to share again to claim it.

## Closing it, or giving it up

- × or Esc — not now. It stays away for the rest of this visit and comes back next time, until the offer expires.
- Don’t remind me again — asks Stop this reminder?; confirming gives up the reward for this account, and it will not come back.
- Clicking outside the window does nothing, so it is not closed by accident.

## Who qualifies

One reward per person. Accounts whose first video was made before the reward started do not get the window. An account created on a device that has already claimed it, or re-registered with an email or phone number that already claimed it, does not qualify; the same network connection gets one reward per day. The site also hands out a limited number of rewards each day, so on a busy day the window may not appear until another day.

## What you share

The shared video carries the small SceneMixer badge in the corner, even if your own clip has none. Why shared videos carry the badge. The link also carries your referral code: people who sign up through it count toward the affiliate program. The affiliate program.

| Action | Counts? |
| Post to X, Facebook, LinkedIn, Reddit, Pinterest, Threads, Tumblr, VK | Yes |
| Phone share sheet, destination picked | Yes |
| WhatsApp, Telegram, LINE, email | No |
| × / Esc | Postpones the window |
| Don't remind me again | Gives up the reward |

## FAQ

### Is the free preview my first video?

No. The first generated segment clip counts; the preview, trims and edits do not.

### I closed the window. Is the reward gone?

No. It comes back on your next visit, within 14 days of your first video.

### Do I have to keep the post up?

No. The reward is for sharing once.

### Why didn't I get the window?

Your first video was made before the reward started, the offer expired, you already claimed or declined it, or another account on your device claimed it.

Updated November 3, 2026


---

# Sheet Description Blocked by the Text Check: Words and Rewrites

URL: https://scenemixer.com/guides/sheet-description-blocked-by-text-check/
Most blocked sheets are stopped by words in the description, not by an image. Which kinds of words trigger the check, how to rewrite them, why nothing was charged.

# “Text Check Failed”: When a Character or Scene Description Is Blocked Before the Image Is Made

Quick answer: When a sheet fails with a message about the text check, the provider refused the description before drawing anything — so nothing was billed and the credits are back. The trigger is a word or phrase in the description: explicit or gory terms, a real person’s name or likeness, a brand or a known character, a political figure, or a word that reads as one of those out of context. Rewrite the description in plain, neutral, visual terms and generate again; the check is a filter, not a judgement of your script.

Open the failed card, read the reason line, and edit the description right there before pressing regenerate.

## Where the block shows

In Step 2 the card turns Failed with a short reason; hover it to see the provider’s original message, which names the text check (or, less often, the image check). Nothing was generated and the credits come back: the provider rejects before it draws. Open the card, edit the description, regenerate. The wider rules and what repeated blocks mean: Content Rules and Moderation.

## The kinds of words that trigger it

Explicit and gory: nudity, sexual acts, wounds described in detail, blood as a feature. Real people: a celebrity’s name, “looks like” a named actor, a politician. Brands and IP: logos, product names, characters from known films, games or comics. Weapons and drugs named as the point of the image. Words that read wrong out of context: a “shot”, a “kill”, a “strip” of cloth, a child in a bath — the check reads words, not stories. SceneMixer checks Chinese and English alike.

## How to rewrite

Describe what the camera sees, in neutral words. A soldier “covered in blood after the massacre” becomes “dust on the uniform, a bandaged left hand”. “Looks like [actor]” becomes the face itself: jaw, brow, eyes, hair. A brand becomes the object: “a red energy drink can, no label”. A named franchise character becomes an original design with its own anchor feature. Keep the identifying features — scars, glasses, a hairstyle — and drop the adjectives that judge. The same discipline makes better sheets anyway: Write a Description That Works.

## When the image, not the text, is blocked

Sometimes the description passes and the generated image fails the check — the message names the image. The cause is still usually the description: it asked for something the model drew too literally. Soften the same words; if it was a photo you uploaded, the photo itself may contain a real person or a logo. Use Your Own Photo.

## IP suspected

A separate check flags likeness to known franchises and brands. It is triggered by names, by iconic outfits and by realistic styles that land on a famous face. Replace the reference with an original design and change one anchor feature; switching to a stylised visual style often clears it. Blocked by Moderation: What to Change.

| Blocked wording | Neutral rewrite |
| covered in blood after the fight | dust on the jacket, a bandaged left hand |
| looks exactly like [famous actor] | square jaw, heavy brow, short grey hair, deep-set eyes |
| holding a [brand] phone | holding a black smartphone, no logo |
| a [franchise] warrior in the iconic armour | a warrior in dark segmented armour with a single red shoulder plate |
| a seductive nurse | a nurse in a plain white uniform, hair tied back |

## FAQ

### Was I charged for the blocked sheet?

No. The provider refused before generating; SceneMixer returned the hold to your balance.

### Nothing in my description is offensive. Why was it blocked?

The check reads words, not intent. A single word that also has a violent or sexual meaning can trigger it; rewrite that word.

### Does a block count against my account?

Not on its own. Repeated blocks in a short time make SceneMixer pause video generation; see the content rules page.

### Can I ask support to allow my description?

No; the check runs at the provider. Rewriting is the only way through.

### Is the check stricter for scenes than for characters?

Same rules. Scenes trip it with gore, drugs and real locations described as crime scenes.

Updated November 3, 2026


---

# Short Drama Ideas: 36 Premises With Episode-One Hooks

URL: https://scenemixer.com/guides/short-drama-ideas/
36 short drama ideas with the line episode one ends on, 20 cliffhanger endings, and how to turn a premise into a one-line brief for a 60 to 120 second episode.

# Short drama ideas: 36 premises, hooks and cliffhangers

Quick answer: A short drama idea is a premise plus an engine: a person with a want, someone in the way and a reason to watch the next 90 seconds. Below are 36 ideas in six genres, each with the line episode one ends on, plus 20 cliffhanger endings you can bolt onto any of them. Paste a premise into AI Create as a one-line brief and SceneMixer drafts the episodes. A Short Drama episode runs up to 240 seconds; write for 60 to 120 so the hook lands early, and close every episode on an unanswered question.

Pick the premise you can say in one breath, copy its episode-one ending, and write the first five seconds as the strangest moment of the episode.

## How a short drama idea is built

| Part | Job | Example (the heiress auditor) |
| The hook | The strangest image or line, in the first five seconds | A badge placed on a desk with her brother's name on it |
| The want | What the lead needs, said or shown | To find who forged her signature |
| The obstacle | Who or what is in the way | Her brother, her father's lawyer, a week |
| The engine | What repeats every episode | One more document, one more lie |
| The ending question | What the last shot asks | Who else knew? |
Hooks and cliffhangers · AI hook generator · Micro drama, short drama, vertical drama.

## Revenge and return

| Premise | Episode 1 ends on | Set it in |
| A disowned daughter returns as the auditor of her family's company, with one week to find who forged her signature. | The auditor's badge is set on her brother's desk. | Urban Opulence |
| A fired chef opens a food truck outside the restaurant that let him go, parked in the owner's own space. | The owner's car does not arrive; his wife does. | Street Life |
| A woman wrongly sent to prison comes out and takes a job as the warden's secretary. | She finds her own case file in the bottom drawer. | Corporate Cold |
| A bridesmaid has the video from the night the groom swore he was somewhere else. | The projector at the reception warms up. | Suburban Mansion |
| A star pupil expelled for cheating returns as the substitute teacher of the class that framed her. | The roll call reaches a name she knows. | Elite Boarding School |
| An heir who died at sea walks into his own memorial and sits in the back row. | Someone in the front row turns around and smiles. | Island Resort |

## Secret identity

| Premise | Episode 1 ends on | Set it in |
| A delivery rider is the missing owner of the tower she delivers to. | The lift takes her to the penthouse, and the key in her pocket fits. | Neon Nights |
| The new cleaner at a failing company is the founder who sold it, and she wants to know who ruined it. | A guard on the camera feed recognises her face. | Corporate Cold |
| A quiet student is the writer of the anonymous blog the whole school is trying to unmask. | The principal reads out the first line of tomorrow's post. | Campus Youth |
| A bodyguard hired to protect a pop star is the star's long-lost sister. | The star hums the lullaby their mother made up. | Retro Hong Kong |
| A village doctor is a retired spy. | A stranger asks for her by her old code name. | Pastoral Healing |
| A waiter at a gala is the only one who knows the bride's real name. | He pours the groom a glass and whispers it. | Classical Elegance |

## Romance with a catch

| Premise | Episode 1 ends on | Set it in |
| Two rivals pretend to date for one family-wedding week. | The photographer catches them actually holding hands. | Soft Romance |
| A shopkeeper signs a one-year contract marriage with her landlord's son to save her shop. | There is a clause she did not read. | Suburban Mansion |
| A translator falls for the man whose letters she has been quietly editing. | He walks into the office holding the last letter. | Corporate Cold |
| A single dad and a night-school teacher share the 6 a.m. bus and a secret. | His daughter says: that's her. | American Small Town |
| The groom's ex is the wedding planner. | She hands the note to the groom, not the bride. | Mediterranean Coast |
| Two strangers keep missing each other by one minute across the city. | Their trains stop at the same platform, on opposite tracks. | Neon Nights |

## Family and inheritance

| Premise | Episode 1 ends on | Set it in |
| Three cousins inherit a vineyard that goes to whoever stays the longest. | The eldest is found packing at midnight. | Mediterranean Coast |
| A grandmother announces at dinner that one of the five at the table is not her grandchild. | She names the child who is. | English Manor |
| A family restaurant's recipe book has a page torn out. | The page turns up in the rival chef's pocket. | Street Life |
| A son returns for his father's funeral and finds a second family at the grave. | A girl his age calls him brother. | American Small Town |
| An estate lawyer reads a will that leaves everything to the one who laughed. | Everyone turns to look at the youngest. | European Palace |
| A mother and daughter swap lives for a week to settle an argument. | The daughter finds the mother's old suitcase under the bed. | Nordic Minimal |

## Work and ambition

| Premise | Episode 1 ends on | Set it in |
| A junior architect's design is presented by her boss; she has the original files and one meeting to say so. | The client asks: who drew this? | Corporate Cold |
| A night nurse notices a pattern in the charts that no one else will read. | The chart turns to a name she knows. | Hospital Ward |
| A chef has one table to impress the critic, and the critic is her mother. | The critic takes off her hat. | Urban Opulence |
| Rival start-ups share a co-working space and a Wi-Fi password. | The router goes dark in the middle of a demo. | Cyber Metropolis |
| A rookie referee must make a call in the final minute against the team her brother plays for. | The whistle is in her mouth; her brother looks up. | Street Life |
| A radio host's last show gets a caller who knows the station's secret. | The caller says the host's real name. | Neon Nights |

## Magic, rebirth and time

| Premise | Episode 1 ends on | Set it in |
| A woman wakes the day before her wedding with a year of memories she has not lived yet, and has 24 hours to change one thing. | The alarm that rings is not hers. | Soft Romance |
| A shop sells second chances by the hour, and the clerk has run out of hours. | The bell over the door rings. | Pastoral Healing |
| A student finds her late mother's school diary, and it writes back. | A new line appears: don't go to the lake. | Magic Academy |
| A witch's apprentice is sent to collect a debt from a king who has forgotten he owes it. | The king pays with the apprentice's own name. | Medieval Castle |
| Every midnight the city repeats the same hour, and a baker is the only one who notices. | The loaf she baked yesterday is still warm. | Parisian Streets |
| A boy finds that his shadow has begun doing things he did not. | The shadow waves from across the street. | Gothic Dark |
How to make each genre · A revenge drama · A contract marriage drama · A rebirth revenge drama · A time-travel romance.

## Twenty cliffhanger endings

Generated video does not write legible text, so a message on a screen is read aloud or shown by the character's face. Pick the type, then write the last shot.

| Type | The last shot to write |
| A name said out loud | Close-up on a mouth; a name nobody expected, then black |
| A door opens | Wide shot of the door and the person in it, cut before they speak |
| The wrong person arrives | Hold on a face that should not be there |
| A phone buzzes | The character reads the message; her face changes; cut |
| A photograph turns over | A hand turns it; stay on the character's face, not the photo |
| A confession cut mid-sentence | I should have told you that I… and black |
| A key fits | A hand, a lock, a click |
| A second set of footsteps | Ambient only; the character stops walking |
| A reflection that does not match | The mirror moves a beat late |
| A countdown starts | A voice: you have until midnight |
| A contract is signed | A pen, a signature stroke, a smile that does not reach the eyes |
| The truth is overheard | A figure behind a door, listening |
| Two people see each other | Across a crowded room, both freeze |
| An object is missing | The empty place on the shelf, then the character's reaction |
| A lie is confirmed | It was true, said quietly, and black |
| A hand on a wrist | A rescue, or a capture; hold one beat |
| A new character is named | Meet my sister, and she is not who the lead expects |
| An offer with a price | I can fix this. You only have to… and cut |
| The call is from inside | The lead looks up at the ceiling |
| A sentence from episode 1, repeated | The same line, now spoken by the villain |

## From a premise to a brief and a 90-second episode

Turn a premise into a brief by adding the lead's want, the obstacle, the ending and the tone, then paste it into SceneMixer's AI Create. For a revenge premise:

A revenge drama: a disowned heiress returns as the family firm's auditor with one week to prove who forged her signature; her brother tries to bury it; she wins the company and forgives no one; cold and fast, every episode ends on a document.

| Part of a 90-second episode | Seconds | Does |
| Hook | 0 to 5 | The strangest image: the badge on the desk |
| Setup | 5 to 25 | Who she is, what she wants, how long she has |
| Turn | 25 to 60 | Someone blocks her; she finds the first document |
| Escalation | 60 to 80 | The stakes rise: her brother knows |
| Cliffhanger | 80 to 90 | The ending question, in one image or line |
Each segment is up to 15 or 30 seconds in one place. Vertical short drama script format · What a micro-drama costs · AI micro drama · AI short drama generator · Make vertical drama-app dramas.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Corporate Cold; Costume Style Power Tailoring, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the heiress, her brother and the family firm's boardroom; the documents are props. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Genre | The engine that repeats | Ends each episode on |
| Revenge and return | One more proof, one more lie | A document, a name or a face |
| Secret identity | Someone almost finds out | A near miss |
| Romance with a catch | The rule of the pretence | A moment that breaks the rule |
| Family and inheritance | A secret in the will | A new heir or a new fact |
| Work and ambition | A meeting, a deadline | A decision the lead did not expect |
| Magic, rebirth and time | The rule of the magic | A cost the lead did not know about |

## FAQ

### What are good short drama ideas?

A person with one want, someone in the way, and an engine that repeats: revenge and return, secret identity, a romance with a catch, an inheritance, a rivalry at work, or a rule of magic.

### How do I come up with a micro drama idea?

Pick a genre, write the lead's want and the obstacle in one line, and write the last shot of episode one first. This page gives 36 premises with their episode-one endings.

### What is a cliffhanger?

An ending that leaves a question open so the viewer wants the next episode. Twenty types are listed above.

### How long should a short drama episode be?

SceneMixer's Short Drama type allows up to 240 seconds; most vertical episodes work at 60 to 120.

### How do I turn an idea into an episode?

Add the lead's want, the obstacle, the ending and the tone, and paste it into AI Create. It writes the episodes' script; reading it gives you the cast and the storyboard.

### Can I use these premises?

They are starting points. Change the person, the place and the ending until they are yours.

### Which format suits a short drama?

9:16 Portrait for phones, with the Short Drama type.

### How many episodes should a short drama have?

Start with the first three. Make episode 1, read the storyboard, then continue. SceneMixer charges only for the episodes you render.

Updated November 3, 2026


---

# Short Drama, TV Series or Film: What the Type Changes | SceneMixer

URL: https://scenemixer.com/guides/short-drama-series-or-film/
The Type setting decides how the script is cut into episodes, how long an episode may run, and how scenes are paced. What each of the three does, and how to pick.

# Short Drama, TV Series or Film

Quick answer: In step 1, under Type, choose Short Drama, TV Series or Film. Short Drama is fast-paced, dense dialogue, hook-driven, with episodes up to 240 seconds. TV Series is series pacing, dialogue-driven, moderate build-up, episodes up to 300 seconds. Film is slow-paced and atmospheric with no episode cap. The type shapes how the script is split into episodes and how each scene is written for the camera, so choose it before you continue; after the read it cannot be changed in that project.

Open a finished short-drama episode and notice the pace: a hook, dense dialogue, a cliffhanger, all inside a few minutes.

## Where it is

Step 1, Set Up Your Project, the Type row. Three cards, each with a one-line description. The hint under the row says what it governs: episode-splitting rhythm, the per-episode length cap and acting grammar.

## Short Drama

The vertical-drama form: every episode opens on a hook, runs dense dialogue, and ends on a nail — an identity revealed, a feeling left hanging, a reversal. Episodes cap at 240 seconds. The script is split so that each episode has its own conflict and its own high point. Pick it for anything that will be watched in a feed, one episode at a time. ReelShort-style dramas.

## TV Series

Longer episodes, up to 300 seconds, with room for a scene to build. Dialogue still drives it, but the cut is less breathless and the episode ends do not all have to be cliffhangers. Pick it for a story with several threads or an ensemble.

## Film

Slow-paced, atmospheric, room to breathe, and no cap on episode length. The script is treated as one piece divided into acts rather than as a run of episodes. Pick it for a short film or a story that does not want to be chopped into hooks. What comes out.

## What it does to your script

The type is read when the script is parsed. It decides where the episode breaks fall, how long each episode may be, and the way scenes are written into shots — how much a look carries versus a line, how quickly the camera moves on. Everything downstream (storyboards, videos) inherits that, which is why the type is fixed once the script is read: the episodes were cut for it. For another type, start a new project with the same script.

## Ads and remakes

Ads and remakes have their own type set for them; the row is not offered. An ad is one short piece; a remake follows the structure of the video it was read from.

| Type | Pace | Episode cap | Pick for |
| Short Drama | Fast, dense dialogue, hook-driven | 240 s | Vertical drama, feeds, one episode at a time |
| TV Series | Series pacing, dialogue-driven | 300 s | Ensembles, several threads |
| Film | Slow, atmospheric | None | Short films, stories that should not be chopped |

## FAQ

### Which type should a first project use?

Short Drama, unless you have a reason not to. It is the form the site is built around and the quickest to see results in.

### Can I change the type after parsing?

Not in that project: the episodes were already cut for the type you chose. Start a new project with the same script and pick another type before it is read.

### Does the type change the price?

No. SceneMixer prices parsing, storyboards and video the same way whatever the type.

### What is the episode cap?

On SceneMixer, the longest an episode may run: 240 seconds for Short Drama, 300 for TV Series, none for Film.

### Does Film mean one long video?

It means no episode cap and a slower cut. You still generate segment by segment and composite at the end.

### I only want one complete video, not episodes.

Keep the script short enough to be one episode, or use AI Create with a short length. If the reading still made several episodes, composite each one once all its segments are done, then join the files in any video editor.

Updated November 3, 2026


---

# Shot List Template for AI Video: Columns, Examples and a Free Draft

URL: https://scenemixer.com/guides/shot-list-template/
A shot list template for short films and AI video: the columns that matter, an example scene, how it differs from a storyboard, and how to draft one free.

# Shot list template for AI video

Quick answer: A shot list is the director's checklist: one row per shot with its number, scene, framing, camera move, action, lines and length. For AI video, add who is in the shot and where, because position has to carry across cuts. SceneMixer writes a shot list for you as the storyboard, shot by shot, and the text is editable, so the template below is also the format you edit.

Copy the columns into a sheet, fill the first scene, then paste a script into SceneMixer and compare its drafted shots with yours.

## Columns that earn their place

| Column | What to write | Why it matters |
| Shot | Number | Order and references |
| Scene / place | Where it happens | One place per segment keeps the set steady |
| Framing | Wide, medium, close-up | Tells the picture what to show |
| Camera | Still, push-in, track, pan | Tells it how to move |
| Who and where | Characters and positions | Carries across cuts |
| Action | One thing | Keeps the shot clean |
| Line | Speaker and words | Spoken in the video |
| Seconds | Length | Adds up to cost |

## An example: two people, one door

| Shot | Framing | Camera | Action and line | Sec |
| 1 | Wide | Still | A corridor; a door at the far end | 3 |
| 2 | Medium | Slow push-in | @Mara stops at the door, hand on the handle | 4 |
| 3 | Close-up | Still | @Mara breathes out. (quiet): I'm back. | 3 |
| 4 | Over the shoulder | Still | @Jon opens the door from inside | 3 |

That is 13 seconds of video, four rows. How to write one shot.

## Shot list vs storyboard

A shot list is words; a storyboard is the same shots as pictures. In SceneMixer the storyboard is a numbered shot list in text, and each segment's shots are rendered into video, so you edit the list and the video follows. The storyboard template · AI shot list generator · AI storyboard generator.

## Draft one for free

Paste a script or one line, read the storyboard it writes and edit it. The first episode's storyboard on your first drama is free, and editing is free. Generate the storyboard for an episode.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Neon Nights; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw Mara and Jon and the one corridor; the door is part of the place. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Shot | Framing | Seconds |
| 1 | Wide | 3 |
| 2 | Medium | 4 |
| 3 | Close-up | 3 |
| 4 | Over the shoulder | 3 |

## FAQ

### What is a shot list?

A row-per-shot checklist with the number, place, framing, camera move, action, lines and length.

### What columns should it have?

Shot, place, framing, camera, who and where, action, line and seconds.

### How is it different from a storyboard?

A shot list is text; a storyboard shows the shots as pictures. In SceneMixer the storyboard is a text shot list that renders into video.

### Can SceneMixer write one?

Yes. Paste a script or one line and the shots are drafted; you edit them.

### Can I paste my own?

Yes, in the numbered shot format.

### Is it free?

Editing on SceneMixer is free; the first storyboard on your first drama is free.

### How long is each shot?

Usually 3 to 5 seconds.

Updated November 3, 2026


---

# Shot Seconds and Timing in an AI Storyboard | SceneMixer

URL: https://scenemixer.com/guides/shot-seconds-and-timing/
Every shot carries a duration and the shots add up to the segment. How to change a shot's seconds, the cap, why dialogue needs its seconds, and what 'over cap' means.

# Shot Seconds: Changing Timing, and the Limits

Quick answer: Every shot marker carries its length: [Shot 2] 3s. To change it, put the cursor on the shot, type @, open the Time tab and pick a duration — @3s is a three-second shot. The shots add up to the length of the segment, and the segment is what you pay for by the second. A segment cannot be longer than the cap of the video tier you selected; go over and the editor says so before you generate. Dialogue needs time: at roughly three words a second in English, a twelve-word line needs about four seconds of shot.

Open a finished episode and read the seconds on the shot markers of one segment. Add them up; compare with the card on the timeline.

## Where the seconds are

Each shot starts with a marker chip: [Shot 1] 4s, [Shot 2] 3s. The number is that shot’s length in seconds. The card for the segment on the bottom timeline shows the total.

## Changing a shot’s seconds

Edit the segment, place the cursor at the shot, type @ and choose the Time tab. Pick the duration you want; the picker inserts it at the cursor (@3s = a 3-second shot). Delete the old duration if it is still there, then save. Nothing is charged for the edit. The @ menu.

Making a shot longer does not stretch the picture; it gives the action more time to happen. Write what happens in the extra seconds, or you get a person standing still.

## Shots add up to the segment

The segment’s length is the sum of its shots. Add two seconds to one shot and the segment is two seconds longer; the price on the generate button follows, because video is priced by the second. Cut a shot and the segment shortens.

There is a cap. Each video tier on SceneMixer has a maximum segment length — some tiers allow longer segments than others, and the limit for each is shown where you pick the tier. When your shots add up past it, the segment card says over cap, and generating is refused with a message asking you to trim the segment or pick a tier that allows longer clips. Choosing a tier.

Rather than squeezing, split: Insert segment after makes a new, empty segment; move the last shots into it. Inserting and deleting segments.

## Dialogue needs its seconds

A line takes as long to say as it takes to say. In English, count on about three words a second; a twelve-word line needs a four-second shot, and a shot that also has to show her crossing the room needs more. Give a line too few seconds and it is rushed or cut off. Give a reaction shot with no dialogue two or three seconds; longer than that, the person just waits.

## A few shapes that work

A shot is one thing happening. Two or three seconds for a look, a turn, a hand on a door. Four to six for a line and the reaction to it. Longer for a walk across the room, a fight beat, a slow reveal. Most segments are three to five shots. How to write one shot.

| Shot | Typical seconds |
| A glance, a turn, a small gesture | 2–3 |
| One spoken line and its reaction | 3–5 |
| A longer line, or a line plus movement | 5–8 |
| Crossing a room, an action beat | 6–10 |
| A whole segment | The sum of the above, within the tier’s cap |

## FAQ

### How do I make a shot longer?

Type @ at the shot in SceneMixer, open the Time tab and pick a longer duration; remove the old one; save. Then write what happens in the added time.

### What does 'over cap' mean on the segment card?

The shots add up to more than the selected video tier allows for one segment. Shorten a shot, split the segment, or pick a tier with a higher cap.

### Does changing seconds change the price?

Yes. SceneMixer prices video by the second, so the generate button's price follows the segment's total.

### How many seconds does a line of dialogue need?

About three words a second in English. Add time for anything else the shot has to show.

### Can a shot be one second?

It can, but very short shots tend to read as a flicker. Two to three seconds is the usual floor for a beat.

Updated November 3, 2026


---

# Shot-to-Shot Continuity in AI Video, Kept Automatically | SceneMixer

URL: https://scenemixer.com/guides/shot-to-shot-continuity/
Posture, what someone holds, where they stand: each shot carries on from the last and each segment opens where the previous one ended. How to write for it.

# Continuity From One Shot to the Next

Quick answer: Continuity is the small stuff: she was holding the cup, so she is still holding it; he was kneeling, so the next shot does not start with him standing; it was dusk, so it is still dusk. SceneMixer keeps those facts from one shot to the next and from one segment to the next — the opening frame of a segment is written from the end of the previous one. Together with the automatic AI 3D staging of each scene, that is what makes an episode play as one continuous scene instead of a row of clips.

Newer storyboards. In storyboards written or regenerated since late October 2026 a segment does not start from a written opening frame. It starts from a running record the site keeps while it reads your text: where people are, posture, what is held, wounds, wet clothes, time and weather. What the system keeps track of between segments · Why the next segment opens like the last one ended.

Open a finished episode and play two consecutive segments. Look at the join.

## What continuity means here

Not consistency — that is the same face and the same room, and it holds across the whole series. Continuity is the moment-to-moment agreement between one shot and the shot right after it: the same posture, the same object in the same hand, the same side of the room, the same light, the same weather.

It is the thing that breaks most easily when video is generated one piece at a time, and the thing viewers notice fastest.

## What holds from shot to shot

Posture. Sitting, kneeling, lying, leaning on the table. A shot that starts on someone who was crouching starts with them crouching.

What is in hand. The letter she picked up in shot 2 is in her hand in shot 3, unless the storyboard says she put it down.

Position. Where each person is, and who they face. Held by the automatic AI 3D staging of the scene. More on that.

Time and weather. Set once in the segment header, applied to every shot in it.

## What holds from segment to segment

A segment is one generation, a few shots long. The next segment is another generation — and its storyboard is written to follow on from where the one before ended: who is in the picture, how they are standing, what has just happened. You can read it: every segment’s storyboard starts with an Opening line describing that frame. Edit it if you want the join to be different.

When you regenerate a middle segment, its neighbours are not touched. The opening line still describes the join, so the new version still fits.

## How to write for it

Say what changes, and let what does not change stay. “She sets the cup down” is a change; the next shot will not have the cup in her hand. If you never mention the cup again, she keeps it.

Say who is in the shot. A named character is carried forward; an unnamed “someone” is not.

Keep the segment header honest. If a segment is later that day, change the Time field; the light will follow. How to write one shot.

## When a join is wrong

Open the second segment and read its Opening line. If it describes the wrong state — he is standing when he should be on the floor — fix the line and regenerate that segment. Only that one. Every version is kept, so you can compare. Fixing one shot.

| Fact | Carried by | You control it with |
| Posture and what is in hand | The shot text, shot to shot | Say when it changes |
| Where people are and which way they face | The scene’s automatic AI 3D staging | Describe the room once |
| Time of day, weather | The segment header | Time and Weather fields |
| The join between two segments | The next segment’s Opening line | Edit that line |

## FAQ

### Does the next segment know how the last one ended?

Yes. SceneMixer writes its Opening line from the end of the previous segment, and you can read and edit it.

### If I regenerate one segment, do the ones around it change?

No. Only the segment you regenerate is remade. The joins are described in the storyboard, so the new take still fits.

### Why did an object disappear between shots?

Usually because a shot mentioned it and the next did not, in a way that read as it being gone. Name the object again, or say what happened to it.

### How do I make time pass between segments?

Change the Time field in the later segment's header.

### Does this cost extra?

No. It is part of how segments are generated.

Updated November 3, 2026


---

# Some Shots Have No Linked Characters: What It Means and How to Fix

URL: https://scenemixer.com/guides/shots-with-no-linked-characters-what-it-means/
Before generating, the editor asks when a shot names a character as plain text, not an @ pill. What that changes, how to link the name, and when to generate anyway.

# “Some Shots Have No Linked Characters”: What the Question Means and How to Fix It

Quick answer: When you press Generate Video on a segment you wrote or edited yourself, the editor may ask: Some shots in segment N have no linked characters. These names are only plain text there… Generate anyway? It means a shot names a character by typing, not with an @ pill, so that character’s sheet will not be used in that shot and the face may come out different. Cancel, edit the segment, replace the typed name with an @ pill, save and generate. If the person is not actually on screen in that shot, generate anyway.

Open a segment you edited, press Edit, and look at the names in each shot: pills are linked, plain words are not.

## When the question appears

Only on segments you wrote or changed yourself: one you inserted, edited, or pasted text into. Segments exactly as the storyboard produced them are not checked. It appears on the website, on a computer or in a phone browser. Nothing is charged until you confirm.

## What a linked name is

A name inserted with the @ menu becomes a coloured pill. A pill tells the shot which character you mean, and that character’s sheet goes into the generation. A name you typed is ordinary text: the video model reads the word, but it has no picture to go with it, so it draws someone who fits the words. The question appears when a shot has no linked character at all, yet its text names one. The @ menu · Who is in frame.

## How to fix it

- Press Cancel.
- Open the segment with Edit and find the shots the question named.
- Delete the typed name, type @ in its place and pick the character. A pill appears.
- Save, then generate.

If the character is not in the @ menu, it is not in this episode yet: add it with Add assets in the editor’s left panel. Add a character to this episode.

## When to generate anyway

When the named person is not in the picture in that shot: someone talked about, heard on the phone, or a shot of a letter or a doorway. Linking a person who is not on screen invites their face into the frame, so leaving the name as text is right there. Directions instead of naming who is off screen.

## Generate all and the same question

Generate all asks the same question for each segment that has such shots. Confirm and that segment is submitted; cancel and only that one is skipped, with Segment N was skipped: some shots have no linked characters. The others carry on. Generate all segments at once.

## Earlier warnings, and pasted text

You may see it coming: after a save, an amber note above the storyboard can say a name was written without @. When you paste text and save, a message saying Names in the text were linked to their characters, scenes and items means the pills are back; if it does not appear, or the question still comes up, link the names with @ yourself. The amber notes after you save · Copying and pasting storyboard text.

| Shot text | Linked? | In the video |
| @Mara pill turns to the window | Yes | Mara's sheet is used; her face holds |
| Mara turns to the window (typed) | No | A woman who fits the words; may not look like Mara |
| She reads Mara's letter (Mara not on screen) | No, and that is right | Generate anyway |

## FAQ

### I already generated, and the person looks wrong. What now?

Link the name with @, save, and regenerate that segment. The earlier clip was delivered, so it was charged; if you think something went wrong on our side, send feedback with the segment number.

### Why was a segment I never touched not checked?

Segments as the storyboard produced them already have their characters linked. The check is for text you wrote or changed.

### How do I use @ on a phone?

Open the segment's editor on the phone and type @ in the text; the same menu opens.

### Does the question cost anything?

No. SceneMixer charges nothing until you confirm a generation.

Updated November 3, 2026


---

# Sign Up and Log In: Google, Email Code, No Password | SceneMixer

URL: https://scenemixer.com/guides/sign-up-and-log-in/
Create an account or sign in with Google or an email code, why there is no password, what the pop-up sign-in window is, and how to log out or switch accounts.

# Signing Up and Logging In

Quick answer: There is no password. Press Sign In / Register at the top right, or any button that needs an account, and the sign-in window opens. Continue with Google or Continue with Apple signs you in with one tap. Or Continue with email: type the address, Send code, type the 6-digit code from the email, Continue. A new address gets an account on the spot; an existing one is signed in. By continuing you agree to the Terms of Service and Privacy Policy. Every new account starts with 50 credits and the free first-drama allowance.

What you get for free is what the new account already has.

## Google

Continue with Google opens Google’s own window; pick the account and you are back, signed in. The first time creates your SceneMixer account with that address. If Google reports the address is already registered with a password from the old sign-up, use the email route instead; Google is not linked to an existing account by itself, but you can link it later in Settings.

## Apple

Continue with Apple (Sign in with Apple on the sign-in page) sits under the Google button. It opens Apple’s own sign-in and brings you back signed in; the first time creates your SceneMixer account. Apple sign-in does not link itself to an account that already exists: if the address is already registered another way, the window says so, and you sign in the way you signed up. If the Apple window is cancelled or has expired, press the button again; if Apple sign-in is unavailable, use the email code.

## Email code

Continue with email. Type the address, Send code. The message says We sent a code to … check your inbox (and the spam folder). Type the six digits and Continue. Resend is available after a short countdown; Change email goes back. New addresses get a SceneMixer account automatically; there is no separate registration form and nothing to remember.

## The window that appears mid-task

Pressing a button that needs an account — Get Started, generate, buy — opens the same sign-in window over the page with Sign in to start creating. What you typed in the script box stays there; sign in and press the button again. The pages at /login and /register are the same form on a full page.

## Link Google later

Settings → Google Account → Link. From then on either route reaches the same account. A Google account already linked elsewhere cannot be linked twice; the message says so.

## Log out, and where you land next time

Click your avatar at the top right, then Log out. Next time, sign in the same way and SceneMixer keeps your projects where you left them — each card reopens at the step you were on. Project cards.

## If the code does not arrive

Check spam; wait for the countdown and Resend; check the address for a typo with Change email. Some company mail servers hold automated mail for minutes. If nothing arrives, sign in with Google, or write to support from the address you tried.

| Route | Steps | Creates an account? |
| Google | Continue with Google → pick account | Yes, first time |
| Apple | Continue with Apple → Apple’s window | Yes, first time; an address already registered another way is refused |
| Email code | Address → Send code → 6 digits → Continue | Yes, for a new address |
| Link Google | Settings → Google Account → Link | No, adds a route |
| Log out | Avatar → Log out | — |

## FAQ

### Where is the password?

SceneMixer has no password. Google, Apple, or a code sent to your email each time.

### Do I register first?

No. SceneMixer gives a new email address an account the moment its code is confirmed; Google does the same.

### Can I use both Google and email?

Yes, once Google is linked in Settings. Both reach the same SceneMixer account.

### The sign-in window opened while I was typing.

A button needed an account. Your text is kept; sign in and press it again.

### What does a new account come with?

A new SceneMixer account comes with the welcome credits and the free first-drama allowance shown on the free-tier page.

### How do I switch to another account?

Click your avatar at the top right, then Log out, and sign in with the other Google account or email address. Projects stay with the SceneMixer account that made them.

Updated November 3, 2026


---

# Sizzle Reel With AI: Make a Proof-of-Concept Reel for a Show

URL: https://scenemixer.com/guides/sizzle-reel-with-ai/
Make a sizzle reel with AI: a 60-second tone piece for a show or film, a 13-shot storyboard, a sound plan, and how it differs from a trailer or a pilot.

# How to make a sizzle reel with AI

Quick answer: A sizzle reel is a 60 to 90 second piece that sells the feel of a show, its world, its look and its energy, rather than its plot. In SceneMixer you draw the cast and three or four places once, write 12 to 18 short shots that rise in pace, let the built-in score carry the mood and end on a narrator saying the title. Use it for a pitch, a crowdfunding page or a proof of concept. It is previs-grade: say so when you share it.

Write four places and three people on one page, then write one shot of each place. If you can see the four places in your head, you have a sizzle reel.

## Sizzle reel, trailer, proof of concept, pilot

| Piece | Length | What it proves | Has plot? |
| Sizzle reel | 60 to 90 seconds | The tone, look and energy of a world | No: moods and moments |
| Pitch trailer | About 60 seconds | The hook and the stakes of a story | Yes: the promise, not the ending |
| Proof-of-concept short | 2 to 5 minutes | That one scene works from start to finish | One complete scene |
| Pilot | Up to 300 seconds (TV Series) | That the first episode works | Yes: a full episode |
A pilot or pitch trailer · A movie trailer · A pitch deck for a TV show · What is previs?.

## The shape of a 60-second sizzle

A sizzle has a rhythm, not a story. Each segment is one place, so each beat gets its own set:

| Segment | Beat | Seconds | Shots | Pace |
| 1 | The world | 0 to 15 | 3 | Slow, wide, one quiet close-up |
| 2 | The people | 15 to 30 | 3 | Medium, moving, a crowd |
| 3 | The tension | 30 to 45 | 4 | Closer, shorter, two faces and a line |
| 4 | The peak and the title | 45 to 60 | 3 | Fastest cuts, then a held final image and the title |

## Thirteen shots, 60 seconds

An original example, a canal-city noir called Neon Tide. One place per segment; the title is spoken by the narrator because generated video does not write legible text.

| Segment | Shot |
| 1. The harbour | [Shot 1] 6s: Extreme wide shot, slow push-in. A flooded neon harbour at dawn, cranes standing in the water, ferries gliding between towers. Ambient: low surf and a distant horn. |
| 1. The harbour | [Shot 2] 5s: Medium shot. @Ana, a river pilot in a yellow coat, steps onto the deck of a small boat, rain beading on her hood. No dialogue. |
| 1. The harbour | [Shot 3] 4s: Close-up on @Ana's eyes in a flash of orange sign-light. Ambient: a heartbeat under the surf. |
| 2. The market | [Shot 4] 5s: Wide shot, tracking. A night market on stilts, lanterns and steam, a crowd moving both ways. Ambient: market chatter. |
| 2. The market | [Shot 5] 5s: Medium shot. @Kade, a smuggler with a silver ear cuff, pushes a crate through the crowd without looking back. |
| 2. The market | [Shot 6] 5s: Close-up. A hand passes a folded paper to another hand under a table. Ambient: a drum, close by. |
| 3. The rooftop | [Shot 7] 4s: Wide shot. A rooftop club at night, a skyline of lit towers, a pulsing blue light. Ambient: bass, building. |
| 3. The rooftop | [Shot 8] 4s: Medium shot. @Ana and @Kade face each other across a table, neither drinking. |
| 3. The rooftop | [Shot 9] 4s: Close-up on @Kade. @Kade (low): You were never meant to find me. (voice: Kade) |
| 3. The rooftop | [Shot 10] 3s: Close-up on @Ana. @Ana (steady): Watch me. (voice: Ana) |
| 4. The canal | [Shot 11] 5s: Wide shot, low angle. Two speedboats tearing down a canal between glass towers, wake spraying neon. Ambient: engines rising. |
| 4. The canal | [Shot 12] 5s: Medium shot. @Ana at the wheel, rain across her face, a siren turning blue-white behind her. |
| 4. The canal | [Shot 13] 5s: The harbour from shot 1 at sunrise, still and empty. @Narrator (voice-over, low): Neon Tide. Ambient: surf. |
That is four segments of 15 seconds and 13 shots. How to write one shot · A segment stays in one place.

## Sound is half the reel

- The score is built in. The site plans the music for each episode and mixes it under the voices; a Music switch turns it off. There is no track picker, so shape the mood with the cut and the ambient lines. Background music.
- Write the sound into the shot. Surf, a heartbeat, bass building: the ambient line is free and does a lot.
- Say the title. The narrator speaks it in the last shot; add any on-screen text in your own editor afterwards.
- If you need a particular song, add it in your own editor after you download the file, and only use music you have the right to use.

## What a sizzle reel cannot do here

- Legible on-screen text. Generated video does not write it; add titles and lower thirds afterwards.
- Real actors or famous faces. Real people cannot be depicted.
- A fixed track. The score is chosen by the site.
- Multi-track editing. The editor is one card per segment, not a timeline of layers. The episode editor, a tour.

## Cost it before you render

Sixty seconds of video is 60 seconds at 10 credits a second on the default tier, plus 1 credit a second to join. Draft at the lowest resolution and finish only the shots you keep; fix single shots rather than redoing a segment; and use the free 5-second preview to check the look first. Draft at 480p, finish at 1080p · The free preview · Spending fewer credits · AI trailer generator · AI movie pitch generator.

Say what it is when you share it: an AI-made sizzle reel or proof of concept, not a finished film. Do AI videos need a label?

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Cyber Metropolis, Neon Nights; Costume Style Near-Future Techwear, Streetwear; Visual Style Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw Ana and Kade and the four places (harbour, market, rooftop, canal); the narrator is an Off-screen character with a voice and no drawing. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Length | Segments of 15 s | Shots (4 to 5 s each) | Use |
| 30 seconds | 2 | 6 to 8 | A social teaser |
| 60 seconds | 4 | 12 to 14 | The standard sizzle |
| 90 seconds | 6 | 18 to 22 | A longer pitch reel |

## FAQ

### What is a sizzle reel?

A short video that sells the feel of a show or film, its world, look and energy, rather than telling its plot.

### How long should a sizzle reel be?

60 to 90 seconds. Thirty seconds works as a social teaser.

### How is a sizzle reel different from a trailer?

A trailer sells a story's hook and stakes; a sizzle sells the tone and the world, with no plot.

### Can I make a sizzle reel with AI?

Yes. Draw the cast and the places once, write 12 to 18 short shots, render the segments and join them with the built-in score.

### Can I choose the music?

No. SceneMixer chooses the score and mixes it under the voices; a Music switch turns it off. Add a specific song in your own editor if you have the rights.

### Can the title appear on screen?

Not as legible text from the video model. Have the narrator say it, or add it in your own editor.

### Is it good enough to pitch with?

Present it as previs or a proof of concept: it shows tone, cast and pacing, not finished footage.

### How much does a 60-second reel cost?

Sixty seconds at 10 credits a second on SceneMixer's default tier plus 1 credit a second to join; the first storyboard on your first drama and a 5-second preview are free.

Updated November 3, 2026


---

# Remake a Video: Source Limits, Read Time and Price | SceneMixer

URL: https://scenemixer.com/guides/source-video-limits-read-time-and-price/
What clip the Remake tab accepts (format, length, size), what the read costs and how long it takes, what the progress page shows, and what comes back at the end.

# Remake: Source Video Limits, Read Time and Price

Quick answer: The Remake a video tab takes one MP4 or MOV clip, within the length and size limits shown on the composer. Press Start remake and the clip is read shot by shot — who is in each, what they do, what they say — and laid out as a project with a cast, sets, props and a storyboard. The read is priced by the clip’s length in minutes, rounded up, and the price is on the button before you press it. It takes a few minutes for a short clip; the progress page shows each phase. If it stops, the credits that were not used come back.

How a remake works, step by step is the overview; this page is the limits and the money.

## What it accepts

One clip, MP4 or MOV, no longer than the length the composer’s hint gives (Step 1: add a clip — up to N minutes), within the megabyte limit it prints under the box (MP4 / MOV up to N MB). Longer or larger clips are refused before anything is charged, with a message to trim or export a smaller version. A file that cannot be read is refused too: export it again as MP4 or MOV.

Use material you have the rights to. The note under the box says it plainly: we read it, then shoot it again — the result is a new picture. The source is never handed to the video model, and the faces in the result are your own cast’s. Reading a clip is not a licence to its story. Rights.

## The one line

Optional: say in one line who to watch and how they are related — the things the picture alone will not tell us. Names, relationships, who the lead is. It helps the read name the cast sensibly; leave it blank and names are invented.

## What it costs

The read is priced by the clip’s length alone, in minutes, rounded up to the next minute; the number of cuts never enters the price. The button shows the figure before you start. It is charged when the read completes. Everything after the read — images, storyboards, video — is priced as in any drama, step by step. Where credits go.

## How long, and what you see

The progress page (Reading your clip) walks through Splitting into shots, Identifying characters, scenes and props, Reading each shot — picture and dialogue with a shot counter, and Checking and laying out the storyboard. Names appear as they are found. A clip of a couple of minutes takes several minutes to read; the page shows elapsed time and an estimate. You can leave; it finishes without you.

## What comes back

Here is what we found: N shots, N characters, N scenes, N props, N lines of dialogue. Then Open the project: step 1 of the wizard, where you choose the visual style before any images are made. After the read.

## If it stops

The page says where (Stopped here) and why. Usually the upstream service was busy and Try again works. Credits that were not used go back to your balance; a read that never finished is not charged. Cancel this remake does the same: unused credits return, and to start over you upload the clip again from the home page.

| | Limit or rule |
| Format | MP4 or MOV |
| Length | The limit in the composer’s hint |
| Size | The megabyte limit printed under the box |
| Price | By length in minutes, rounded up; shown on the button |
| Charged | When the read completes; unused credits return if it stops |
| Rights | Only material you may use; the source is never shown to the video model |

## FAQ

### How long a clip can I remake?

As long as the composer's hint says, within the size limit printed under the box. The limit can change, so the composer is the source.

### What does the read cost?

It is priced by the clip's length in minutes, rounded up. The button shows the price before you start.

### Is the read charged if it fails?

No. SceneMixer returns the credits that were not used, and a read that never finished is not charged.

### Is my clip used to generate the video?

No. It is read and set aside; the result is shot again with your own cast.

### How long does the read take?

Several minutes for a clip of a couple of minutes. The progress page shows an estimate, and you can leave.

Updated November 3, 2026


---

# Spatial Consistency in AI Video: Everyone Stays Put | SceneMixer

URL: https://scenemixer.com/guides/spatial-consistency-everyone-stays-put/
An automatic AI 3D model of every scene keeps seats, doors and walls in place from one cut to the next, and the camera only cuts to angles that exist. Nothing to set up.

# Everyone Stays Where They Sat

Quick answer: Most AI video falls apart at the cut: the man who was on the left is now on the right, the window moved, the reverse shot shows a wall that was never there. SceneMixer builds an automatic AI 3D model of each scene — the room, the furniture, and where every person is placed — and generates every shot in that scene against the same model. Seats, doors, walls and who stands next to whom hold across every cut. There is nothing to set up; it happens when you generate.

Open a finished episode, find a scene with two or more people at a table or in a car, and watch the cuts between them.

## The problem this solves

A conversation is three or four shots: her, him, the two of them, her again. In a real room those shots agree with each other. She is on the left of the table in the wide shot, so in her close-up the window is behind her, and when we cut to him the door is on his side. That agreement is what makes cuts invisible.

Generated shot by shot with nothing holding them together, those four shots each invent their own room. The viewer feels it before they can say what is wrong.

## What SceneMixer does about it

For every scene, SceneMixer builds an automatic AI 3D model: the space, the furniture that matters (the table, the bed, the front seats of the car), and a place for each character who is present. Every shot in the scene is then generated against that one model. The wide shot, the close-ups and the reverse shot all describe the same room from different positions.

You never see the model and you never touch it. It is built from the storyboard you already have and from the scene’s image.

## What holds across cuts

Seats and positions. Whoever sat at the head of the table is still there in the next shot. The person in the passenger seat does not migrate to the back.

Walls, windows, doors. The window that was behind her stays behind her. A reverse shot shows what is actually on the other side of the room, not a new room.

Who is next to whom. Left and right stay left and right. When the storyboard says she faces him, the camera is placed where that is true.

Who is in the picture. A shot framed on two people shows those two, and the person standing behind them appears where she would really appear — or not at all if she is out of frame.

## What it means for how you write

Write the storyboard the way you would describe a real room. “Mara sits at the head of the table, Jules on her right.” “The door is behind him.” SceneMixer reads those sentences, and the staging follows them. Where you say nothing, the scene is laid out sensibly for you and kept that way.

You do not have to write camera positions. If you do — “over her shoulder”, “from the passenger window” — they are honoured where the room allows it. How to write one shot.

## Where it applies

Rooms, vehicles, tables, corridors, stairwells, halls: any scene where people have positions. Wide outdoor spaces with people walking freely get less from it, because there is less to hold. It is on for every project; there is no switch and no extra cost.

New projects on paid plans go one step further. Each scene also gets a world built from its scene image: a 360° set with the walls, doors and furniture measured, and every camera is placed inside it. While a shot renders you can drag around that set yourself. How it works: What is a world model in AI video? What you see in the editor: Look around the set while a shot renders.

| Without a shared model | With SceneMixer’s automatic AI 3D staging |
| Characters swap sides between cuts | Left stays left, right stays right |
| The reverse shot invents a new wall | The reverse shot shows the real other side of the room |
| Someone at the table vanishes or doubles | Everyone present stays present, in their seat |
| A car scene: driver and passenger trade places | Driver drives, passenger rides, back seat stays back |
| Every shot needs its own description of the room | You describe the room once; the shots agree |

## FAQ

### Do I have to set up the 3D scene myself?

No. SceneMixer builds it automatically from your storyboard and the scene image. You never see or edit it.

### Does this cost extra credits?

No. It is part of generating a shot on SceneMixer.

### Can I control where people sit?

Yes, by writing it in the storyboard: who is at the head of the table, who is on whose right, who is in the passenger seat. SceneMixer’s staging follows what you write.

### Does it work outdoors?

On SceneMixer, it works anywhere people have positions relative to each other and to fixed things. A crowded street with people moving freely gives it less to hold on to than a kitchen or a car.

### What if a shot still comes out with someone on the wrong side?

Regenerate that segment. Check that the storyboard names everyone in the shot and does not contradict itself about who is where.

Updated November 3, 2026


---

# Spend Fewer Credits on an AI Drama: Habits That Save the Most

URL: https://scenemixer.com/guides/spend-fewer-credits-on-an-ai-drama/
Credits go on video, so savings come from what you generate and how often: fix text first, draft low, test one segment, trim instead of regenerating, use free items.

# How to Spend Fewer Credits Making an AI Drama

Quick answer: Most credits in a drama go on video, and most wasted credits go on regenerating video that the text could have fixed. Fix the storyboard before you generate, because editing text is free. Draft on a low tier and finish on a higher one. Generate one segment before the whole episode. Repair a finished clip with the cheapest tool that fixes it: Trim for extra seconds, Redub for a line, regeneration only when the picture itself is wrong. And use the free items your account comes with before paying for the same thing.

Open your credit history and read the last week: the biggest lines are almost always video regenerations.

## Fix text, not video

Editing the storyboard costs nothing; generating it into video costs every time. Check names, seconds, states and props before the first generation, and when a clip comes out wrong, find the cause in the text before regenerating. Checklist before you generate · When a shot comes out wrong.

## Draft low, finish high

Video is priced per second, and lower-resolution tiers cost less per second. Work out timing, faces and blocking on a low tier, then generate only the segments you keep on a higher one. Draft at 480p, finish at 1080p.

## One segment before Generate all

Generate all submits every segment at once. If the sheet or the wording is off, every one of them is off. Generate one, watch it, fix, and then generate the rest. Generate all.

## Use the cheapest fix that works

- A stray second or word at the start or end: Trim, a small fixed price.
- A wrong line, everything else fine: Redub with lip sync, priced per second at a low rate.
- One visual detail wrong, the rest right: AI visuals; it is priced at a multiple of the tier’s per-second price, so for a clip that is wrong in several ways, regenerating is cheaper.
- Framing, people or action wrong: edit the text and regenerate that segment only.

What each clip fix costs.

## Do not re-parse to fix one shot

Re-parsing an episode replaces its whole storyboard, deletes its videos and charges the storyboard again. To change one shot, edit that segment. What you lose when you regenerate a storyboard.

## Use what is free

SceneMixer gives your account free character, scene and prop sheets, your first project’s first episode storyboard is free, and you get one free video preview. Use them before paying for the same thing. Most failed generations are refunded automatically, so a failure is not a reason to buy more. What you get for free · When a generation fails.

## Seconds you do not need

A segment costs by its length. Do not pad shots with seconds the story does not use; give time to lines and reveals, and keep the rest tight.

| Situation | Cheaper choice |
| Unsure the sheets or wording are right | Generate one segment, not all |
| Trying out timing and blocking | A low tier |
| Extra second at the start or end | Trim |
| One wrong line | Redub with lip sync |
| One shot is wrong | Edit that segment and regenerate it, not the storyboard |

## FAQ

### What uses the most credits?

Video, priced per second. Reading the script and images cost far less by comparison.

### Are failed generations charged?

SceneMixer refunds most failures automatically; the task's line in your credit history shows Refunded. A few are held for review first and show Under review.

### Is editing the storyboard free?

Yes. Only generating video, images, storyboards and paid clip fixes cost credits.

### Where do I see the prices?

On each button before you press it, and on the pricing page.

Updated November 3, 2026


---

# State Image Messages: Main Image First, Describe the Difference

URL: https://scenemixer.com/guides/state-image-messages-main-image-first-describe-the-difference/
States are second looks of a character, scene or prop, generated from the main sheet. The seven messages you can meet while making one and what each needs.

# State Image Messages: ‘Needs Its Main Image First’, ‘Describe How It Differs’, ‘Can’t Use Reference Images’

Quick answer: A state (a costume change, a burnt room, a drawn blade) is generated from the card’s main sheet, so it needs three things: the main sheet must exist, the state must have a name, and it must have a sentence saying what differs. Each missing piece has its own message. Two others are about limits: a card can hold only so many states, and a state that is still generating cannot be replaced until it lands. One is about your settings: if your chosen image model cannot work from a reference image, states cannot be made with it; pick another image model in settings and generate again. And one means the feature is switched off site-wide for the moment.

On a card with no image yet, add a state and generate it: the message asks for the main image first; generate the sheet, then the state.

## How a state image is made

The main sheet is the reference; the difference sentence is the instruction; the model edits the sheet into the new look. That is why every message here is about one of those inputs. What states are for: Costume Change and Scene States.

## The messages, one by one

This asset needs its main image first — the card has no sheet yet, or it is still generating. Generate the sheet, wait for it, then the state. Describe how this state differs from the main image first — the difference field is empty. Write what changes, in structural words. Write the Change, Not the Picture. Give this state a name — the name field is empty. This asset already has a state with that name — names are unique per card. You’ve reached the maximum number of states for this asset — delete one you do not use, or merge two into one. This state is still generating — wait for it to land or fail before replacing its image. Asset states are turned off right now — site-wide switch; while it is off, clips use the base sheets instead of any state images.

## The image model message

The current image model can’t use reference images: you selected an image model in settings that generates from text only. States are always edits of a reference, so they cannot be made with it. Open settings, choose an image model that accepts references (the default does), generate the state, and switch back if you prefer the other model for plain sheets.

## If the state did not come out as described

The difference sentence is the lever: shorter, structural, one change. A state that came out as the original means the sentence restated the whole sheet. Use a State in One Scene once it looks right.

## Credits

A state image is a generation at the sheet price; the first project has a separate free allowance for states. Messages above are refusals and charge nothing; a state that fails after starting is settled by the usual rules.

| Message | Needs |
| Needs its main image first | Generate the sheet, wait for it |
| Describe how this state differs | A difference sentence |
| Give this state a name | A name |
| Already has a state with that name | Another name |
| Maximum number of states | Delete or merge one |
| Still generating | Wait for it to land |
| Image model can't use reference images | Pick another image model in settings |
| States are turned off | Wait; clips use the base sheets meanwhile |

## FAQ

### Can I upload my own image as a state?

Yes; then no difference sentence is needed for generation, though a short one helps the storyboard.

### Do states share the free sheet allowance?

No. SceneMixer gives states their own first-project allowance.

### Why can't the text-only image model make states?

A state is an edit of the main sheet; a model that cannot read references cannot edit.

### Does deleting a state affect clips already made?

No. It only stops future scenes from selecting it.

### Where do I pick which scene uses the state?

In the editor, per segment; or per episode in Step 3.

Updated November 3, 2026


---

# Still Generating, Syncing or Rendering: The Five Wait Messages

URL: https://scenemixer.com/guides/still-generating-still-syncing-the-five-wait-messages/
Five messages refuse a click because something on that episode or segment is still running. What each waits for, how long it takes, whether you can stop it.

# ‘Still Generating’, ‘Still Syncing’, ‘Still Rendering’, ‘Still Being Prepared’: What Each Message Is Waiting For

Quick answer: These messages are refusals, not failures: the site will not start a second job on top of one that is still running on the same episode or segment, and nothing is charged for the refused click. Storyboard still generating: the episode’s storyboard job is running; wait for it, or stop it from the episode card. Edit still syncing: you edited a segment in a translated view and the change is being written back; a minute or two. Video still generating: that segment already has a clip in progress. Some segments still rendering: you asked to regenerate the storyboard while clips are in flight; wait for them to land. Episode still being prepared: on a long script, this episode’s outline is being completed; the storyboard button waits and then continues by itself.

Click generate on a segment whose card shows a spinner: the message names the running job.

## The five messages and what each is waiting for

This episode’s storyboard is still generating — the storyboard job for that episode is running (a few minutes). Until it lands, segments cannot be generated or edited. On a new project the first part of that wait can be the Building the world stage, while the scenes are prepared before the storyboard is written (what it is). This segment’s edit is still syncing — an edit made in a translated view is being written into the working text; generation on that segment waits until it lands (normally well under two minutes). This segment’s video is still generating — a clip for it is already in progress; a second one would be a second charge. Some segments in this episode are still rendering — regenerating the storyboard would orphan clips in flight; wait for them to land. This episode is still being prepared — long scripts get their episode outlines completed before a storyboard can be written; this one is being finished now. Task states in general: What Each Task Status Means.

## How long, and what to do

Storyboard: minutes; watch the episode card. Sync: under two minutes; the status line above the segment clears itself. Video: the tier’s usual time; the card shows the estimate. Rendering before regenerate: until the last in-flight clip lands. Prepared: usually about a minute; the storyboard button waits and then continues on its own. None of these needs a refresh, though a refresh never hurts.

## Stopping instead of waiting

A storyboard, parse or AI Create job can be stopped: SceneMixer refunds it if it had not been sent to the model yet, and charges it otherwise. A clip that is generating cannot be stopped; let it finish, then regenerate if you must. Stop a Generation.

## Why the sync message exists

When you read a storyboard in a language other than the project’s working text and edit it, the edit has to be carried into the working text before a clip can use it, or the clip would be generated from the old text. The wait guarantees that what you see is what gets shot. The languages on this site.

## Why the prepared message exists

Scripts above a length are read in two passes: first the cast and scenes, then the episodes as outlines. SceneMixer completes the first three episodes in the background; any other episode is completed when you press Generate storyboard on it (the next one comes along), and the message is that gap. Long Scripts Are Read in Two Passes.

| Message | Waiting for | Typical wait | Can stop? |
| Storyboard still generating | Episode storyboard job | Minutes | Yes, from the episode card |
| Edit still syncing | Your edit reaching the working text | Under 2 min | No; it clears itself |
| Video still generating | Clip on this segment | Tier's estimate | No; let it finish |
| Segments still rendering | In-flight clips before a storyboard regenerate | Until they land | No; let them land |
| Episode still being prepared | Outline completion after a long parse | About a minute | No; the button continues by itself |

## FAQ

### Was I charged for the refused click?

No. SceneMixer never charges for a refusal.

### The sync message has been there for ten minutes.

Refresh; if it persists, the write-back failed and the segment shows a status line with a retry.

### Can I edit a segment while its video generates?

Text edits are allowed; a new generation waits for the running one.

### Why does regenerating the storyboard care about clips?

The new storyboard replaces the segments; clips landing afterwards would have nowhere to go.

### Does 'still being prepared' cost anything?

No. Outline completion is part of the parse you already paid.

Updated November 3, 2026


---

# Stop a Generation: What Can Be Stopped, What Comes Back | SceneMixer

URL: https://scenemixer.com/guides/stop-a-generation/
Stop exists for AI Create, script parsing and storyboards, not for video clips. What the confirmation says, when credits come back, and what to do about a clip instead.

# Stopping a Generation: What Can Be Stopped, and What Comes Back

Quick answer: A Stop button appears while text is being written: AI Create, the script parse in Step 1, and an episode’s storyboard in Step 3. Press it and confirm. If the job had not been sent to the model yet, the held credits come back; if it had, they are spent, and the confirmation says so before you press. A video clip that is generating has no Stop: it runs to the end and lands as a new version of the segment, which you can ignore or replace. Deleting an episode or project refunds the work that had not reached the model; what happens to work already sent is below.

Open a finished episode to see a segment card and its version history.

## What can be stopped

AI Create, in its writing window. The script parse, in Step 1 while the script is being read. An episode’s storyboard, on the episode card in Step 3. Each shows a small Stop button while the text streams in. Pressing it asks Stop this? with a sentence worth reading: credits already spent are not refunded, because the request has already gone to the model; stopping just frees the screen so you can start over.

## What comes back

Not yet sent to the model. The job was still queued, or had started but had not called the model yet. Stop, and SceneMixer returns the held credits to your balance.

Already sent. The model was writing. Stop, and those credits are spent; what you get is a free screen to edit and run again.

The line between the two is not visible from the screen, which is why the confirmation warns you every time. If the job had in fact just finished, Stop says it could not stop it, and the result is yours.

## A clip cannot be stopped

Once you press generate on a segment, the clip runs to the end. If you spotted a mistake in the storyboard after pressing generate, edit the text while it renders; the edit is saved and the running generation is unaffected. When the clip lands, it becomes a new version; regenerate from the new text if you need to. SceneMixer charges for two clips instead of one, but you keep both.

## Deleting while generating

A segment whose video is still generating cannot be deleted; wait for it to finish. Deleting a whole episode or project while videos generate is allowed: work that had not started, or had not reached the model yet, is refunded. Work already sent is charged if you delete an episode; if you delete a project, clips already sent are settled by what the provider reports (charged if delivered, refunded if it confirms a failure), and other work already sent is held for review. Deleting projects and episodes.

| What is running | Can you stop it | Credits if you do |
| AI Create, not yet sent to the model | Yes | Returned |
| Script parse or storyboard, not yet sent | Yes | Returned |
| Any of the three, already sent | Yes | Spent |
| A video clip | No | Charged when it lands; the clip is yours |

## FAQ

### I stopped a storyboard. Where are my credits?

If it had not been sent to the model yet, back in your balance. If it had, spent — the confirmation says so before you press it.

### Why is there no Stop on a generating clip?

A clip is paid for by the rendering itself; it runs to the end and lands as a version you can keep or ignore.

### Can I edit the storyboard while a segment generates?

Yes. The running generation uses the text it started with; your edit is saved for the next one.

### Can I delete a segment that is generating?

No. Wait for it to finish, then delete.

### What happens if I delete the project while videos are generating?

Work that had not started, or had not reached the model, is refunded. Clips already sent are settled by what the provider reports: charged if delivered, refunded if it confirms a failure. Other work already sent is held for review.

Updated November 3, 2026


---

# Story Bible Template for a Series: Cast, Places, Rules and Arc

URL: https://scenemixer.com/guides/story-bible-template/
A story bible template for a series: premise, cast, places, rules, season arc and episode engine, and how each part becomes sheets, voices and sets in SceneMixer.

# Story bible template for a series

Quick answer: A story bible is the one document that keeps a series consistent: premise, cast, places, the rules of the world, the season arc and the engine that repeats every episode. Write it before episode one. In SceneMixer each part has a home: the cast becomes character sheets and voices, the places become locations drawn once, and the whole thing is reused from episode to episode, so the bible and the screen agree.

Fill the table below in one page for a series you want to make, then read the cast list that SceneMixer finds in your first script and compare.

## The seven parts

| Part | Write | Becomes, in SceneMixer |
| Premise | One sentence | The brief you type into AI Create |
| Cast | Name, role, want, look, voice | Character sheets and a voice each |
| Places | Three or four that recur | Locations drawn once and reused |
| Rules | What can and cannot happen | Lines in the script and storyboard |
| Season arc | What changes by the end | The episode outline |
| Engine | What repeats every week | The shape of every episode |
| Tone and look | Three adjectives and a reference look | The visual style chosen once in Step 1 |

AI character sheet generator · Episode outline generator.

## A cast entry that works

MARA OKAFOR, 34, night-shift nurse. Wants to leave the city and tells no one. Tall, close-cropped hair, a green cardigan she wears every episode, a chipped front tooth. Voice: low, dry, quick to deflect. Keep each cast entry to one line of want, one line of look, one line of voice.

## Keeping the bible and the series in sync

- Pick characters from your library for each new episode so nothing is redrawn. Reuse a cast across projects.
- Set a costume change as a state, not a new description. Costume changes and states.
- Keep recurring places the same. The same room in every episode.

## Start the series

AI TV series creator · AI mini series maker · Making a multi-episode series.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Hospital Ward, Urban Opulence; Costume Style Professional Uniforms, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the cast entries of your bible first; they are what every episode reuses. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Cast entry | Length |
| Want | One line |
| Look | One line: face, hair, build, one detail |
| Voice | One line |

## FAQ

### What is a story bible?

A document that fixes a series' premise, cast, places, rules, arc and engine so every episode agrees.

### Do I need one for a short film?

No; a one-page version is enough for a series of three or more episodes.

### How does SceneMixer use a bible?

The cast and places become sheets and locations reused across episodes; the arc becomes the outline.

### Can I paste it in?

Yes, as text. SceneMixer reads the cast and places out of it.

### Where do voices go?

Each character gets a voice in the SceneMixer project, and members can upload a sample.

### Is it free?

Writing on SceneMixer is free; reading costs 4 credits per 1,000 characters.

### Can I reuse the cast in another series?

Yes, from your library.

Updated November 3, 2026


---

# Storyboard Checks: Two Drafts, What Is Saved, When It Fails

URL: https://scenemixer.com/guides/storyboard-failed-after-two-drafts-what-to-do/
Each storyboard draft is checked for structure, language, dialogue and timing. A failing draft gets one rewrite; the better draft is saved. When the task fails instead.

# Storyboard Checks: Two Drafts, What Gets Saved, and the Rare Case Where the Task Fails

Quick answer: Every storyboard draft is checked before it is saved: the segment structure is complete, the text is in the project’s working language, every spoken line from the outline is present, shot timings add up. A draft with problems gets one rewrite with the list of problems, and the better of the two drafts is saved — so a storyboard can arrive with a small problem left in it, which you fix in the editor. The task only fails when neither draft comes back as a usable storyboard at all: nothing is saved, you are not charged, and the credits stay held until support reviews the task and returns them.

After a storyboard is saved, read it once against the outline: a missing line or a shot in the wrong language is the kind of leftover the second draft did not fix.

## What the two drafts are

The model writes a draft; the site checks it; if there are problems, they go back with the draft for one rewrite; the rewrite is checked too. The site keeps whichever draft has fewer serious problems. There is no third attempt. How a storyboard is generated: Generate the Storyboard for an Episode.

## What is checked

Structure: every segment has its header and its shots. Language: the text is in the working language; a translated interface does not change this. Dialogue: every line from the episode outline appears, on a shot, with its speaker. Timing: shot seconds are within range for the segment length. Output cut off (the model stopped mid-text) counts against the draft: Output Was Cut Off.

## When the task fails

Only when neither draft is a usable storyboard — both came back unreadable or cut off. Then nothing is saved for the episode and you are not charged; the storyboard credits are held rather than returned on the spot, because the site cannot always tell how much of the work the service already did. Support reviews held tasks and returns the credits; you can also send feedback to support from the Ask window, with the episode name.

## When a saved storyboard has a leftover problem

Open the segment and fix the text: add the missing line to the shot where it belongs, rewrite a shot that slipped into another language, adjust a shot’s seconds. Editing is free. Regenerate the whole storyboard only if most of it is off.

## If it keeps going wrong

Too much in one episode: a very long outline with many scenes and lines is more than one storyboard lays out consistently; shorten or split it. Dialogue language: outline lines in one language while the project expects another make the model slip; the dialogue language cannot be changed after the script is read, so for another language start a new project with the same script. Odd structure: an outline that is mostly narration gives the model nothing to shoot; rewrite it as scenes with things happening. Transient: the model service was unstable; retry a few minutes later.

| What you see | Likely cause | Do |
| A line from the outline is missing | Very long outline | Add it in the editor; trim the outline next time |
| A shot in the wrong language | Outline lines in another language | Rewrite the shot; fix the outline’s lines |
| Storyboard task failed | Neither draft was usable | Retry; credits held until support returns them |
| Output cut off | Too much text | Split the episode |

## FAQ

### Was I charged for a failed storyboard?

No. SceneMixer does not charge a failed storyboard task; its credits are held until support reviews and releases them.

### Why not try a third time automatically?

Two drafts almost always produce a usable storyboard; when they do not, the material usually needs a change first.

### Does a failed storyboard delete my clips?

No. Existing segments and clips stay until a storyboard succeeds and replaces them.

### How long can an episode outline be?

There is no fixed limit here; very long outlines are the most common reason for leftover problems.

### Can I write the storyboard by hand instead?

Yes, segments can be inserted and edited manually.

Updated November 3, 2026


---

# Storyboard Template for AI Video, With a Worked Example

URL: https://scenemixer.com/guides/storyboard-template/
A storyboard template you can copy: number, seconds, framing, action, line and sound for every shot, with a worked example in the format SceneMixer reads.

# Storyboard template for AI video

Quick answer: A storyboard for AI video needs five things per shot: its number and length in seconds, the framing, one action, any spoken line with who says it, and the sound. SceneMixer writes shots in exactly that shape, [Shot 1] 3s: Wide shot, a small town at dawn under fog., and you can paste your own in the same style. Below is a blank format you can copy and a worked example from a one-minute trailer.

Copy the blank row, fill it for your first five shots, and paste them into the storyboard of a new episode. Editing storyboard text is free.

## The blank format

| Part | Write | Example |
| Number | [Shot n] | [Shot 3] |
| Seconds | A whole number of seconds | 4s |
| Framing | Wide, medium, close-up | Close-up on |
| Action | One thing happening | @Dr Ellis closing her eyes |
| Line | @Name (tone): the line (voice: Name) | @Sheriff Dunn (uneasy): How could you possibly know that? |
| Sound | Ambient: … | Ambient: a heartbeat, then a whisper |

Use @Name for a character so the shot uses their drawing. The @ menu in the storyboard.

## A worked example: a one-minute trailer

| Beat | Shot |
| The town | [Shot 1] 3s: Wide shot, a small town at dawn under fog. |
| The lab | [Shot 2] 3s: Medium shot. @Dr Ellis pulls off her gloves in a cold morgue lit by tube lights. |
| The gift | [Shot 3] 4s: Close-up on @Dr Ellis closing her eyes. Ambient: a heartbeat, then a whisper. |
| The line | [Shot 4] 4s: Close-up on @Sheriff Dunn. @Sheriff Dunn (uneasy): How could you possibly know that? (voice: Dunn) |
| The title | [Shot 5] 3s: Wide shot, the town at night, every porch light off but one. |

That is five shots and 17 seconds. The full trailer guide.

## Rules that make a storyboard render well

- One action per shot. Two actions are two shots.
- One place per segment. Keep a scene in one place.
- Keep lines short. A shot of 3 to 5 seconds holds about one line.
- Say who is where. Position carries across cuts.
How to write one shot · Everyone stays put.

## Let the tool write the first draft

You do not have to fill the template by hand. Paste a script or one line and the storyboard is written for you, then edited in the same format. AI storyboard generator · Storyboard to video · AI shot list generator.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style American Small Town, Hospital Ward; Costume Style Professional Uniforms, Smart Casual; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the coroner and the sheriff; the town and the morgue are the two places. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Shot | Seconds | Framing | Action | Line |
| 1 | 3 | Wide | The town at dawn | — |
| 2 | 3 | Medium | The doctor in the morgue | — |
| 3 | 4 | Close-up | She closes her eyes | — |
| 4 | 4 | Close-up | The sheriff reacts | How could you possibly know that? |
| 5 | 3 | Wide | The town at night | — |

## FAQ

### What goes in a storyboard template?

For every shot: number, seconds, framing, one action, any spoken line with the speaker, and the sound.

### What format does SceneMixer read?

Numbered shots such as [Shot 1] 3s: Wide shot, … with @Name for characters and (voice: Name) after a line.

### Can I paste my own storyboard?

Yes. Edit or paste the storyboard text into SceneMixer in the same format; editing is free.

### How long should a shot be?

Usually 3 to 5 seconds; a segment of up to 15 or 30 seconds holds several shots.

### Do I have to write it by hand?

No. SceneMixer drafts the storyboard for you once you paste a script or one line.

### Is the storyboard free?

Editing is free; the first episode's storyboard on your first drama is free on SceneMixer.

### Can I use the template for a live shoot?

Yes. The five parts work for any shoot; the @Name tags are for SceneMixer.

Updated November 3, 2026


---

# Anime and Stylised Looks Across the Tiers | SceneMixer

URL: https://scenemixer.com/guides/stylised-and-anime-looks-across-tiers/
How a project’s visual style (anime, cel, 3D, painterly) travels through the image and video tiers, what each tier changes, and where stylised projects go wrong.

# Anime and stylised looks across the tiers

Quick answer: The look is a project setting chosen in step 1 — a visual-style preset, or your own on a paid plan — and every tier renders it: the image tiers draw the sheets in that style, the video tiers keep the shots in it. Switching tiers does not switch styles. What tiers change is resolution, price and how much of a long description they read; for stylised projects the two things that go wrong are a style picked in a hurry — it cannot be changed once you continue — and a face too generic to hold across shots.

Choosing the visual style walks through the presets and the three style axes.

## Where the style lives

In the project, set in step 1 before the script is read: a visual-style preset (anime, cel-shaded, 3D, painterly, live-action and more) plus a scene style and a costume style. The character sheets, location images and every video segment inherit it. There is no per-tier style; a tier renders whatever style the project has.

## What the image tiers change

| Tier | On a stylised project |
| Qwen-Image 3.0 / Pro | Follows the preset; keep the identity features first in the description |
| SeedDream 5.0 Pro / Lite | 2K sheets in the style; reads long descriptions to the end |

## What the video tiers change

Nothing about the style itself. The Wan, Seedance and H3 tiers on SceneMixer all keep the project’s look; they differ in segment length, resolution, references and price, exactly as on a live-action project. A flat, cel-shaded look survives 480P better than a painterly one with fine texture, which is a reason to draft stylised projects on the cheapest tier and finish only the keepers.

## Where stylised projects go wrong

Style picked in a hurry. The visual style is fixed once you continue from Step 1 and shapes every sheet and clip; to try another, start a new project with the same script. Generic faces. Anime faces without a distinguishing feature drift between shots; give each character one unmistakable trait in the description. Mixed axes. A period costume style with a modern scene style reads as a mistake; set all three axes together.

## FAQ

### Is there an anime video tier?

No. Anime is a project style on SceneMixer; every video tier renders it.

### Can I change the style after generating?

No. The visual style is fixed once you continue from Step 1; for another style, start a new project in SceneMixer with the same script.

### Which image tier is best for anime sheets?

SceneMixer’s default and Pro Qwen tiers follow the presets well; SeedDream Pro for 2K.

### Does a stylised project cost less?

No; SceneMixer charges per sheet and per second regardless of style.

### Do the video tiers handle 3D styles?

Yes, as a project style, on any SceneMixer tier.

Updated November 3, 2026


---

# Subscribe, Upgrade or Cancel a Plan: What Happens When | SceneMixer

URL: https://scenemixer.com/guides/subscribe-upgrade-cancel/
How to cancel or turn off auto-renew, how to subscribe monthly, quarterly or yearly, when credits arrive, what an upgrade charges, and why there is no downgrade.

# Subscribing, Upgrading and Cancelling

Quick answer: Open Membership in the sidebar, pick a plan and a cycle (monthly, quarterly, yearly; longer cycles are discounted), and Subscribe. Paid plans start from $7.99/month. The plan’s credits are granted when each period starts — monthly, even on a yearly plan — on top of anything you bought. Upgrade to a higher plan takes effect now: you see a confirmation with the prorated difference before anything is charged. There is no downgrade; subscribe to a lower plan after the current one ends. Cancel subscription stops the renewal and keeps everything until the period ends.

What a paid plan changes lists the perks; this page is the mechanics.

## Subscribe

Each plan card lists its monthly credits, the sitewide discount, the concurrent-task limit, storage and the perks. Pick Monthly, Quarterly or Yearly; the card shows the price per cycle and the saving. Subscribe opens the payment window: card or wallet, or PayPal, renewing automatically; or the Airwallex tab, where Card / Google Pay renews automatically and a local method such as Alipay or WeChat Pay buys one period that does not renew (how it works). The summary shows Credits per period, First charge, Next charge, and a First subscription bonus when it applies. Before you pay.

## When the credits arrive

At the start of each period: the plan card shows Granted monthly, next on a date. On SceneMixer, a quarterly or yearly plan grants its credits month by month, not all at once; the first-subscription bonus, when there is one, is paid in full on day one. The credits go to the same balance as bought credits and are spent the same way. Seeing the grant in your history.

## Upgrade

Upgrade on a higher plan shows Upgrade to … with three lines: what is charged now to your card (the prorated difference for the rest of this period), the credits added right away, and the new price from the next period. Nothing is charged until you press Confirm and pay. The higher plan starts immediately and your billing date stays the same. If the price changed while the dialog was open, it is shown again rather than charged.

## Why no downgrade

Plans only go up while one is running. To move to a lower plan, let the current one end — cancel the renewal, use it to the end of the period — then subscribe to the lower one. Picking a plan of the same level is refused with a message to switch after the current one ends.

## Cancel

On the current plan card, next to the renewal date, there is one small Cancel button. Pressing it stops the renewal straight away — there is no confirmation step — and the card’s date line changes to Ends <date>: the plan stays usable until then, nothing is refunded and nothing further is charged. Changed your mind before that date? For a card subscription the same button now reads Resume; press it and the renewal is back on, with no new charge. A PayPal subscription cancels with the same button, but PayPal stops the subscription itself rather than scheduling an end, so no Resume follows — subscribing again is the way back. To upgrade a cancelled plan, the upgrade resumes it in the same step. If the automatic renewal did not go through and the plan dropped back to free, that is a separate screen: Renewal Payment Failed: How to Restore Your Plan.

## Changing the card

There is no update card screen on the site itself: the card on file belongs to the payment company, and the membership page never asks for card details. Where a change can be made:

- After a failed renewal — the payment page the dialog opens can take a different card, and for card payments it also lets you update the card on file. Renewal Payment Failed: How to Restore Your Plan walks through the dialog.
- At any other time — contact support and they will help you change the card on the plan; How to get help shows the way in.
- Paying with PayPal — the card or funding source lives in your own PayPal account; change it there, under automatic payments, or from the button in the failed-renewal dialog.

| Action | Charged | Takes effect |
| Subscribe | The cycle price now | Now; credits granted each period |
| Upgrade | The prorated difference now, after you confirm | Now; billing date unchanged |
| Cancel | Nothing | Renewal stops; perks until the period ends |
| Resume auto-renew | Nothing now | Renewal is back on |
| Lower plan | — | Only after the current plan ends |

## FAQ

### Do I get all the yearly credits at once?

No. A yearly plan grants its credits month by month. The first-subscription bonus, when there is one, is paid in full on day one.

### What does an upgrade cost?

The prorated difference for the rest of the period, shown before you confirm. The new price applies from the next period.

### Can I downgrade?

Not while a plan is running. Let it end, then subscribe to the lower plan.

### What happens when I cancel?

Renewal stops. You keep everything until the period ends; SceneMixer refunds nothing and charges nothing more.

### Can I undo a cancellation?

For a card subscription, yes, before the period ends: press Resume on the plan card, with no new charge. A cancelled PayPal subscription is stopped at PayPal and has no resume step; subscribing again is the way to continue.

Updated November 3, 2026


---

# Swap One Character for Another in a Finished AI Video Clip

URL: https://scenemixer.com/guides/swap-one-character-for-another-in-a-clip/
Put a different character from your project into a finished clip: same acting, same position, same lines, the new character's face, hair and outfit from their sheet.

# Swap One Character for Another in a Clip You Already Have

Quick answer: Open Edit this clip → AI visuals → Swap person. Under Who pick the person to replace, under Swap with pick the character who should be there, and press Edit into a new version. The acting, the position and the lines stay as they were; the face, hair and clothes come from the new character’s sheet. It costs 2× the per-second price of the clip’s tier for the length of the clip.

Typical use: the right scene, the right performance, but the wrong face — for example two similar-looking characters got mixed up in one clip.

## What you need first

The replacement must be a character of this project with a character sheet, because the sheet is where the new face comes from. Swap with lists only characters that have one, and never the person you are replacing. If the project has fewer than two characters, or nobody has a sheet yet, Swap person is greyed out. To bring in someone new, add them to the asset library and generate or upload their sheet first. Add a character to this episode.

## Step by step

- Select the segment and open Edit this clip → AI visuals.
- For a clip over 15 seconds, choose the half to edit.
- Click Swap person.
- Under Who, click the person to replace; people who appear in this segment are listed first.
- Under Swap with, click the character who should be there.
- Press Edit into a new version; the price is on the button.

## What carries over and what changes

The performance carries over: the same movements, the same place in the frame, the same timing and the same lines. What changes is the person: face, hair and outfit follow the new character’s sheet — including the outfit, so if the sheet shows a coat, the coat comes along. If the voice should change as well, run Redub on the result and pick the new character’s voice. Change a character's voice.

## It changes this clip only

The storyboard still names the original character, so if you regenerate the segment the original comes back. For a real recast — the part should be played by someone else from now on — change the character’s sheet instead and regenerate the segments, or edit the storyboard so the other character is named in those shots. Replacing a character sheet.

## When a swap will not help

A swap replaces one person with another in the same place. It cannot take out an extra person who should not be in the shot, and it cannot fix a clip where the same character appears twice. Those come from the storyboard: check who the shot names and regenerate. Who is in frame.

| Stays as in your clip | Comes from the new character's sheet |
| Movements and acting | Face |
| Position in the frame and framing | Hair |
| Timing and spoken lines | Clothes and accessories on the sheet |

## FAQ

### Can I swap in someone from another project?

Not directly. Copy the character into this project first so it has a sheet here, then swap. Reuse a character across projects.

### Why is Swap person greyed out?

A project on SceneMixer needs at least two characters, and the one you swap in needs a character sheet.

### Does the voice change?

Swap changes how the person looks. To change how they sound, redub the result.

### Can I swap two people at once?

No. One swap per edit; run another edit on the result for the second person.

Updated November 3, 2026


---

# Target Region: Relocating a Story to Another Country | SceneMixer

URL: https://scenemixer.com/guides/target-region-relocate-a-story/
Target Region moves a story's cast, names, streets and props to a market such as Western or East Asia. What changes, what your script keeps, and when it applies.

# Target Region: Relocating a Story

Quick answer: Target Region in step 1 tells the reading which world the story should live in. Pick Western and a Chinese web novel comes out with Western names, faces, architecture, street signs and props; pick East Asia and the reverse. Anything your script states explicitly — a Tokyo setting, a character called Ahmed — is kept. Follow script changes nothing. The region is applied when the script is read, so choose it before you continue; after the read it cannot be changed in that project.

A finished Western-region episode: the cast, the street, the signage all belong to one place.

## Where it is

Step 1, Set Up Your Project, the Target Region row. Hover the label for a short description: it sets character ethnicity, names and scene styling, and cultural settings explicit in your script are preserved.

## The options

Western, East Asia, Southeast Asia, South Asia, Central Asia, Middle East, Mediterranean, Eastern Europe, Latin America, Africa, and Follow script. The default depends on the site you are on; Follow script means the story keeps whatever world the text describes.

## What changes

Character names become local names. Faces and builds follow the region’s people. Buildings, streets and interiors look like that part of the world. Visible text in the picture — signs, labels — is in the local language. Props follow the culture: the tea set, the car, the phone. The story, its structure and its dialogue are untouched.

## What your script overrides

Anything explicit. If the script says the scene is in Tokyo, it stays in Tokyo. If a character has a name the story depends on, keep it in the text and it is kept. The region fills in what the script leaves open; it does not overrule what the script decides.

## When it takes effect

At parse time. The region is read together with the script, and the cast and locations are produced from it. After the read, Step 1 shows the summary and SceneMixer no longer lets you change the region in that project. To move an existing cast, edit the descriptions and names in Step 2 and regenerate the images; for a whole project in another region, start a new project with the same script. Checking the summary.

## Region, language and voice

The region moves the world. The Dialogue language row on the same screen decides what the cast speaks; the two are independent, so a Western-region cast can speak Japanese. Voices are matched to characters afterwards, in step 2. Dialogue language and cast localization.

## Ads and remakes

An ad follows the language of its dialogue rather than a region setting. A remake applies the region when it tidies the descriptions of the cast and places it read from your video. After a remake read.

| Element | Follows the region | Stays as written |
| Character names | Yes — local names | Names you keep explicit in the script |
| Faces, builds | Yes | Traits the script states |
| Buildings, streets, interiors | Yes | Places the script names |
| Signs and visible text | Yes — local language | — |
| Props | Yes — local culture | Objects the script names |
| Story, structure, dialogue | No | Always as written |

## FAQ

### Does the region change the story?

No. It moves the world the story happens in. Plot, structure and dialogue stay as written.

### Can I change the region after the script is read?

Not in that project. Edit the descriptions and names in Step 2 and regenerate the images, or start a new project with the same script.

### Can I keep one character's original name?

Yes. Keep it explicit in the script and it is preserved.

### Does the region set the language the cast speaks?

No. That is the Dialogue language row. They are independent.

### What does Follow script do?

Nothing is relocated; the story keeps the world the text describes.

Updated November 3, 2026


---

# Storyboard Text History: Restore an Earlier Version or the Original

URL: https://scenemixer.com/guides/text-history-restore-an-earlier-storyboard/
Every segment keeps its original storyboard text plus the text from before each of your last five changes. How to compare, restore, and what clears the history.

# Text History: Restore an Earlier Storyboard Text, or the Original

Quick answer: Press History on a segment’s storyboard card. The Original text is always there, together with the text from before each of your last five changes, each labelled Before an AI edit, Before a manual edit or Before a restore and showing how many places differ from the current text. View shows the difference; Restore puts that text back, and the current one goes into the history, so a restore can be undone too. It is free.

Try an AI edit you are unsure about: save it, then open History and restore the text before it if you change your mind.

## What is kept

The original — the text the storyboard was generated with, recorded the first time the segment is changed and never overwritten. The last five changes — each time you save a change, the text it replaced is stored. A sixth save pushes the oldest of the five out; the original stays.

Every save of the text counts: a hand edit, a saved AI edit and a restore. Saving without changing anything does not add an entry. Until the segment has been changed, the panel says This segment hasn’t been changed yet.

## Restoring

- Press History in the row of buttons under the storyboard card.
- Find the version by its label and the note on how many places differ; press View to see the first difference and the text around it.
- Press Restore, and confirm: Restoring replaces the current text. The current text is kept in history.

An entry marked Same as the current text would change nothing. Restoring changes the text only; regenerate the segment to see it on video.

## What history does not keep

Regenerating the episode’s storyboard rebuilds its segments from scratch, so their text histories go with them. Deleting a segment removes its history. History covers the storyboard text, not the videos — clips have their own Version History. Re-parse an episode: what you lose and Segment versions.

## Translated storyboards

If your storyboard is shown translated into your interface language, History shows the history of the text on your screen, the same text AI edit and hand edits change. AI edit the storyboard text.

| Label | What it is |
| Original | The generated text, kept for good |
| Before an AI edit | The text an AI edit replaced |
| Before a manual edit | The text a hand edit replaced |
| Before a restore | The text a restore replaced |
| Same as the current text | Restoring it would change nothing |

## FAQ

### How far back does it go?

SceneMixer keeps the original plus the texts before your last five changes.

### Does restoring cost credits?

No. It is a text save.

### Can I restore after regenerating the storyboard?

No. Regenerating rebuilds the segments and their histories. Copy anything you want to keep before regenerating.

### Does restoring bring back the old video too?

No. Videos are in Version History; switch there if you want the clip made from the old text.

Updated November 3, 2026


---

# Text in the Picture and Wrong Dialogue: Two Fixes in the Storyboard

URL: https://scenemixer.com/guides/text-on-screen-and-wrong-dialogue-fixes/
Text on screen, wrong or missing dialogue, a name said wrongly, an extra word: each is fixed in the storyboard or with Trim and Redub. Where to look and what to change.

# Text Appearing in the Picture, and Dialogue Read by the Wrong Voice or with the Wrong Words: Both Fixed in the Storyboard

Quick answer: Two complaints, one place to fix them. Text in the picture — subtitles, signs, captions you never asked for — comes from a shot that mentions writing, a screen or a caption, or from a description with lettering; the video is generated as clean picture with dialogue carried by sound only, so remove the mention and regenerate. Wrong dialogue — the wrong character speaking, or words changed — comes from the line’s speaker tag and voice tag, or from a line the model paraphrased; fix the tag, put the exact words in the line, and regenerate that segment.

Select the segment, press Edit, and read the shot: the speaker tag, the bracketed delivery note, the line, and the voice tag at the end.

## Text that appears in the picture

The site asks the video model for clean picture: dialogue is carried by sound, not subtitles. When lettering shows up anyway, something in the input invited it: a shot that says “a subtitle reads…” or “the sign says…”, a phone or a screen described with its content, a prop described with a label, or a character sheet with writing on the clothes. Delete the mention (describe a blank screen, an unlabelled bottle) and regenerate. On-screen captions are a separate site-wide feature that is not switched on at the moment; add readable text in a video editor after download.

## Dialogue read by the wrong character

Every spoken line in a shot has a speaker tag in front and a voice tag at the end. The voice tag decides whose voice reads it; a line without one, or with the other character’s, is read by the wrong person. Press Edit on the segment, put the right character in both tags (the @ menu inserts them), save, regenerate. A character who speaks from outside the frame needs the off-screen note in the bracket, or the model will put them in the picture. Dialogue Lines, Voice-Over and Captions.

## Words changed or a line skipped

The model reads the line as written and nothing else: if the words are wrong, the storyboard has the wrong words. Check the line in the segment, not in the script — the storyboard may have paraphrased. Put the exact sentence in, keep it short enough for the shot’s seconds (the editor does not check this, so give a spoken shot the time it takes to say), regenerate. A line that is missing from the video was missing from the shot text or had no speaker tag.

## The voice itself is wrong

If the right character speaks with a voice that does not suit them, the fix is in Step 2: rematch or pick another voice on the character, then regenerate the segments. Changing the voice does not touch existing clips. How Character Voices Work.

## Kanji, names and pronunciation

Names written in kanji or with unusual readings can be read wrongly by the voice. Write the name in the line the way it should be pronounced — kana or romanisation for Japanese, pinyin-friendly characters for Chinese — and keep the display name on the sheet as you like; only the line is read aloud. Language rules for dialogue: Dialogue Language and Cast Localization.

| Symptom | Where it comes from | Fix |
| Subtitles or signs in the picture | A mention of writing, a screen or a caption in the shot or a description | Remove the mention; describe blank surfaces |
| Wrong character speaks | Voice tag or speaker tag on the line | Set both tags to the right character |
| Words changed | The line as written in the storyboard | Put the exact sentence in the line |
| Line missing | No line or no speaker tag in the shot | Add the line with its tags |
| Voice does not suit | The character's voice in Step 2 | Rematch or pick another voice |

## FAQ

### Can I add real subtitles later?

Yes, in any video editor after download. SceneMixer delivers clean picture on purpose.

### The script has the right words but the video does not.

Check the segment's shot text; the storyboard may have shortened it. The video reads the storyboard.

### Why did the off-screen character appear on screen?

The line lacked the off-screen note. Add it in the bracket before the line.

### Do these fixes cost credits?

Editing is free; regenerating the segment costs its normal price.

### Can two characters speak in one shot?

Yes, each line with its own tags, in order.

### The clip has a word or a line I never wrote.

If it sits at the start or end, cut it with Trim; in the middle, remove that moment. To replace what is said, use Redub. Trim · Redub.

Updated November 3, 2026


---

# Text to Video: A Map From the Text You Have to the Video You Want

URL: https://scenemixer.com/guides/text-to-video/
Text to video, goal first: a one-minute story, a series, a short film or an ad. Which door each uses, how much text, what it costs, and what text cannot become.

# Text to Video: Start From the Video You Want

Quick answer: Text to video starts from the goal, not the text. A one-minute story for a feed, a series, a short film and a product ad are four different jobs, and they use different doors, type settings and amounts of text. Here a story of any length goes in as text; an ad does not, because it starts from a photo; and text itself never appears as moving type, because video models do not write legible words. This guide maps each goal to a door, a setting, an amount of text and a price shape, then lists what text cannot be turned into.

## Pick the goal first

| You want | Door and setting | Text you bring | What you get | Price shape |
| A one-minute vertical story for a feed | AI Create or Upload Script; Type Short Drama; 9:16 Portrait | A sentence up to two pages | One or a few episodes, each up to 240 seconds | A reading fee by length, a storyboard (the first is free on your first drama), then video by the second. A 60-second episode on the default tier is about 660 credits with the join, around $4.11 |
| A series with a hook in every episode | Upload Script or AI Create at Medium or Long; Type Short Drama or TV Series | A season’s worth of story | Episodes up to 240 seconds (Short Drama) or 300 seconds (TV Series) | The reading once, then each episode has its own storyboard and its own video |
| A short film | Upload Script; Type Film; 16:9 Landscape | Prose or a screenplay | One piece divided into acts, with no cap on episode length | The same fees, with fewer, longer segments |
| A promo for a product, a shop or an app | Ad Video | One line, plus one to four photos | A 15, 30 or 60 second ad with an end card | One all-in price that is a cap, or expert mode priced step by step |
In SceneMixer, the type decides how your text is cut. Short Drama is fast, dense with dialogue and built around a hook per episode; TV Series is paced for several threads; Film is slow and atmospheric and is treated as one piece in acts. It is chosen in step 1 and fixed once the text has been read, because the episodes were cut for it. To get the other type, start a new project with the same text. Short Drama, TV Series or Film has the full comparison.

## Why an ad is not a text-to-video job here

If your text is a product description, the ad route still needs at least one picture. The composer will not start without it (“Add at least one image first”), because an ad is built around something the camera can see: the first photo is the subject, the end card shows it and the other photos become the places the shots are set in. Your words still matter, but as one line under the photos about what it is and who it is for, which wins over the picture when they disagree.

So the practical advice for a text-only product idea is to get one photo or screenshot, however plain, and put your sentence under it. If you have no photo and want a story about the product instead, that is a script project and the text is the whole input. Four ways to start a project · How to make an ad from one photo.

## What kind of text you have

The amount of text decides the path, and create AI videos from text covers the three sizes. The kind of text decides how much of your own wording survives, and the site has a guide for most kinds.

| Your text is | How it is handled | Start with |
| Your own screenplay, with scene headings and character cues | Layout is read directly, so dialogue and action keep your wording | Use a screenplay |
| A web novel chapter | Read as prose into a cast, places and episodes | Turn a web novel into a drama series |
| A Reddit post or other short true-story text | Prose, usually with one voice; dialogue has to be added or lifted from quotes | Turn a Reddit story into a drama |
| A podcast or interview transcript | Spoken text becomes dialogue; scene and place have to be invented or implied | Turn a podcast episode into a drama |
| A stage play | Acts and speeches map naturally to scenes and lines | Turn a stage play into a short film |
| Only an idea | Written into a script from a brief, then read | Write the one-line brief |
Whatever the kind, write what must be said as lines (Name (delivery): words) and leave the rest as prose. A line folded into narration is summarised instead of spoken. Write dialogue so it is picked up.

## From words to seconds

Text length does not set video length on SceneMixer; the story does. Spoken English runs at about three words a second, so a scene whose lines add up to 180 words has about a minute of speech before any pause or movement is added. Each episode has a ceiling by type, and each segment is shot as one 15 or 30 second take, joined into the episode afterwards. Video is bought by the second, so the total follows the finished length rather than the page count: the default tier is 10 credits a second, and joining the segments is 1 credit a second. Shot seconds and timing · Where the credits go in one episode.

The text is read once for its whole length, and the fee follows it: 4 credits per 1,000 counted characters, with English counted lightly. A single pass covers up to 10,000 counted characters, and longer texts are read in two passes up to 300,000 English characters. How characters are counted.

## What text cannot become

| If you hoped for | What happens instead |
| Your words appearing on screen as animated type, captions or a lyric video | Not offered. Video models do not draw legible text, so on-screen words are unreliable. The one place readable words are guaranteed is the end card of an ad. Subtitles of speech are a separate switch for accounts that have bought credits or a plan |
| A link to an article turned into a video | Links are not fetched. Paste the text itself |
| A PDF as the source | Refused. Export it to plain text or Word first; .txt, .docx and .md files work |
| A table, a list or a set of bullet points | Not a story, and it can fail repeatedly. Turn it into sentences first |
| Narration laid over stock footage | Every picture is a generated shot of your cast and places. A narrator can speak as a voice-only character, but there is no footage library |
| A famous character or a real person | Refused by the IP check when a picture or clip is made. Invent your own |
The pattern behind the table: text goes in as story, and what comes out of SceneMixer is acted by a generated cast in generated places. Anything that depends on exact typography, exact people or exact footage is outside it. What you can feed in.

## Two minutes to choose the door

- Do you have photos of the thing you are selling? Use Ad Video with your sentence as the line under them.
- Is the story already written? Paste or upload it on Upload Script and pick the type. The reading is the first charge.
- Do you only have an idea? Write one line on SceneMixer’s AI Create, or press Inspire Me if you have not even that, and pick a length from Mini to Long.
- Do you have a clip rather than text? That is Remake a video, a different door with a reading priced by the clip’s minutes.
All four end in the same kind of project: a cast, places, props, episodes, a storyboard of numbered shots and then video, and every step after the first is priced the same way. For the text routes in more depth see text to video AI for the brief and AI text to video generator for the controls.

## FAQ

### How do I turn text into a video here?

Paste or upload the text into SceneMixer, or have AI write it from one line. It is read into a cast and episodes, each episode gets a storyboard, and each segment becomes a video clip that is joined into the episode.

### Can text become an ad?

Not on its own. An ad needs at least one photo or screenshot, with your text as one line about what it is and who it is for.

### How much text do I need?

SceneMixer needs about a paragraph at the least, roughly 330 English characters, and up to 300,000 English characters, which is read in two passes.

### How long will the video be?

The story decides, not the page count. A SceneMixer Short Drama episode runs up to 240 seconds, a TV Series episode up to 300 and a Film has no cap.

### Can I turn a link or a PDF into a video?

Not directly. SceneMixer does not fetch links and refuses PDFs. Paste the text, or export the document to text or Word.

### Can I make a video of animated text or captions?

No. Video models do not write legible words, so on-screen type is unreliable. Subtitles of the speech are available separately to accounts that have bought credits or a plan.

### Does the type setting change what I get?

Yes. In SceneMixer it changes where episodes break, how long each can be and how scenes are written for the camera, and it cannot be changed after the text is read.

### Is any of it free?

Reading is a small charge on SceneMixer, 15 setup images and the first storyboard are free, and one preview shot of up to 5 seconds is free. Video is charged from the first second.

Updated November 2, 2026


---

# Text to Video AI: How to Write a Brief That Becomes a Watchable Video

URL: https://scenemixer.com/guides/text-to-video-ai/
Text to video AI turns words into a clip, but a watchable video comes from a chain of text: brief, script, storyboard, shot. Each link with an example you can copy.

# Text to Video AI: Writing the Brief

Quick answer: Text to video AI turns words into moving pictures, but one sentence makes one clip, not a watchable video. A watchable video comes from a chain of text where each link is more specific than the last: a brief, a script, a storyboard, a shot. In SceneMixer you write the brief, or paste a script; the site writes the cast and the storyboard; and you edit the shot text, which is free, before any video is paid for. This guide shows each link with an example.

Copy this brief: A one-minute short: at 11:58 on New Year’s Eve a courier delivers a letter to the wrong flat and stays for the countdown; one stairwell, one balcony, fireworks off screen.

## Four links, from one sentence to a shot

| Link | What it is | Size | Who writes it |
| Brief | Who, where, and what goes wrong | One line | You |
| Script | The story in scenes and lines | 1,000 to 20,000 characters on AI Create; or your own | AI Create, or you |
| Storyboard | Shot by shot: framing, who is in frame, what happens, seconds | One per episode | The site, then you edit |
| Shot | One framing, one action, a line if there is one | 3 to 5 seconds | The site, then you edit |

Everything above the shot is text, and editing text in SceneMixer costs nothing. Only the last step, turning a segment into video, is charged, per attempt. Fixing one shot without redoing everything.

## The brief: person, place, what goes wrong

The one-line brief matters more than the length you pick. On SceneMixer’s AI Create tab the options are Mini (about a thousand characters), Brief (2,500), Short (5,000), Medium (10,000) and Long (20,000), each with its own price on the button. Mini is a test, Short is the usual first choice, and a short script gives a few episodes while a long one gives many. Choosing script length and cost.

## If you paste your own text

SceneMixer has a floor: about a paragraph, roughly 330 characters of English. Below it there is nothing to break into episodes. There is also a ceiling, and it is a hard stop rather than a silent cut, because a script that comes back looking complete when it was truncated is worse than a refusal. Long books are read in two passes. What you can feed in · Long scripts, read in two passes.

## The shot: what the camera sees

A shot is a marker with its seconds, then a few sentences about what a viewer would see and hear, in the order they would notice it: the framing, whether the camera moves, who is in the picture, the one thing that happens, and the line. Named characters are tagged with @. Keep to one action per shot, and give it enough seconds to happen.

| Weak | Strong |
| Mara arrives and is upset and talks to Teo. | [Shot 1] 4s: Wide shot, a stairwell at night. @Mara, in a wet jacket, checks a door number and knocks. |
| There is a big emotional moment. | [Shot 3] 3s: Close-up on @Mara. She looks at the letter, then at the paper hat. @Mara: Thirty seconds. Can I watch from your balcony? |
| They watch fireworks together. | [Shot 4] 4s: Two-shot on a small balcony, city lights below. Both count the last three numbers aloud. The sky flashes off screen. |
How to write one shot · Write dialogue in the storyboard.

## Six mistakes that waste credits

- Two actions in one shot. Split it.
- No seconds, or seconds that do not add up. The seconds are real; a line cut without changing them leaves dead air.
- Feelings instead of things you can see. “A big emotional moment” gives the model nothing to draw.
- Dialogue not marked as spoken. A line has to be written as a line to be picked up. Write dialogue so it is picked up.
- Expecting readable text in the picture. Signs and prices are not drawn reliably. Put them in the spoken line. Text on screen: fixes.
- Regenerating before editing. Fix the text first, look at it, and only then spend. Spend fewer credits.

## From text to video, step by step

On SceneMixer, the read costs 4 credits per 1,000 characters. Video is 10 credits a second on the default tier, and the first 15 setup images and the first storyboard are free. Generate one segment before the rest. Text to movie · Script to video.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: the two leads with one mark each and a voice; the stairwell and the balcony as locations; the letter as a prop. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Text you write | What the site does with it | What it costs |
| A brief | Writes a script on the AI Create tab | Priced by the length you pick, shown on the button |
| A script | Reads it into a cast, locations, props and episodes | 4 credits per 1,000 characters |
| A storyboard edit | Re-reads only the shots you touched | Nothing |
| A segment | Generates the video for that segment | 10 credits a second on the default tier |

## FAQ

### What is text to video AI?

A model that turns words into moving pictures. One sentence makes one clip; a watchable video needs a chain of text from brief to shot.

### How do I write a good text to video prompt?

Describe what the camera sees: the framing, who is in the picture, the one thing that happens, how many seconds, and the line if there is one.

### How long can the text be?

SceneMixer needs at least about a paragraph, roughly 330 characters of English. There is a hard ceiling, and long books are read in two passes.

### Can I write the script myself?

Yes. Paste it or upload a .txt, .docx or .md file. Or write a brief and let SceneMixer’s AI Create draft it.

### Why does my video not match my text?

Usually the shot text asks for more than one action, has no seconds, or describes a feeling. Edit the text first; editing is free.

### Does it make speech from my text?

Yes. SceneMixer speaks lines you write as dialogue in the video in any of 15 languages.

### How much does text to video cost?

SceneMixer charges 4 credits per 1,000 characters for reading, video is 10 credits a second on the default tier, and the first storyboard is free.

### Can I try it free?

A new SceneMixer account has signup credits, 15 free setup images, a free first storyboard and one free short preview.

Updated November 3, 2026


---

# The Ad One-Price Is a Cap: What You Actually Pay, Stage by Stage

URL: https://scenemixer.com/guides/the-ad-one-price-is-a-cap-what-you-actually-pay/
Hand an ad to AI and the button shows one price. It is the maximum: each stage charges its real price as it runs and the total stays at or below it.

# The Ad’s One Price Is a Cap, Not a Bill: What Is Actually Charged, Stage by Stage

Quick answer: When you hand an ad to AI, the button shows a single price. It is the most the ad can cost: the sum of every stage priced at its upper bound. As the ad runs, each stage freezes and charges its own actual price — fewer sheets than the bound, a shorter shot — so the total lands at or below the number you saw. In advanced mode you pay only the first stages up front (photo clean-up and script), and every later button carries its own price like any project. If a stage fails, that stage’s credits are refunded by the usual rules and the retry is quoted on the remaining stages; if a price rises before a stage starts, the ad pauses and asks. Ads are never the free first project.

Start an ad and open its progress page: each stage shows its own price as it runs; the sum stays under the one price on the start button.

## What the one price is

Each ad stage has a price: reading the photo is free, cleaning it is one sheet, writing the script is a writing charge, the breakdown is a parse, sheets are per image up to a bound, shots are per second by the tier, the final cut is per second. The one price is those at their maximum, added up. Stage list and prices: Ad Options.

## What is actually charged

Each stage freezes its own price when it starts and settles when it delivers, exactly as the same action would in a drama project. The bound on sheets is an upper limit; an ad with two people and one scene uses fewer. The shots are priced by the real seconds and tier. So the credit history shows several rows for one ad, and their sum is what the ad cost. Reading Your Credit History.

## Advanced mode

You pay the photo clean-up and the script up front; then the ad is handed to you as a project and every generate button shows its price. Nothing later is charged until you click it. Hand It to AI or Advanced Mode.

## Failures, retries, price changes

SceneMixer refunds a failed stage on the normal rules (provider not reached, content blocked: refunded; delivered: charged). The retry button shows a new quote for the stages still to run. Before each paid stage the ad checks the price against the quote; if it rose (a tier repriced, your discount ended), the ad pauses with a new number to confirm. Price Changed.

## Balance

Starting the ad checks that your balance covers the one price; each paid stage re-checks the remaining stages before it starts and pauses if the balance has dropped below them. A paused ad resumes from the same stage after a top-up. Refund Rules.

| Stage | Priced as | In the one price |
| Read the photo | Free | 0 |
| Clean the photo | One sheet (0 for screenshots, places and vehicles, unless people have to be removed) | One sheet |
| Write the script | Writing charge | Writing charge |
| Break it down | Parse | Parse |
| Sheets | Per image, actual count | Upper bound |
| Shots | Per second, real length and tier | Full length |
| Final cut | Per second | Full length |

## FAQ

### Why did the ad cost less than the one price?

Fewer sheets than the bound, or shorter shots. SceneMixer treats the one price as a cap.

### Can it cost more than the one price?

No. If a stage's price would exceed the quote, the ad pauses and asks first.

### Why several rows in my history for one ad?

Each stage is its own charge; the sum is the ad.

### Is the first ad free?

No. SceneMixer never applies the free first project to ads.

### Does regenerating a shot later count against the one price?

No. After the ad is done, regenerations are priced per action like any project.

Updated November 3, 2026


---

# The AI Ad Progress Page: Stages, Retry, Rebuild, Cancel | SceneMixer

URL: https://scenemixer.com/guides/the-ad-progress-page/
What the stages on an ad's progress page mean, what Try again and Rebuild final cut do, when Cancel this ad is allowed, and what each failure message asks of you.

# The Ad Progress Page

Quick answer: Every ad has a progress page that walks through its stages: Preparing images, Ad script, Breaking down, Cast & set, Shot list, Shooting, Final cut, Done, with an estimate of the time left. A stage that fails says why and offers Try again, which continues from that stage — nothing already made is redone. Rebuild final cut re-joins the finished shots without reshooting. Cancel this ad is available before shooting starts; credits spent on started steps are not refunded.

The whole ad flow, from photo to finished spot.

## The stages

Preparing images: the product shot is made from your photo, or a place is drawn from it. Ad script: the story is written. Breaking down: the script is read into cast, set and beats. Cast & set: the people, places and props get images. Shot list: the shots are written. Shooting: each fifteen-second segment is filmed. Final cut: the segments are joined, the end card added, the sound levelled. Done: download, share, make another.

In expert mode the page stops after Ad script at Workbench: Your images and ad script are ready, with Open the ad workbench. From there the stages are yours to run. The two modes.

## How long

The page shows About N min left. A fifteen-second one-click ad on SceneMixer is typically ready in about ten minutes; longer ads take longer because they have more segments to shoot. You can leave; the page keeps its state and the home card for the ad shows its stage.

## Try again

When a stage fails, the page names the stage (Stopped here) and the reason, and offers Try again. It continues from the failed stage; finished stages are kept. If the price of the remaining work has changed since you started, the button says The rest now costs ✦ N and you confirm the new figure. If a previous attempt’s credits are on hold for review, the page says so, and trying again is charged separately.

## The failure messages

Couldn’t make the product shot, Couldn’t write the story, Couldn’t break down the story, Couldn’t create the cast and set, Couldn’t build the shot list, A shot couldn’t be made, Couldn’t put the final cut together: retry that stage. This step took too long: retry. Not enough credits: top up, then the button below continues. The price or your selected model changed: continue at the new price. The quality you picked is temporarily unavailable: continue with an available one. A shot was blocked by content review: edit that shot in the editor first, then come back and retry — retrying an unchanged shot is refused. Blocks.

## Rebuild final cut

On a finished ad, Rebuild final cut joins the same shots again with the current end card. Use it after Edit end card, or after fixing one shot in the editor. It does not reshoot. While it runs the page says the new version will appear; if the rebuild fails, the previous cut stays. The end card.

## Fix this shot in the editor

A finished one-click ad can be opened in the editor to change a single shot: edit its text, regenerate that segment, then Rebuild final cut. The ad script itself stays as written. Fixing one shot.

## Cancel this ad

Available until shooting starts. The confirmation is plain: SceneMixer does not refund credits spent on steps that have started, and a cancelled ad cannot be resumed. Once shooting has begun, the button is gone — let the ad finish and decide then. In expert mode, what was already made is kept and you can continue in the workbench.

| Button | When | Does |
| Try again | After a failed stage | Continues from that stage; finished stages kept |
| Rebuild final cut | On a finished ad | Re-joins the same shots with the current end card |
| Fix this shot in the editor | On a finished ad | Opens the editor on that shot |
| Edit end card | On a finished ad | Changes brand, name, tagline, call to action |
| Cancel this ad | Before shooting starts | Stops the ad; started steps not refunded |
| Make another | On a finished ad | Back to the composer |

## FAQ

### Does Try again redo the whole ad?

No. It continues from the stage that failed; everything finished before it is kept.

### Why does Try again show a new price?

The price of the remaining work changed since you started. Confirm to continue at the new figure.

### A shot was blocked and Try again is refused.

Edit that shot in the editor first; the retry is refused until the blocked shot has changed.

### Can I cancel during shooting?

No. Cancel is available until shooting starts. Let it finish and decide then.

### Does Rebuild final cut reshoot anything?

No. It joins the finished shots again with the current end card.

Updated November 3, 2026


---

# The AI Badge Switch: Keep or Remove the AIGC Label | SceneMixer

URL: https://scenemixer.com/guides/the-ai-badge-switch/
New videos carry a small AIGC badge in the corner. Where the switch is, what it does to new videos, why members can remove it, and what the download carries.

# The AI Badge Switch

Quick answer: Every new video carries the SceneMixer AIGC badge in its bottom-right corner unless you turn it off. The switch is a small pill beside the model picker in the editor: AI Badge means the next clip will carry it, No Badge means it will not. Removing the badge is a membership benefit; members get badge-free videos by default and can switch it back on. The choice applies only to videos generated from then on; clips you already have are unchanged, and regenerating a clip applies the current setting.

Open a finished episode and look at the bottom-right corner of a clip.

## What the badge is

A small SceneMixer AIGC mark in the bottom-right corner of the video, telling anyone who watches that the clip was generated. The panel behind the switch says it plainly: new videos carry the badge; this only affects videos generated from now on; existing clips are unchanged. Rights and labels.

## Where the switch is

In the episode editor, next to the model picker above the preview pane, as a pill that reads AI Badge or No Badge. Click it to open the panel AIGC Badge on videos, which shows a preview of the badge and the state. On a phone the same control is at the bottom of the model sheet. The account settings dialog has the same switch.

## Who can turn it off

Removing the badge is a membership benefit. SceneMixer makes a member’s videos badge-free by default; the panel says so, and offers to keep the badge if you prefer. Without a membership the panel shows Upgrade to remove the Badge and the badge stays on. What a paid plan changes.

The panel also carries a sentence worth reading before you press Remove Badge: once removed, you assume all compliance risks and legal liability for how the video is labelled wherever you publish it.

## When it takes effect

The switch is read at the moment a segment is generated. Flip it, and the next clip follows; the clips you already have do not change. To change an existing clip, regenerate it with the switch set the way you want; the earlier version stays in Version History.

## What you download

The download button gives you the full-quality file, and it carries whatever the switch said when that clip was made. SceneMixer stitches the episode composite from the clips as they are, so a mixed episode has a badge on some segments and not others. Generate them all under one setting before compositing. Original and play versions.

## Where it cannot be turned off

Some regions require AI-generated video to be labelled. On a site that operates under such a rule, the switch is locked and the panel says so; the badge stays on for everyone, members included.

| You are | Default | Can change it |
| Not a member | Badge on | No — the panel offers an upgrade |
| A member | Badge off | Yes — keep or remove, per your account |
| On a site with a labelling requirement | Badge on | No |

## FAQ

### Where is the badge on the video?

Bottom-right corner, a small SceneMixer AIGC mark.

### I turned it off, but my old clips still have it.

The switch applies to clips generated from then on. Regenerate a clip to apply the current setting.

### Can I remove the badge without a membership?

No. Badge-free video is a SceneMixer membership benefit.

### Does the composite have the badge?

The composite on SceneMixer is stitched from the clips as they are. If some clips carried the badge, those parts of the composite do too.

### Is the badge in the downloaded file?

Yes, if it was on when the clip was made. The download from SceneMixer is the full-quality file with whatever the clip carries.

Updated November 3, 2026


---

# The Asset Library: Tabs, Uploads, Downloads and the Quota | SceneMixer

URL: https://scenemixer.com/guides/the-asset-library-tabs-uploads-quota/
What the Assets page holds: the six tabs, what Uploaded by me is, how to upload (a paid-plan perk), how to download, and why only uploads count against the quota.

# The Asset Library

Quick answer: Assets in the sidebar is everything you have made or uploaded, across every project, in six tabs: Characters, Scenes, Items, Videos, Novels, Other. Each entry shows the project it belongs to; uploads that belong to no project sit under Uploaded by me. Upload asset takes JPG or PNG up to 20 MB and is part of a paid plan. Anything can be downloaded. Only uploads count against the plan’s storage quota — generated sheets and videos never do.

Where your files live is the storage side; this page is the library you see.

## The six tabs

Characters, Scenes, Items: every sheet from every project, with the project name under each; open one to see it large or download it. Videos: every segment clip and assembled episode, with a download on each, plus standalone videos. Novels: the scripts you fed in, one per project. Other: uploads that are not sheets — reference images, product photos. The count on each tab is the total.

## Uploaded by me

Images you uploaded here rather than inside a project. They are available to every project: when you replace a sheet or add a reference in the editor, the picker can take them. Using your own photo.

## Upload asset

Upload asset or drop a file onto the page: JPG or PNG, Over 20 MB is refused, Images only. Uploading is a paid-plan perk (Uploads · custom style · AI fine-tune on the plan card); on a free account the button opens the plan window. Uploaded images can be optimised into sheets from inside a project. What a paid plan changes.

The first time you upload a file, a window titled Before you upload asks you to confirm that you have the right to use the material (scripts, images, video, audio) and that every real person whose face or voice appears in it has given consent. Press Confirm and upload to go on; you only need to confirm once, several files in one go ask once, and cancelling stops the upload.

## Download and delete

Every entry has Download; videos download the original-quality file. There is no delete on this page: it is for finding and downloading. Sheets, clips and episodes are deleted inside their project (delete a segment, an episode or the whole project), which keeps an episode from losing a sheet it still uses. Downloading.

## The quota

Each plan has a storage quota, listed on the plan card. SceneMixer counts uploads only: your photos, reference images, product shots, voice samples. Generated sheets, clips and episodes are not counted, however many you make. When the quota is full, uploads are refused with Storage is full — delete some assets or upgrade your plan; deleting the projects that hold uploads frees their space.

| Tab | Holds | Counts against quota? |
| Characters / Scenes / Items | Sheets from every project | No |
| Videos | Segment clips and assembled episodes | No |
| Novels | The scripts you fed in | No |
| Other / Uploaded by me | Your uploads | Yes |

## FAQ

### Is everything from every project here?

Yes. Sheets, videos and scripts from all projects, each labelled with its project.

### Why can't I upload?

Uploading is part of a SceneMixer paid plan. The button opens the plan window on a free account.

### What counts against the storage quota?

Only uploads. Generated sheets and videos never do.

### Can I delete anything from here?

No. This page finds and downloads; deleting happens inside a project.

### How big can an upload be?

Up to 20 MB, JPG or PNG.

Updated November 3, 2026


---

# The @ Menu: Cast, Scenes, Props, Voices and Seconds | SceneMixer

URL: https://scenemixer.com/guides/the-at-menu-in-the-storyboard/
Type @ while editing a storyboard and a picker opens with five tabs. What each tab inserts, where to put it, and the two mistakes people make with voices.

# The @ Menu in the Storyboard

Quick answer: While editing a segment, type @. A picker called Insert a reference opens with five tabs: Cast, Scenes, Props, Voice and Time. Pick an entry and a chip drops in at the cursor. Cast, Scenes and Props chips tell the shot which person, place or object you mean; a Voice chip goes after a line of dialogue and says whose voice speaks it; Time inserts a shot duration, so @3s is a three-second shot. Esc closes the picker.

Open a finished episode and look at any segment: every name, voice tag and duration you see is a chip that came from this menu.

## Opening it

Click Edit Script on a segment, put the cursor where you want the reference, and type @. The picker opens right there. Start typing to filter — the search box reads Search characters, scenes, items… — then click the entry, or press Esc to close without inserting.

Only entries that exist in this project appear. If the person or thing you want is not there, add it first: in the editor, Add assets on the left panel; or in step 2 of the project.

## Cast

A character chip. Put it wherever the character is in the picture: @Mara turns from the window; In frame only: @Mara. The shot uses her image. A character who is mentioned but not visible should not get a chip in the picture description — who is in frame.

Note the hint the picker shows: a Cast chip references the image only. It does not give the character a voice. That is the next tab.

## Voice

A voice chip goes right after a line of dialogue and reads (voice: 🔊Mara). It says which character’s voice speaks the line. Every line needs one. If a line has none, the editor warns you after saving that its audio may drift — meaning the line may be spoken in a voice that is not hers.

Two mistakes to avoid: using a Cast chip where a Voice chip belongs (the picker tells you; they are different), and tagging a line with the wrong character. Writing dialogue lines.

## Scenes and Props

A Scenes chip normally sits in the Location field of the segment header — that is what tells the segment where it takes place. A Props chip goes into the Props line of the header and into the shot text where the object is handled: @Sealed letter is slid across the table. Props and locations in a shot.

## Time

The Time tab inserts a duration chip: @3s = a three-second shot. Use it to change how long a shot runs. The seconds of all shots add up to the length of the segment, and SceneMixer keeps the segment within the cap of the video model you have selected. Shot seconds and timing.

## After you save

Three warnings can appear, and none stops the save. …is not a character, scene or item in this project: you typed a name by hand instead of using a chip. …has no reference image yet: generate that entry’s image in step 2. Some dialogue lines are missing voice tags: add a Voice chip after each line. To find which line, read each line of dialogue to its end: the ones without a (voice: 🔊Name) chip after them are the ones to fix.

An amber note may also appear above the storyboard: a name written without @ in a shot, or a “do not show…” sentence about a person. And when you press Generate Video on a segment you wrote or edited, the editor asks first if some shots have no linked characters. The fix is the same each time: delete the typed name, type @ in its place and pick the entry. The amber notes · Shots with no linked characters.

| Tab | Inserts | Where it goes |
| Cast | A character chip | In the picture text, wherever the character is visible |
| Scenes | A scene chip | The Location field of the header |
| Props | A prop chip | The Props line, and the shot text where the object is handled |
| Voice | (voice: 🔊Name) | Right after a line of dialogue |
| Time | @3s and the like | On the shot whose duration you are changing |

## FAQ

### Nothing happens when I type @.

Make sure you are in edit mode (Edit Script) and the cursor is inside the storyboard text.

### The person I want is not in the picker.

They are not in this project yet. Add them with Add assets in the editor, or in step 2.

### Do I need a voice chip on every line?

Yes. A line without one may be spoken in the wrong voice, and SceneMixer will warn you.

### Can I just type the character's name?

You can, but the shot will not know which entry you mean and no image will be attached. Use the chip.

### Can I delete a chip?

Yes. Put the cursor next to it and delete, or cut and paste it elsewhere. Discard undoes everything since you started editing.

Updated November 3, 2026


---

# AI Edit's Eight Quick Requests, With Examples That Work

URL: https://scenemixer.com/guides/the-eight-quick-ai-edit-requests-with-examples/
Change a line, an action, the mood, time and weather, the camera, remove someone, add a voice-over, change the location: eight AI edit starters, with examples.

# The Eight Quick Requests in AI Edit, With Examples That Work

Quick answer: Under the AI edit box sit eight buttons: Change a line, Change an action, Change the mood, Time and weather, Change the camera, Remove someone, Add a voice-over and Change the location. Each types the start of a request into the box — “Change the line in shot 1 to:” — and you finish the sentence. They are only starters: you can edit the shot number, combine two, or write your own request from scratch. Each revision costs 1 credit.

Click a starter, change the shot number to the shot you mean, then finish the sentence. Be concrete: what, where, how.

## Change a line

Starts: Change the line in shot 1 to:. Finish with the new words, or with how the line should change. Good: “Change the line in shot 2 to: ‘You kept the key. Of course you did.’” or “…make her line in shot 2 colder and shorter”. Keep an eye on length: the review warns you when a line looks too long for its shot.

## Change an action

Starts: In shot 1, change the action to:. Describe what the person does, physically. Good: “In shot 3, change the action to: he tears the letter in half and drops it”. It changes the words of the shot, not its length.

## Change the mood

Starts: Change the mood of this segment to:. Name the feeling and, if you can, how it shows. Good: “…tense and quiet: short lines, nobody raises their voice”. It adjusts the lines, the performance notes and the actions across the segment.

## Time and weather

Starts: Change the time and weather to:. Good: “…late night, heavy rain”. It updates the header’s time and weather and the descriptions that depend on them. For a whole scene, change each of its segments so they match. For a clip that is already made, AI visuals can change the time or weather on the video itself. Day to night on a finished clip.

## Change the camera

Starts: Change the camera in shot 1 to:. Good: “…a close-up on her hands, then the face” or “…a wide shot from the doorway”. This is the way to change framing and angle — things the video fixes cannot do — followed by a regeneration.

## Remove someone

Starts: Keep this person out of the segment:. Finish with the name. The shots are rewritten without them. If they speak, their lines go too — and a revision that adds or drops too many lines of dialogue at once is dropped (free), so for someone who talks a lot, remove them in steps or by hand. Good for a character who crept into a scene they should not be in.

## Add a voice-over

Starts: In shot 1, add a voice-over line that says:. Finish with the words, and say whose voice if it matters: “…in Mara’s voice: ‘I should have left then.’”. A voice-over is heard, not spoken on screen. If you do not say who speaks, AI edit gives the line to someone already in the segment; if it ever names someone who was not, the review points that out. Lines, voice-over and captions.

## Change the location

Starts: Move this segment to this location:. Finish with a location from your project, or describe a new one. A location that is not in your project is written as plain text, without a location image; for a consistent look across segments, add it to the asset library and generate its image. Add a location to this episode.

## Writing a request that works

- Name the shot: “in shot 2”. Unclear requests are refused (free) with a note asking which shot.
- Say the result, not a process: “she answers without looking at him” beats “make it more dramatic”.
- One or two changes per request. A revision that adds or drops too many lines of dialogue at once is dropped (free); do it in steps.
- More or fewer shots, or different seconds? Ask for it in the same box — “split shot 2 into two shots”, “drop shot 4”, “make shot 3 two seconds longer”. The review shows the old and new shot count and the new total seconds before you save.
- Use names that exist in the project. Write in any language; the text keeps its own language.

| Starter | Types into the box | Finish it with |
| Change a line | Change the line in shot 1 to: | The new words, or how it should change |
| Change an action | In shot 1, change the action to: | What the person physically does |
| Change the mood | Change the mood of this segment to: | The feeling and how it shows |
| Time and weather | Change the time and weather to: | When and what weather |
| Change the camera | Change the camera in shot 1 to: | Framing and angle |
| Remove someone | Keep this person out of the segment: | The name |
| Add a voice-over | In shot 1, add a voice-over line that says: | The words, and whose voice |
| Change the location | Move this segment to this location: | A location from the project |

## FAQ

### Can I click two starters?

Yes. A starter replaces another starter's unfinished text, but if you have written something yourself it is added on a new line below.

### Do I have to use a starter?

No. They only type the beginning; any request up to 500 characters works on SceneMixer.

### Why was 'make it better' refused?

The AI could not tell what to change. Say which shot and what should be different; refusals cost nothing on SceneMixer.

### Can I add a character this way?

Only one that is already in the project. Add new ones in the asset library.

Updated November 3, 2026


---

# The AI Ad End Card: Brand, Tagline, Call to Action | SceneMixer

URL: https://scenemixer.com/guides/the-end-card-brand-tagline-cta/
Every ad ends on an end card: the main image plus four lines. Where the lines come from, how to edit them, the limits, and why the card carries your brand name.

# The End Card

Quick answer: The final cut ends on an end card: your main image plus four lines — Brand, Product name, Tagline, Call to action. The first two are read from your photo and your brief; the tagline and call to action are written with the script. Any line can be left blank. Edit them with Edit end card and Assemble to rebuild the cut; the shots are not reshot. The card uses the ad’s dialogue language, and because video models do not write legible text, the end card is where your brand name actually reaches the viewer.

The whole ad flow ends on this card.

## What is on it

The main image — the first photo you added, cleaned up — and up to four lines: Brand (the name read off your signage or package, or what you typed), Product name (the name from the reading dialog), Tagline (one line to remember it by), Call to action (Shop now, Book a table, Download the app). The colours follow the ad vibe. If brand and product name are the same words — a shop whose sign is its name — only one line is printed.

## Where the lines come from

The brand is read from the photo when a name is printed on a shopfront, a package or a car, copied letter for letter. The product name comes from the reading, corrected by you in the dialog. The tagline and call to action are written with the ad script, and the tagline is spoken in the ad before the final freeze so the card repeats what was just heard.

## Editing

On the progress page of a finished ad, Edit end card opens the four fields. Change what you like, leave what you like blank, then Assemble: the final cut is rebuilt with the new card, without reshooting. In expert mode the same fields appear when you assemble from the editor. The hint says it: changing them after assembling means assembling again.

## Length limits

Each field has a limit that depends on the language: shorter in languages counted by character (Chinese, Japanese, Korean), longer in languages counted by word. The field stops you at the limit and the preview shows how the line fits. A tagline that does not fit is better cut than squeezed.

## Language

The card is in the ad’s dialogue language, the one shown as Dialogue & end card in the composer. A few scripts cannot be drawn on the card yet; the composer warns that such a line will be left out or shown in English. Language in the two modes.

## Why the card matters

Video models do not write legible text, and ads are made without on-screen text for that reason. The shop sign in your photo is read, but it is not something the shots can reproduce. The end card is the one place the name, the tagline and the call to action are delivered as readable text — treat it as the line that closes the deal. Signs and photos.

| Line | Comes from | Editable | Blank allowed |
| Brand | Signage or package in the photo, or your brief | Yes | Yes |
| Product name | The reading dialog | Yes | Yes |
| Tagline | Written with the script; spoken in the ad | Yes | Yes |
| Call to action | Written with the script | Yes | Yes |

## FAQ

### Can I change the end card after the ad is done?

Yes. Edit end card, then Assemble. The cut is rebuilt; nothing is reshot.

### Why is only one line printed for my shop?

When the brand and the product name are the same words, one line is printed instead of two.

### Why did my long tagline get cut?

Each line on SceneMixer has a length limit that depends on the language. Shorten it; the preview shows the fit.

### Which language is the end card in?

The ad's dialogue language. Some scripts cannot be drawn yet and fall back to English or are left out; the composer warns you.

### Can the shop name appear in the shots?

Not reliably; generated video does not write legible text. The end card is where the name is delivered.

Updated November 3, 2026


---

# The Episode Editor, on Desktop and Phone: A Tour | SceneMixer

URL: https://scenemixer.com/guides/the-episode-editor-a-tour/
Where everything is in the editor: assets, the storyboard and its shot list, the preview and Generate, version history, the timeline, shortcuts and the phone tabs.

# The Episode Editor: A Tour

Quick answer: The editor is three columns over a timeline. Left: the assets this episode uses — characters, scenes, items. Middle: the selected segment’s storyboard, readable as a Shots list or as plain Text, with Edit, AI edit, History and the black Generate button. Right: the preview player, then Version History, share and download, and Edit this clip (AI visuals / Redub / Trim / Enhance). Bottom: the timeline — one card per segment, the free-preview card, and the composite button. The top bar carries the model picker, the AI badge switch, Generate all and Composite Episode. On a phone the panes become tabs — Preview, Script, Assets — plus a model tab.

Open a finished episode: this is the editor with every segment done.

## Getting in

From step 3, press Edit on an episode that has a storyboard (an episode without one asks you to parse it first). The top bar holds the back arrow and the episode title on the left; the model picker, the 3/12 done counter, Generate all and the composite button in the middle; your account menu on the right.

## The page at a glance

Three columns and a timeline underneath. The assets panel takes the left, the storyboard the middle, the preview the right, and the timeline runs along the bottom. For a portrait (9:16) project the right column runs from the top of the page to the bottom and the timeline sits under the middle column only; for a landscape project the timeline runs the full width and the side columns stop above it. Layout for vertical and landscape projects.

| Area | Holds | Main controls |
| Left | This episode’s characters, scenes and items | Add assets, details, Edit (this episode), Show all, Collapse |
| Middle | The selected segment’s storyboard | Segment picker, Shots / Text, Edit, AI edit, History, Insert, Delete, Generate |
| Right | The segment’s clip or its state | Player, Play continuously, Music, Share, Download, Edit this clip, Version History |
| Bottom | Every segment, in order | Select, scroll arrows, free preview, + segment |

## Left: the assets this episode uses

Characters sit at the top as portrait cards; scenes and items follow as thumbnail rows. Click any of them for its details: the appearance text, and for a character the voice section (Auto-match voice, Change, Re-match, Pick from library, Upload Voice). Edit (this episode) opens an editor whose changes apply to this episode only — the project’s step 2 entries stay as they are. The + at the top right of the panel is Add assets: it lists everything in the project, and entries already in the episode are marked Added and can’t be picked twice. When a group is long, Show all (N) reveals the rest and turns into Collapse. Adding assets · Character voices.

## Middle: the storyboard

The top strip is the segment picker: a window of numbered pills with ‹ › arrows on either side. The highlighted number is the segment you’re looking at; a small violet dot under a number means that segment already has a clip. On the right of the same strip sit the length, the price and the status pill — Pending, Done, Generating or Failed.

### Reading: Shots or Text

When the text reads as a storyboard, a small toggle at the top right of the pane switches between two views of the same content. Shots lays it out as a list: each shot with its number and seconds, the action as lines, and every line of dialogue in its own block with the speaker, the delivery hint and the voice. Text is the exact text as stored. Both are reading views — editing always opens the text — and copying from either keeps the original structure, so a paste back into the editor doesn’t lose the shot markers. How dialogue is written.

### Editing by hand

Edit turns the pane into an editor. Cancel and Save sit at the bottom with a character count, and by the toolbar a running total shows the shot seconds you have written against this segment’s cap — it turns red if you go over. The @ pill (or typing @) opens the reference picker: Cast, Scenes, Props, Voice and Time tabs; picking an entry inserts it at the cursor, and the Time tab inserts shot lengths like @3s. If you click another segment with unsaved changes, a dialog asks Save & Switch, Discard or Cancel. After a save, amber notes appear at the top of the pane when something needs a look. Editing the storyboard · The @ menu · The amber notes.

### AI edit and text history

AI edit takes one sentence about what to change — wording, action, camera, sound, or the shots themselves: adding, removing or merging shots and changing their seconds. When the shots or timing change, the review shows the old and new count and seconds; you review the changes before saving, and a failed request isn’t charged. On a segment that has no standard storyboard yet — blank, or written in plain prose — the same button says AI write and turns your sentence into a standard storyboard. History keeps the original text and the five most recent changes: view a version, see how it differs from the current text, and restore it. AI edit · Eight quick AI edit requests · Text history.

### Insert and delete

Insert adds a blank segment right after this one — free; write its script and generate it like any other. Delete removes the segment after a confirmation that says what goes with it: its clip and version history. SceneMixer does not refund credits already spent, and a segment that is generating can’t be deleted. Insert or delete a segment.

## Right: the preview

The player shows the selected segment’s clip, or the state it is in. It has its own controls along the bottom (play, seek, sound, full screen).

Below it, one row carries the everyday switches. Play continuously (on by default): when a segment ends, the next one plays on its own — segments without a clip are skipped, and their numbers are named under the player. Music appears when this episode has background music: on means play and download with the music, off means the dialogue-enhanced version without it. A small pill shows the active version and its length (v1 · 13.5s). Then Share and Download, greyed while the segment has no clip. Background music · Downloads.

### Edit this clip

Under the switches sits Edit this clip with four tabs — AI visuals, Redub, Trim, Enhance — and it opens straight on Trim for a clip that exists.

- Trim cuts the start or the end. Keep / Remove decides whether the range you mark stays or goes; drag either end of the bar, or nudge with the small − / + steppers (half a second a press). Preview plays just the kept part without submitting anything, and Save makes it a new version — the price is on the button. A clip that is too short says so instead.
- AI visuals changes a detail of the picture — a colour, the time of day, the weather, a scar, an expression, swapping one person for another. On a clip over fifteen seconds it asks which half to work on and stitches the result back automatically. Framing, camera and where people stand can’t be changed this way; those come from the storyboard.
- Redub re-records the talking without touching the storyboard. It lists this segment’s lines, lets you pick a voice for each, and offers two modes: Lip sync (the picture stays, only the voice changes) or Re-perform (the whole clip is re-acted, so the picture may change). A segment with no dialogue says so.

One change per clip at a time: while one is running, the other tabs answer “another change to this clip is still running”. A clip made by an older pipeline has to be regenerated before it can be edited this way. Edit this clip, tab by tab · Trim · Redub · Lip sync or re-perform · Enhance.

### Version History

Once a segment has more than one take, its versions are listed below, newest first, each tagged with where it came from — Generated, Trimmed, Reused, AI edit, Redubbed, Enhanced — and a thumbnail. Click one to make it active; the active one is marked Active. Switching is free and instant: the next playback, download and composite use the new version. Versions.

### Before a clip exists

The box is never empty. No clip yet: No video generated with the Generate button and its price. Generating: Generating video with a time estimate built from past generations, so it tracks the real average rather than a fixed promise. Failed: a plain-language reason, Retry, and a link that carries the message into the help assistant. What each status means.

## Bottom: the timeline

The header row counts the episode (12 segments · 96s), reminds you of the arrow keys, and draws one small dot per segment: violet done, orange pulsing generating, red failed, grey not yet. Then one card per segment with its number, a thumbnail once there is a clip, its length, and a status dot; click a card to select it. When the cards overflow the strip, small arrows appear at either end to scroll it.

Two special cards bookend the strip. First a dashed violet card — the account’s free five-second preview; click it to watch it in the preview pane, or to make one if you haven’t yet. Last a + card that appends a blank segment, same as Insert. Timeline icons · The free preview.

## Generating video

Pick the model in the top bar first — Select Model lists each option with its resolution and its length cap, and options whose cap is shorter than this segment are greyed out with the reason. Then press the black Generate button, which carries the segment’s price. Each segment is charged on its own, and a failed generation is refunded. Generating a segment · Choosing a tier.

Generate all in the top bar (it appears while some segments have no clip) opens a dialog that states the plan before you commit: how many segments, their total seconds, the model, which segments already have a clip and stay as they are, and any segments longer than the model allows that are left out. It notes that every segment is charged on its own, refunded if it fails, and that segments beyond your account’s concurrency limit wait in line and start automatically. The dialog submits them one by one and can be stopped part-way with Stop here. Generate all.

## Waiting: the tab title and notifications

While anything in this episode is generating, the browser tab’s title gains a quiet prefix: (3/12) generating. You can leave the page, watch something else, or close the laptop — the work continues on the server, and the state is there when you come back.

On a desktop browser, the first time you press a Generate button the site asks once for permission to send notifications. If you allow it and switch to another tab, a segment finishing raises a small system notification — Segment 4 is ready, and clicking it brings you back to the editor. Phones don’t get this prompt; the tab title still tracks progress. Why a generation takes time.

## Composite and download

Composite joins every segment into one episode. It stays disabled until all of them are done — the button shows the count, like Composite (5/12), and explains what is missing. While it runs, the button becomes Compositing… and opens the history; once a composite exists it becomes View Composite.

Composite History lists every composite: its date, length, credits and status. Play it right there, Composite Again if you have changed segments since, Share or Download it — and set the Music switch for the whole episode, which decides whether the played and downloaded file carries the music. A failed composite shows its reason instead of a file. Composite and share · Which segments are joined.

## Keyboard

- ↑ / ↓ — previous / next segment, from anywhere on the page. Not while you are typing, and not while the AI edit panel holds an unsaved result.
- In full-screen playback: Space or K play / pause, ← / → back / forward five seconds, M mute, F full screen.
- Esc — closes the @ picker and any open panel.

## On a phone

The editor fits the phone as tabs along the bottom: Preview, Script, Assets, and a fourth tab showing the current model, which opens the model sheet (the AI badge switch lives inside it). Above the tabs sit Generate all and the composite button, with the reason written out under it when composite is disabled. The same account, the same data, the same clip — fine editing of storyboard text is the one thing the editor itself tells you to do on a desktop: Viewing and generating work on mobile; for detailed storyboard editing, use a desktop.

- Preview tab — the player filling the width, with version, Play continuously, Music, full screen, Share and Download on the video; below it the Edit this clip row (Trim opens inline under the player, AI visuals, Redub and Enhance open as sheets), then the timeline as a horizontal strip.
- Script tab — segment chips, the meta pills, the Shots / Text toggle and the text; Edit, AI edit and Generate in the first button row, Insert, Delete and History in the second. Editing opens a full-screen sheet with the same @ picker.
- Assets tab — the same library, with details and per-episode edits in sheets.

For a landscape project, the preview tab shows this segment’s shot list under the player, so you can glance at it while you watch.

## Ads

An ad in advanced mode opens the same editor with ad wording: shots instead of episodes, Product & props instead of items, Assemble ad instead of composite, and an end-card step before assembling. Advanced mode.

## FAQ

### Where is the Generate button?

Under the storyboard in the middle column, and again in the preview box while a segment has no clip. Both carry the segment’s price.

### Where do I change the video model?

The model picker in the top bar, labelled Select Model. On a phone it is the last tab at the bottom.

### Can I edit the storyboard on my phone?

Yes — the Script tab opens the same text in a full-screen editor. Fine editing is easier on a desktop.

### How do I go back to the episode list?

The back arrow in the top bar: Back to episodes.

### Where are a segment’s older versions?

Version History in the right column, once a segment has more than one take. On a phone it is behind the version pill on the player.

### What is the difference between Shots and Text?

Two reading views of the same storyboard: Shots lays out each shot with its seconds and its dialogue, Text is the exact text as stored. Editing always opens the text, and copying from either keeps the original structure.

### Can I leave the page while segments generate?

Yes. Generations run on the server. The tab title shows quiet progress, and on a desktop browser a finished segment can raise a notification if you allowed it.

### Why is Generate greyed out?

Either this segment is already generating, or a submission is still in flight. Wait for it to finish; the button comes back on its own.

### Do I pay again when I make changes?

On SceneMixer, each paid step is charged on its own: generating a segment, trimming, AI visuals, redub, compositing. Reading, playing, switching versions and downloading are free, and a failed generation is refunded.

### Does everything have to be generated before I can composite?

Yes. Composite stays disabled until every segment has a clip; the button and the phone bar say how many are done.

Updated November 3, 2026


---

# The Free Preview: A Few Seconds, Once — What It Is and Isn't

URL: https://scenemixer.com/guides/the-free-preview-a-few-seconds-once/
The free preview (the trial or test video) generates one shot of your episode, a few seconds long, once per account. Which shot, where the button is, how to skip it.

# The Free Preview: Why It Is Only a Few Seconds, Once, and How It Differs From Generating

Quick answer: When you open a parsed episode’s storyboard, the editor offers a free preview: it generates one shot of the episode with your own characters and script, preferring a shot where someone speaks, about 5 seconds long, at the lowest 480P tier, for nothing. It is a taste of what a finished shot looks like with your cast, not a cut of your project: one account gets one preview, and the full episode is generated segment by segment at the tier and length you choose. Until you use it or choose Don’t show this again, the offer pops up when you open a parsed episode; a preview card also stays at the start of the timeline, and once used it plays your preview.

Open the editor of a parsed project; the preview card sits at the start of the timeline, before the first segment; once you have used your preview, the same card plays it.

## What the preview is

The preview is a real generation, made from your own character sheets and one shot of the episode, on the cheapest 480P tier that is switched on. It picks a short shot where someone starts talking, so you hear your cast as well as see them; when the episode has no such shot, it uses the opening shot of the first segment. It exists so that you see a finished shot with your cast before spending credits. It is not a shortened version of your episode and not a sample of every segment: one shot, a few seconds, once. The clip lands in the preview pane like any other and can be shared; downloads come with the segments you generate.

## Why only a few seconds, and only once

Video is priced per second and the preview is free, so SceneMixer caps it at 5 seconds and at one per account — not one per project. The full run of an episode is the paid part: each segment on the timeline generates separately at the tier you pick, for the seconds you set, and the price tag on each card tells you the cost before you press. Generating a Video Segment.

## Where the button is

Until you use it, the offer pops up as a card each time you open a parsed episode. Generate free preview starts it; Maybe later closes the card without using the preview and leaves a preview card at the start of the timeline, before the first segment; Don’t show this again hides the pop-up but keeps that card. After you use it, the timeline card stays and plays your preview. On the phone the entry is a button in the top bar. There is no preview entry when no 480P tier is available for it.

## If the preview looks wrong

SceneMixer makes it from that shot and the sheets as they are: a poor sheet or a thin shot shows up here first. That is the point — fix the sheet in Step 2 or that segment’s text, then generate the real segment. The preview is not regenerated. When a Shot Comes Out Wrong.

## The rest of the project

Nothing else is generated for free on the video side. Your account’s 15 free sheets (on projects whose parse you paid for) and the first project’s first storyboard are the other allowances; after that, every segment costs its tier’s price per second (10 credits a second on the default tier). What SceneMixer Gives You for Free.

| | Free preview | Generating a segment |
| Length | About 5 seconds | The seconds you set, up to the tier's cap |
| Content | One shot, preferably with dialogue | Any segment of the storyboard |
| Tier | Lowest 480P tier | Your choice in the selector |
| Cost | Free, once per account | Per second of the tier |
| Download | Share only | Original and play versions |

## FAQ

### Why did I only get a few seconds?

That was SceneMixer's free preview, not your project. Generate the segments on the timeline for the full episode.

### Can I preview again in a new project?

No. SceneMixer's preview is once per account.

### I clicked Later. Did I lose it?

No. The preview card stays at the start of SceneMixer's timeline; after you use it, the same card plays your preview.

### Can I choose which shot is previewed?

No. SceneMixer picks a shot from the episode you are in, one with dialogue when there is one, otherwise the opening shot.

### I pressed a button and it asked for credits. Was that the preview?

No. The preview card sits before the first segment and is marked Free preview. The Generate button on SceneMixer's segment card is a normal, paid generation.

### My preview will not play.

Click the preview card again to load it in the player. If it still does not play, the steps in the video-will-not-play guide apply to it too.

### Does the preview use my credits?

No. It is free on SceneMixer and holds nothing.

Updated November 3, 2026


---

# The Notification Bell and the Announcements Page

URL: https://scenemixer.com/guides/the-notification-bell-and-the-announcements-page/
What arrives in the bell at the top right, what the red number means, where each kind of message comes from, how to reply from the bell, how read and unread work, what the announcements page shows, and whether any of it is emailed.

# The Notification Bell and the Announcements Page: What Arrives, and What You Can Do About It

Quick answer: the bell at the top right holds two different kinds of message. Announcements go to everyone: new features, improvements, updates, maintenance notices and promotions. Replies meant only for you arrive the same way: someone answered a feedback you sent, or answered a question you asked in the Ask window. The red number on the bell counts everything you have not opened yet. Clicking an announcement takes you to the announcements page; clicking a reply opens the whole answer right there, with a link to your feedback history or to continue the question. Opening a message is what marks it read, and pressing the small cross on a row counts it as read too; nothing else does.

## Where the bell is

The bell sits in the top bar of every page, on the right-hand group of icons: next to your credits (the ✦ number) and just to the left of your round avatar. On a wider screen it comes after the Ask button and the settings gear; on a narrow one it is simply one of the three things the top bar keeps — credits, bell, avatar. The episode editor has the same top bar, so you can check it while you are working on a storyboard.

You have to be signed in. There is no bell for visitors, and the announcements page asks you to sign in before it shows anything. Everything in the bell belongs to the account you are signed in with — on another device, or in another browser, the same messages are there with the same read state.

## What the red number means

The small red badge on SceneMixer's notification bell is the number of messages you have not opened. It counts everything published for you: the announcements that go to everyone, plus the replies written to you. It shows only when the number is above zero; read everything and the badge disappears.

Two things about the number are worth knowing. First, it is not limited the way the list is — the bell shows only the most recent messages, while the number counts every unread one, so after a quiet week the number can be larger than the rows you can see. Second, it refreshes about every half minute while you have a page open. The bell gives a short ring when the number goes up, and while unread messages remain it repeats roughly every minute and a half; that is a reminder, not an alarm.

## What can arrive

| The label you see | What it is | Where it comes from | What you can do |
| New | A newly added feature | An announcement written by the site's team and published to everyone | Read it; the detail view opens from the announcements page |
| Improved | Something that got better or was fixed | Same | Same |
| Update | A change worth knowing about | Same | Same |
| Maintenance | A period when something is being worked on | Same | Same |
| Promo | An offer or campaign | Same | Its detail has a Top up button that opens the top-up page |
| Feedback | “Your feedback got a reply” | Someone from the team answered a feedback you submitted | Read the answer in the bell; History opens your feedback records |
| Help | “Your question got a reply” | You asked something in the Ask window, the answer did not settle it, and a person wrote back | Read the answer next to your own question; Still not solved? Tell us more continues the same conversation |

SceneMixer publishes announcements in the interface language you are using. If no translation exists yet for that language, you see the original wording rather than an empty message. Switching the interface language switches announcements to that language too, the next time the list loads.

Two of the labels are coloured the same way on purpose: Feedback and Help are both answers to you personally, while the other five are announcements. Only the last two behave differently when you click them.

## Reading messages in the bell

- Click the bell. A panel opens with Notifications at the top, View all in the corner, and a small cross to close it. An empty list says No notifications.
- Read the rows. Each row has a purple dot if it is unread, a title, a coloured label, up to two lines of the message, and the date it was published. Rows you have already read look dimmer.
- Announcement rows take you to the announcements page. Clicking one marks it read and leaves the bell; the announcements page opens on its list. That message is not opened for you there — click its card on the page to read it in full.
- Feedback and Help rows open in place. Clicking one marks it read and expands the whole answer inside the panel — the panel stays open and nothing else on the page moves. For a Help reply, your original question is shown first under Your question, and the answer under Official Reply.
- Read more for a long answer. The bell shows about five lines of a reply. If the text is cut off, a Read more link appears and opens that one message on the announcements page in a wider window, where the links inside the text can be clicked.
- Answer from there. Under a feedback reply, History opens your feedback records so you can see the whole exchange. Under a question reply, Still not solved? Tell us more reopens the Ask window on the same conversation, carrying over what was asked, so you can push further without starting over.
- The small cross on the right of a row hides that message from the bell for the rest of this visit and counts it as read. It is not deleted: the announcements page still lists it, and it is back in the bell after you reload.

## Read and unread, exactly

Three actions mark a message read on SceneMixer: clicking its row in the bell, clicking its card on the announcements page, and pressing the small cross on a row in the bell. Hovering, scrolling the list, opening the panel and closing it again do not. There is no “mark all as read” button, so if you want the badge to clear, open the messages you care about — or press the cross on the ones you do not.

Where you see the difference: in the bell, an unread row has a purple dot on the left and full contrast, a read one is dimmed and has no dot. On the announcements page, an unread message has a purple dot on the date line and a New label at the right of its card; a read one has a grey dot and no label. Read state is kept on your account, so opening a message on your phone clears it on your desktop too.

## The announcements page

Open it with View all at the top of SceneMixer's bell panel. It is also a page of its own: once you are signed in, /announcements goes straight there.

The page shows the same messages as the bell, newest first, on a dated timeline. It shows the most recent ones rather than everything ever published. When more than one kind of message is on the page, a row of filter buttons appears above the list — All, followed by one button per label present on the page; press one to see only that kind, press All to come back.

Clicking a card opens it in a popup: the label, the full date, the title and the whole text, with links you can click. Closing it is the small cross in the corner, the Esc key, or a click outside the popup. A Promo announcement has a footer with a Top up button that opens the top-up page. A Feedback or Help announcement has a footer link instead — History or Still not solved? Tell us more — which does the same thing as the link in the bell. With nothing to show, the page says No announcements.

## Does any of this arrive by email?

Announcements do not. They live in the bell and on the announcements page, and nothing about them is emailed; the unsubscribe link in SceneMixer's emails does not apply to them either.

A reply to your feedback is different: it always appears in the bell and in your feedback records, and an email copy is sent only if, at the moment you sent the feedback, you chose an address to be written to — your account address or another one. If you chose Just show the reply here, there is no email. A reply to a question you asked in the Ask window stays in the bell and is never emailed. The emails we send, and how to stop the reminders lists everything that does arrive by mail.

## What each part is not

| You want | Where to look |
| A reply to a feedback you sent | Bell, label Feedback — and your feedback records via History |
| A person answering a question in the Ask window | Bell, label Help; the same conversation reopens from the link under the answer |
| News about features, maintenance and offers | Bell, then View all for the full list and the filter buttons |
| How a feature is used | The guides — the Ask window answers from them, with links |
| Progress of a running generation | The project and episode cards, not the bell |
| Something about your account that needs action | Email, when it is one of the things we email about |

## If something does not look right

The bell is empty but you expected something. Check that you are signed in with the account the message was sent to; the bell on SceneMixer only works while signed in, and the panel shows the most recent messages, not a history. The announcements page shows the same most recent messages — it is not a longer archive, so anything older than the list is not there either.

The red number will not go away. The number counts every unread message, while the bell and the announcements page both show only the most recent ones. Open — or dismiss with the cross — the unread rows the list shows and the count drops with each. Any unread messages older than the list are shown on neither page, and the number still counts them.

A reply you read yesterday is gone. A person can withdraw a reply they sent. When that happens the message is removed from the bell and from the announcements page, and it no longer counts as unread.

The quote of your question has vanished from a Help reply. Help conversations are kept for a limited time. After the original is cleared, the reply still shows the answer; only the quotation of what you asked disappears.

You are waiting on a video and nothing rings. The bell is not a progress display — a running generation shows its state on the project and episode cards. What each task status means explains those.

## FAQ

### Do I need to be signed in to see the bell?

Yes. On SceneMixer, the bell only exists once you are signed in, and the announcements page asks you to sign in too. There is no bell for guests.

### Why is the number on the bell bigger than the number of messages I can see?

The number on SceneMixer counts every message you have not opened, including older ones. The bell and the announcements page both show only the most recent ones, so when a lot arrived while you were away, some of them are older than the list.

### Does opening the bell mark everything as read?

No. Only the message you click is marked read. SceneMixer has no mark-all-as-read button.

### I pressed the small cross on a row. Will it come back?

Yes, after you reload the page. The cross only hides a message from the SceneMixer bell for that visit and counts it as read; the announcements page still shows it.

### Do announcements arrive by email?

No. Announcements live in the SceneMixer bell and on the announcements page only. A reply to your feedback is emailed only if, when you sent the feedback, you chose an email address to be written to.

### Where does a reply to my feedback appear?

Under the label Feedback, a reply shows in SceneMixer's bell and in your feedback history. Clicking the row in the bell opens the whole answer there; the link underneath opens your feedback records.

### A reply I read yesterday is gone. What happened?

A person can withdraw a reply they sent. When that happens SceneMixer removes the message from the bell and from the announcements page, read state included.

### The quote of my question has disappeared from the reply. Why?

Help conversations are kept for a limited time. Once the original is cleared, the reply still shows the answer but no longer quotes what you asked.

### Is the bell different on a phone?

The bell and its panel behave the same. On a phone SceneMixer's top bar keeps only your credits, the bell and your avatar, and the message list scrolls inside the panel.

### Does the bell tell me when a video is finished?

No. SceneMixer shows generation progress on the project and episode cards, not in the bell. The bell carries announcements and replies addressed to you.

Updated November 3, 2026


---

# The Affiliate Program: Your Link, Commission and Payouts | SceneMixer

URL: https://scenemixer.com/guides/the-referral-program/
How the SceneMixer affiliate program works: your link and code, changing your code, what each payment earns you, partner cash rates, payouts, and refunds.

# The Affiliate Program

Quick answer: Open Invite friends in the top navigation and copy your link, or share your code: people can type it when they sign up. Anyone who signs up within 60 days of clicking your link, or with your code, is bound to you, once and for good. For 12 months after they sign up, every payment they make — credit top-ups and membership plans alike — earns you 20% of what they paid, as credits, the moment the payment clears. Creators with an audience can apply to the Affiliate program and earn 15–25% in cash instead. If a payment is refunded or charged back, the matching commission is taken back.

Reading your credit history shows the Referral reward lines.

## Where it is

Invite friends in the top navigation; on a phone it is in the menu. The page shows your rate in large type, Your link with Copy link, the Opens on choices, a row of Share to buttons, Performance with Clicks, Sign-ups and Paying, Credits earned, and Earnings line by line. If the page says Referrals are not available right now, the program is switched off for the moment. The program for creators with an audience has a page of its own: Affiliate program, linked in the footer of every page and at the bottom of Invite friends.

## How the link works

Under Opens on, pick where the link should land: Home, Pricing, Samples, Guides, or Custom for any other page on this site. Then copy it, or use a Share to button. Someone who opens the link and creates an account within 60 days is bound to you at sign-up. Each SceneMixer account can have only one referrer, and it never changes, so an existing account cannot be claimed later.

## Your code

Your code is shown next to the link, after Code. It is the same code your link carries, so people who never click a link can still be bound to you: when they create an account, they tap Have a referral code? and type it in. That makes it the one to say out loud in a video, a stream or a podcast. Typing is forgiving: lower case and spaces are fine. If someone has opened several referral links before signing up, the most recent one carrying a real code counts, and a code typed at sign-up counts over any link.

## Changing your code

Your code starts as a random one we pick for you. To make it easier to remember, select Change next to it and type your own: capital letters A–Z and numbers only, 4 to 20 characters. You can change it once every 30 days. Your old code keeps working for you, so links and videos you already shared still count, and it is never given to anyone else. Codes that look like ours or a payment provider’s, such as SceneMixer, Official, Support or PayPal, and codes that match an existing promo code cannot be used.

## What you earn and when

20% of the amount your referral actually pays, on every payment they make: credit top-ups and membership plans, the first payment and the renewals. It is paid in credits — every $1 of commission is worth 100 credits — and it lands the moment the payment clears, as a Referral reward line in your Credit History and a row in Earnings that shows what they paid, the rate and what you earned. Payments count for 12 months after the person signs up.

## The cap on credits

When you are paid in credits, a single referral can earn you at most the number of credits stated under Rules on SceneMixer's referral page. When a person reaches it, their later payments show (capped) in Earnings and earn nothing more; your other referrals are unaffected.

## Partners: cash instead of credits

If you have an audience — a channel, a blog, a community — open the Affiliate program page and select Apply. The form asks where you promote, how many followers you have on your biggest platform, who your audience is, and where payouts should go: PayPal or Wise. Every application is reviewed by hand, and the result shows on that page. Once you are approved, each payment earns you cash at a rate set by the size of your audience:

| Audience | Cash commission |
| Under 10,000 followers | 15% of each payment |
| 10,000 to 100,000 followers | 20% of each payment |
| Over 100,000 followers | 25% of each payment |

The rate that applies to you is shown in large type on your partner dashboard: Partner dashboard in your account menu, or the same Affiliate program page once you are signed in.

Partners also get a promo code for their audience: 300 bonus credits on a credit top-up of $1.49 or more, once per person. Create it yourself in one click under Promo code for your audience on your partner dashboard; it uses the same name as your referral code, so there is only one word to remember. Your viewers type it when they create an account, which binds them to you, and again into Promo code (optional) when they buy credits on the Billing page (see buying credits step by step).

A cash commission becomes withdrawable 30 days after the payment. Payouts go out once a month, once your withdrawable balance reaches $50. You can switch between cash and credits whenever you like; the switch applies to new orders only. You can also convert a withdrawable cash balance into credits, which cannot be undone.

Partners tell their audience that they earn a commission, and do not bid on our brand name in search ads, send spam or make misleading claims.

## Refunds and chargebacks

If a payment is refunded or charged back, SceneMixer takes back the commission it produced in the same proportion. For credits that is a Referral reward reversed line in your history, and the row in Earnings is marked clawed back; if that leaves your balance short, later commission covers it first. For cash, a commission that has not been paid out yet is reduced, and one that has already been paid out appears as a Refund adjustment row and comes off your next payout.

## What does not count

Sign-ups alone earn nothing; only payments do. Referring yourself does not count: commission from an account created on your own device is held for review, and commission from your own extra accounts is voided.

| | Everyone | Partners |
| How to join | Open Invite friends | Apply on the Affiliate program page; reviewed by hand |
| Paid in | Credits | Cash, or credits if you prefer |
| Rate | 20% of each payment | 15%–25% of each payment, by audience size |
| Your code | Typed at sign-up; change it once every 30 days | Typed at sign-up; change it once every 30 days |
| For your audience | — | A promo code you create yourself: 300 bonus credits on a top-up of $1.49 or more |
| What counts | Credit top-ups and membership plans | Credit top-ups and membership plans |
| For how long | 12 months after sign-up | 12 months after sign-up |
| When | The moment the payment clears | Withdrawable after 30 days, paid monthly from $50 |
| Refund | The matching commission is taken back | The matching commission is taken back |

## FAQ

### Where is my referral link?

Invite friends in the top navigation; the link is at the top of SceneMixer's Invite friends page with Copy link. Partners find it on their Partner dashboard.

### Can people use my code without a link?

Yes. When they create a SceneMixer account they tap Have a referral code? and type your code.

### Can I change my code?

Yes, once every 30 days on SceneMixer: capital letters A–Z and numbers, 4 to 20 characters. Your old code keeps working for you and is never given to anyone else.

### What do I earn?

SceneMixer pays 20% of every payment the people you referred make in the 12 months after they sign up, paid in credits. Approved partners earn cash instead: 15% under 10,000 followers, 20% from 10,000 to 100,000, and 25% over 100,000.

### Do membership plans count?

Yes. Credit top-ups and membership payments both count on SceneMixer, renewals included.

### When does it arrive?

Credits land in your SceneMixer account the moment the payment clears. Cash becomes withdrawable 30 days after the payment and is paid out monthly from $50.

### How do I get paid in cash?

Apply on SceneMixer's Affiliate program page, linked in the footer of every page. Applications are reviewed by hand, and payouts go to PayPal or Wise.

### Is there a perk for my audience?

Yes. Once approved, create a promo code for SceneMixer in one click on your partner dashboard: 300 bonus credits on a credit top-up of $1.49 or more, once per person, typed into Promo code (optional) when buying credits.

### Does a sign-up alone earn anything?

No. Only payments do.

### What if they refund?

SceneMixer takes back the commission from that payment.

Updated November 3, 2026


---

# The Segment Header, Field by Field: Newer and Older Layouts

URL: https://scenemixer.com/guides/the-segment-header-fields/
Two header layouts: newer storyboards show Location and a Time and Weather line you can change; older ones have six editable fields. How to tell which you have.

# The segment header, field by field

Quick answer: There are two layouts. Storyboards written or regenerated since late October 2026 have a header with just Location and one line, Time: Dusk · Weather: Rain (from the script), with a Change link. The site reads the time and weather from your script once for the scene, and keeps track of the room, who is in it, props and crowds itself. Older storyboards, and ad and remake projects, have six editable fields: Location, Time, Weather, Opening frame, Set dressing and Props. Open a segment and look: if you see Time: and Weather: as one line with Change, it is the newer layout.

Open a finished episode and read the header of any segment. What the system tracks in the newer layout: between segments.

## Which layout do I have?

- Newer: under Location there is one line, Time: … · Weather: …, followed by (from the script) or (changed by you) and a Change link. There are no Opening frame, Set dressing or Props lines in the text, and the shots do not carry In frame only or Ambient lines.
- Older: the header text has Time, Weather, Opening frame, Set dressing and Props as lines you edit like any other text.

On SceneMixer, storyboards written from October 21, 2026 use the newer layout in new projects, and from October 23 in older projects too, whenever an episode’s storyboard is written or regenerated. Storyboards written before that keep the older layout until you regenerate them, and ad and remake projects always use the older one. A project can have both at once, one episode each.

## The newer header: Location, then Time and Weather

Location is a scene chip from the project’s Scenes, the one place the whole segment happens in. To change it, put the cursor in the field, type @ and pick another scene. Two places means two segments: a shot that cuts to somewhere else is stopped before it is generated. One place, one timeline.

Time · Weather is one line under it, for example Time: Dusk · Weather: Rain (from the script). SceneMixer reads it once for the whole scene from the scene heading and the script, so every segment in that scene starts from the same sky. A dash means the script did not say. In the text view the line sits in a small box above the text.

## Changing time and weather

- Press Change next to the line. Two menus open.
- Time: From the script, Dawn, Morning, Noon, Afternoon, Daytime, Dusk, Night, Late night. Weather: From the script, Clear, Cloudy, Overcast, Rain, Storm, Snow, Fog, Windy, Sandstorm. Leave a menu on From the script to keep the script’s value.
- If the scene runs over several segments the box says Applies to all 3 segments of this scene — the change is made for all of them at once.
- Press Save. The line now says (changed by you).

The box adds: Used the next time you generate the video; videos already made stay as they are. Changing it on SceneMixer is free and rewrites nothing in your storyboard text; it changes the light and the weather in the next video you generate for those segments, so regenerate the ones you want to see differently. To make a whole scene night, change the time once here rather than editing every segment. For a single finished clip, use Edit this clip instead: change day to night in a finished clip.

## What happened to the other fields

| Older field | In the newer layout |
| Opening frame | Not written. Who is where when a segment starts comes from the story so far. What carries over. |
| Set dressing (the room, the light, Background people:) | Not written. The room comes from the scene image and the 360° room; the light follows the time. Whether a crowd is there is decided from your shots, so describe the crowd (or its absence) in the shot. |
| Props | No Props line. A prop appears when a shot names it with @ (Props tab), and stays in the hand it is in until the text says otherwise. Props and locations. |
| In frame only: in each shot | Not written. Each shot’s cast is worked out and shown under the text as In frame. Notes under a segment. |
| Ambient: in each shot | Not written. The sound of the place follows from the scene and the weather; a sound that matters, such as a dog barking, goes into the shot’s action. |

You can still edit the shots, add and delete shots, and use AI edit as before. SceneMixer's layout only changes which lines are in the header.

## The older header: six fields

If your segment has the older layout, the header is text like the rest of the storyboard: click Edit Script, change it, Save. Free.

- Location — a scene chip; the one place the segment happens in.
- Time — time of day, and the moment if it matters: Night, active air raid. It sets the light for every shot. If the next segment is later, change its Time and the light follows.
- Weather — plain words: Clear, Falling ash and smoke, Rain on the glass. It applies to every shot.
- Opening frame — what the first frame shows: framing, who is in the picture and how they stand or sit, where they are, what is in their hands. It is written from the end of the previous segment so the join is continuous; only people who are visible in that frame belong here. Why the next segment opens like the last one ended.
- Set dressing — the room in words: furniture that matters, the light source, the mess, parked cars outside. At the end, background people if any: Background people: diners at the other tables, quiet. Delete that clause and the room has no extras. Who is in frame.
- Props — the objects the segment uses, as prop chips. Name each again in the shot that handles it; one listed but never used is dropped when you save. Why props stay the same.

Keep the header consistent with the shots: a header that says night and a shot that describes sunlight fight each other. What you can change.

## Getting the newer layout for an older episode

Regenerate the episode’s storyboard: SceneMixer writes the new text in the newer layout. It rewrites every segment of the episode and discards their videos, so only do it when the breakdown is wrong throughout, not to get a different header. What you lose.

| Field | Newer layout | Older layout |
| Location | A scene chip | A scene chip |
| Time and Weather | One line, read from the script, with Change; one setting for the whole scene | Two lines of text you edit; apply to every shot of the segment |
| Opening frame | Not written | Editable text, from the end of the previous segment |
| Set dressing and background people | Not written | Editable text; Background people: at the end |
| Props | No line; named in the shots with @ | Prop chips, also named in the shots |

## FAQ

### Why does my segment header only have Location and a Time and Weather line?

It is a storyboard in the newer layout. SceneMixer keeps track of the opening frame, the room, the props and the crowd itself, so those lines are not written into the text.

### How do I change the time of day in a newer storyboard?

Press Change next to the Time and Weather line on SceneMixer, pick a time, and save. It applies to all segments of that scene and is used the next time you generate their videos.

### Does changing the time cost credits?

No. Only generating the video on SceneMixer does.

### Where did Opening frame, Set dressing and Props go?

In the newer layout they are not lines in the text. SceneMixer carries who is where from the previous segment, builds the room from the scene image, and puts props in a shot when the shot names them.

### How do I add a crowd in a newer storyboard?

Describe it in the shot where it matters. For a clip already made, use SceneMixer's Edit this clip, AI visuals, Background people.

### Can I get the older six-field header back?

Not in the newer layout. SceneMixer's ad and remake projects use it, and so do storyboards written before late October 2026.

### What does a dash in Time or Weather mean?

The script did not say. Use Change on SceneMixer to pick a value.

### A prop I listed vanished from the header.

In the older layout a prop listed but never used in a shot is dropped on save. Name it in a shot and it stays.

Updated November 3, 2026


---

# TikTok Video Length, Shorts and Reels: Limits Checked Nov 2026

URL: https://scenemixer.com/guides/tiktok-video-length/
How long a video can be on TikTok, YouTube Shorts and Reels: the official figures we could read on November 2, 2026, what we could not, and how to plan a length.

# TikTok video length (and Shorts and Reels)

Quick answer: Platforms change their length limits often, so this page quotes only what we read on the platforms' own help pages on November 2, 2026. YouTube Shorts: up to 3 minutes, with a square or vertical frame. Instagram Reels: the help page says you can record and edit up to 20 minutes, and that reels over 3 minutes will not be recommended to new audiences. TikTok: its help-centre pages would not display for us, so no TikTok figure is stated here; the upload screen in the app tells you your current limit before you post. The one number that does not move is the shape: vertical, 9:16.

Plan from the platform, not the other way round: open the upload screen of the app you will post to, note the longest video it accepts today, then work back to segments: a segment is 15 or 30 seconds, so three minutes is 12 segments of 15 seconds or 6 of 30.

## What the official pages say (read November 2, 2026)

| Platform | What the page says | Official page |

| YouTube Shorts | Shorts can be up to three minutes long. Videos uploaded on or after October 15, 2024 with a square or vertical aspect ratio up to three minutes are categorised as Shorts. The upload page lists the same two conditions: up to 3 minutes, square or vertical. | Understand three-minute YouTube ShortsUpload YouTube Shorts |

| Instagram Reels | You can record and edit videos up to 20 minutes, as one clip or several clips that add up to that. Reels over 3 minutes won't be recommended to new audiences. The reel size page gives the shape: between 1.91:1 and 9:16. | Record a reel on InstagramReel size and aspect ratios |
| TikTok | Not stated here. We could not read the help-centre pages on November 2, 2026, and we do not repeat figures second-hand. Check the upload screen in the app, or the help centre while signed in. | TikTok Help Center |
One small thing worth knowing about quoting limits: on the day we checked, the search results page of Instagram's help centre still showed a one-line summary of the reels article saying 3 minutes, while the article itself said 20. Read the article, not the blurb, and read it again before a launch.

## The part that does not change: the shape

Whatever the length, short-video feeds are vertical. YouTube's Shorts rule is square or vertical; Instagram's range runs from 1.91:1 down to 9:16; TikTok's feed is a phone screen held upright. A 9:16 video is inside all three. In SceneMixer the shape is chosen once, in Step 1 (9:16 Portrait or 16:9 Landscape), and cannot be changed after you continue, so choose vertical before the script is read. Vertical or landscape · Posting to TikTok, Shorts and Reels.

## What lengths SceneMixer makes

Lengths are built from segments, and a segment is one generation of up to 15 seconds, or up to 30 seconds on the tiers that take long takes:

| What you make | Length | How it is built |
| An ad | 15, 30 or 60 seconds, plus an end card of a few seconds | One, two or four 15-second segments, joined with the end card |
| A Short Drama episode | Up to 240 seconds | Up to 60 segments of 15 or 30 seconds |
| A TV Series episode | Up to 300 seconds | The same, with a slower open allowed |
| A Film | No set limit | Sequences, each made and paid for on its own |
Two consequences. A Short Drama episode can run to 240 seconds, which is longer than the three minutes YouTube's page allows for a Short, so for Shorts keep an episode to 180 seconds or less, or cut it in two. And an ad is a few seconds longer than its label because the end card is added after the segments. Choosing script length and cost · Ad length, format and quality.

## Counting back from a limit

| Target length | Segments of 15 s | Segments of 30 s | A fit for |
| 15 seconds | 1 | — | One beat: a hook, an ad, a teaser |
| 30 seconds | 2 | 1 | A small scene or a one-product ad |
| 60 seconds | 4 | 2 | A whole mini story or a four-part ad |
| 90 seconds | 6 | 3 | A short episode with a turn and a hook |
| 3 minutes | 12 | 6 | The longest Short on the YouTube page above |
Video on SceneMixer is priced by the second, so cost follows length: 10 credits a second on the default tier plus 1 credit a second to join. Fifteen- or thirty-second segments cost the same per second; a 30-second segment is just a bigger retake if it misses. 15 or 30 seconds · AI video cost calculator.

## Before you post

- Check the limit on the day. Open the upload screen of the app. It is the only source that is current.
- Post the downloaded original, not a share link; the download is the full-quality file. Original or play version.
- Turn on the platform's AI label. Free accounts also carry a small SceneMixer label in the corner of the video. The AI label switch.
- Open on the hook. Length matters less than the first seconds. Hooks and cliffhangers.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Neon Nights, Campus Youth; Costume Style Streetwear, School Uniform; Visual Style Digital HD, Soft Glow; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the two leads and the one or two places; a 60-second episode is four segments. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Platform | Length (official page, November 2, 2026) | Shape |
| YouTube Shorts | Up to 3 minutes | Square or vertical |
| Instagram Reels | Up to 20 minutes recorded; over 3 minutes not recommended to new audiences | 1.91:1 to 9:16 |
| TikTok | Not quoted: check the upload screen | Vertical, 9:16 |
| SceneMixer | Segments of 15 or 30 seconds; Short Drama episodes up to 240 seconds | 9:16 or 16:9 |

## FAQ

### How long can a TikTok video be?

TikTok publishes its limit on its own help pages and in the upload screen, and changes it often. We could not read the help centre on November 2, 2026, so this page does not quote a number; check the app before you post.

### How long can a YouTube Short be?

Up to three minutes, with a square or vertical aspect ratio, according to YouTube's help pages read on November 2, 2026.

### How long can an Instagram Reel be?

Instagram's help page says you can record and edit up to 20 minutes, and that reels over 3 minutes will not be recommended to new audiences.

### What length should my AI video be?

As long as the story needs and no longer than the platform accepts: 15 seconds for one beat, 30 to 60 for a small story, up to 3 minutes for a full short.

### Can I make a video longer than 60 seconds?

Yes. A Short Drama episode on SceneMixer runs up to 240 seconds, and a series continues across episodes.

### How many segments make a 3-minute video?

Twelve segments of 15 seconds or six of 30 seconds.

### Does a longer video cost more?

Yes, by the second on SceneMixer: 10 credits a second on the default tier plus 1 credit a second to join.

### Which aspect ratio should I choose?

9:16 Portrait for short-video feeds. It is chosen once in SceneMixer's Step 1 and cannot be changed afterwards.

### Why does this page not give a TikTok number?

Because we could not verify one on the platform's own page, and we do not repeat figures from third-party lists. A wrong limit is worse than none.

Updated November 3, 2026


---

# Timeline Cards: Every Icon, Price Tag and State | SceneMixer

URL: https://scenemixer.com/guides/timeline-cards-every-icon/
The row of cards along the bottom of the editor is the episode. What the thumbnail, seconds, status word, price tag, Free, over cap, version count and card menus mean.

# Timeline Cards: Every Icon and What It Means

Quick answer: Every segment of the episode is a card on the bottom timeline, in order. Each card shows a thumbnail once there is video, the segment’s length in seconds, a status (Pending, Generating, Done, Failed), and the price the generate button would charge — or Free. A card that says over cap is longer than the selected tier allows. Click a card to select the segment; its menu holds Insert segment after and Delete segment. At the end of the row: Composite Episode.

Open a finished episode and hover the cards on its timeline.

## Thumbnail and number

The card’s number is the segment’s place in the episode; it shifts when you insert or delete. The thumbnail is a frame from the active version of the segment’s video; a card without video shows no thumbnail.

## Seconds

The segment’s length: the sum of its shots. Change the shots and the number changes. Shot seconds.

## Status

Pending: no video yet. Waiting: behind your other jobs. Generating: running, with the usual time in the preview pane. Done: video delivered. Failed: read the message in the preview pane. Every status.

## The price tag, and Free

SceneMixer's price tag shows what generating this segment would cost right now: its seconds times the selected tier’s per-second rate. Change tier and every tag changes. Free means this segment is covered by a free allowance. Tiers.

## over cap

The segment is longer than the selected tier allows; generating is refused until you shorten it, split it, or pick a tier with a higher cap. Segment length.

## Versions

A segment that has been generated more than once shows how many versions it has. Open Version History in the preview pane to switch; the thumbnail follows the active one. Versions.

## The card menu

Insert segment after adds a blank segment behind this one. On SceneMixer, Delete segment removes it with its video and versions, after a confirmation; it is refused while the video generates and on the last segment. Inserting and deleting.

## No Stop on a clip

A segment that is generating has no Stop: the clip runs to the end and lands as a new version, which you can keep or ignore. What can be stopped.

## The end of the row

Composite Episode, available when every card is Done; until then its tooltip counts the remaining segments. Composite History lists the files you have made. Compositing.

| On the card | Means |
| Number | Position in the episode |
| Thumbnail | A frame of the active version |
| Seconds | Sum of the segment’s shots |
| Pending / Waiting / Generating / Done / Failed | Where the video stands |
| Price tag | Cost to generate now, at the selected tier |
| Free | Covered by a free allowance |
| over cap | Longer than the tier allows |
| Version count | How many takes exist |
| Insert segment after / Delete segment | The card menu |

## FAQ

### Why did the price on every card change?

You changed the tier in the model picker. Tags on SceneMixer follow the selected tier's per-second rate.

### What does over cap mean?

The segment is longer than the selected tier allows on SceneMixer. Shorten, split, or change tier.

### Why does a card say Free?

SceneMixer's free allowance covers that segment.

### Why is there no thumbnail?

No video has been generated for that segment yet.

### Where is the delete button?

In the card's menu on SceneMixer: Delete segment. It is refused while the video generates and on the last segment.

Updated November 3, 2026


---

# "Too Many Requests", "Too Soon": Limits That Aren't About Credits

URL: https://scenemixer.com/guides/too-many-requests-too-soon-rate-limits-that-are-not-about-credits/
A few actions are limited by how often you can do them, not by your balance: sign-in codes, image recognition, voice matching, ad and remake starts.

# ‘Too Soon’, ‘Too Many Requests’, ‘Busy Right Now’: The Limits That Count Clicks, Not Credits

Quick answer: Some actions are free or nearly free for you but cost the site something each time — sending a code, recognising an image, matching a voice, translating an edit, starting an ad read or a remake read. Those are limited by frequency: so many per minute, per hour, or one at a time per account. The message names the action; the fix is always to wait, from a few seconds (codes) to the top of the hour (recognition, matching). Nothing is charged and nothing is queued: the click is simply refused. This is different from the concurrency limit on generations, which queues your jobs rather than refusing them.

Request a sign-in code twice in a row: the second is refused for a few seconds; the first is still valid.

## Which actions have a rate limit

Sign-in codes: one every few seconds, a handful per hour per phone or address. Image recognition (ad photos, descriptions from images): per person per hour, plus a site-wide budget. Voice matching: per person per hour. Translated edits: many edits to translated storyboards in a short time. Ad and remake starts: per hour, and one read in progress at a time per account. Site-wide daily limit: a shared budget that, when reached, refuses until the next day; rare.

## What the messages look like

‘Too soon — wait a few seconds’, ‘Too many codes’, ‘Image recognition is busy’, ‘Too many image recognition requests’, ‘Too many voice matches’, ‘Too many edits in a short time’, ‘This ad is busy right now’, a remake that is already reading. Each one is SceneMixer's refusal of that click; the page state is unchanged. Every Error Message.

## How long to wait

Codes: seconds; the hourly cap resets on the hour. Recognition and matching: the hour. Translated edits: minutes. Ad and remake: the read in progress has to finish, or the hour has to turn. The message does not show a timer; a retry after the stated interval works.

## Not the same as the concurrency limit

Generations (sheets, storyboards, clips) are never rate-refused; beyond your concurrent-task limit they queue and start as slots free up. Rate limits are for the auxiliary actions above. Generating Many Segments at Once.

## Why these limits exist

Each of these actions calls a model or sends a message at SceneMixer’s cost, with nothing charged to you. A per-person cap keeps a runaway script or a stuck button from spending that budget; the numbers are set so that normal use never touches them.

| Action | Limit | Wait |
| Sign-in code | Per few seconds; per hour | Seconds; the hour |
| Image recognition | Per hour per person; site budget | The hour |
| Voice matching | Per hour per person | The hour |
| Translated edits | Per short window | Minutes |
| Ad or remake start | Per hour; one read at a time | The read, or the hour |
| Generations | Not rate-limited; queued | — |

## FAQ

### Did the refused click cost credits?

No. SceneMixer's rate refusals never charge or queue anything.

### I keep getting 'too many codes'.

Use the code you already received; it is valid for its lifetime. Then wait for the hour.

### Is my account flagged?

No. SceneMixer's limits are ordinary counters that reset by themselves.

### Can a member get higher rate limits?

The concurrency limit on SceneMixer rises with membership; the auxiliary rate limits do not.

### The remake says busy but I see no read running.

A read in progress on another device or tab; wait for it, or refresh the projects list.

Updated November 3, 2026


---

# "Translating n/m": The Overlay After a Storyboard Lands

URL: https://scenemixer.com/guides/translating-n-of-m-the-overlay-after-a-storyboard-lands/
When you read a project in another language than its working text, a fresh storyboard is translated segment by segment first. What the counter means and how to skip it.

# ‘Translating n / m’: Why the Editor Waits After a Storyboard Lands, and the Three Ways Through

Quick answer: If your interface language differs from the project’s working text, every new storyboard is translated for you segment by segment. The editor shows a counter — segments ready out of segments total — and opens by itself when they are all ready, usually within a minute. It is a soft wait, never a lock: the button under the counter opens the original text at once, and the wait releases itself after three minutes whatever the state, so a slow translation can never keep you out. A segment that fails to translate shows a status line above it with a retry; the original is shown there meanwhile. On the Chinese site, and with an English interface on an English project, there is nothing to translate and no overlay.

Generate a storyboard with the interface set to a third language: Step 3 shows “translating n/m” on the episode card, and the editor opens when it reaches m/m or when you choose to read the original.

## What the overlay is

The storyboard is written in the project’s working text (Chinese or English). To let you read and edit it in your language, each segment is translated after it lands; segments are done a couple at a time, in parallel, and the counter rises as they finish. Which languages are which: The languages on this site.

## The three ways through

Wait: the counter reaches the total and the editor opens; typically under a minute for an episode. Read the original: the button under the counter opens the editor with the working text; translated segments swap in as they finish. Time: three minutes after the storyboard landed the wait releases on its own, server side and in the page, even if some segments are not ready. SceneMixer's overlay exists to spare you a half-translated page, not to hold you.

## Where else it shows

Step 3 episode cards carry the same counter while an episode is being translated. Editing an already open storyboard is not gated; only a freshly generated one. Remakes behave the same when their working text differs from your language; ads are shown in their working language and have no translation step.

## A segment that never turns

If translation of a segment fails after its retries, a status line above the segment says so and offers a retry; the segment shows the original text until then, and you can generate its clip from that. Editing a segment while it is untranslated edits the working text directly. The five wait messages.

## Cost and speed

Translation on SceneMixer costs you nothing and does not use your concurrent-task slots. Speed depends on segment length and the translation service’s load; a very long episode takes longer, which is why the three-minute release exists.

| You see | Means | Do |
| Translating 3/8 | Three segments readable in your language | Wait, or read the original |
| Counter stuck | A segment is retrying | Read the original; it swaps in when done |
| Overlay gone at 6/8 | Three minutes passed | Remaining segments show original until ready |
| Status line: not translated, retry | Translation of that segment failed | Retry, or work from the original |

## FAQ

### Can the overlay lock me out?

No. The button on SceneMixer's overlay opens the original at once, and it releases by itself after three minutes.

### Is translation charged?

No.

### Why is there no overlay on the Chinese site?

Working text and interface are both Chinese; nothing to translate.

### Can I generate a clip from an untranslated segment?

Yes. SceneMixer generates clips from the working text either way.

### Why did my edit take a moment to save?

SceneMixer writes edits made in a translated view back to the working text; that is the 'syncing' wait.

Updated November 3, 2026


---

# Trim an AI Video Clip: Cut the Start or the End | SceneMixer

URL: https://scenemixer.com/guides/trim-a-clip-cut-the-start-or-end/
Trim a finished segment in the editor: drag two handles, preview what is kept, save it as a new version. Frame-accurate, picture and sound together, original kept.

# Trim a Clip: Cut the Start or the End Without Regenerating

Quick answer: Select the segment, and under the player open Edit this clip → Trim (it opens there by default). Leave the mode on Keep, drag the two handles to where the clip should start and end, press Preview to watch only the kept part, then Save. The trim is cut at the exact frame, picture and sound together, and saved as a new version of the segment. It costs 5 credits and the clip you started from is kept.

Good first use: a clip whose last line ends a second before the clip does. Pull the end handle back to just after the line and save.

## Step by step on a computer

- Open the episode and click the segment on the timeline. Its clip plays in the right-hand pane.
- Under the player, Edit this clip is open on Trim. The two words at the top, Keep and Remove, set what the handles mean; leave it on Keep.
- Drag the left handle to where the clip should start and the right handle to where it should end. The player jumps to that frame while you drag, so you can see exactly where you are.
- For fine moves use − and + beside Start and End: each press moves that handle by half a second.
- Press Preview. Only the part between the handles plays, then it stops. Press Stop to end early.
- Press Save. The trim button in SceneMixer shows the price (5 credits). While it works the panel says Trimming…; when it is done you see Trimmed and saved as a new version.

## On a phone

Open the Preview tab, pick the segment, and tap Trim under the player. The strip of segments below the player turns into a trimming track, while the player stays in view and follows the handles as you drag. The same controls are there: Keep or Remove, the track, − and + for each end, Preview and Save. Close the trimming mode to get the segment strip back.

## What changes and what does not

The new version contains exactly the frames between the handles, with the sound cut at the same points — SceneMixer generates nothing again, so the picture and voices are the ones you already had. It appears in Version History labelled Trimmed. If the clip you trimmed is still the one showing when the trim finishes, the trimmed version becomes active straight away; otherwise it waits in the history.

Your storyboard does not change, and neither does the segment’s planned length. If you regenerate the segment later, the new clip comes out at the full planned length again and is priced by it. How the versions fit together.

## Trimming again

A segment keeps one trimmed version. If you trim again, the new trim is cut from whatever version is on screen — which may be the trimmed one — and replaces the previous trimmed version. The generated clip is never touched, so the way back is always there: switch to it in Version History.

## How a trim affects the episode

Composite on SceneMixer joins each segment’s active version and is priced by the total length, 1 credit per second. A trimmed segment counts at its trimmed length, so trimming dead air also makes the episode a little cheaper to composite. Download and share also use the active version. Which segments are joined.

## Limits

At least one second must remain; when a handle reaches that limit the panel says Keep at least 1s. A clip that is one second or shorter cannot be trimmed. You cannot trim while the segment is generating or while the episode’s storyboard is being generated, and one trim per segment runs at a time. There is also an hourly limit on trims per account. Every message and what to do.

| Control | What it does |
| Keep / Remove | Keep = the handles mark what stays. Remove = they mark what is cut out |
| Handles on the track | Drag to set the start and the end; the player follows |
| − / + next to Start and End | Move that handle half a second back or forward |
| Preview | Plays only what will be kept, then stops |
| Save · price | Makes the trimmed version; 5 credits, refunded if it fails |

## FAQ

### Does trimming lower the quality?

Not visibly. SceneMixer saves the kept frames again at high quality; nothing is generated again.

### Can I trim the same clip twice?

Yes. The second trim on SceneMixer replaces the first trimmed version; the generated clip stays in Version History.

### Why did my trimmed version not become active?

The active version changed while the trim was running, for example because you switched versions. The trimmed version is in SceneMixer's Version History; switch to it.

### Can I add a fade or a transition?

No. Trim only cuts. Composite on SceneMixer joins segments with straight cuts.

### Can I trim a clip in a sample project?

No. SceneMixer's sample projects are read-only: the panel is shown but greyed out. Trim your own copy of a project instead.

Updated November 3, 2026


---

# Trimmed, AI-Edited and Redubbed Versions of a Segment, Explained

URL: https://scenemixer.com/guides/trimmed-ai-edited-and-redubbed-versions/
Every fix to a finished clip is saved as a version. How Generated, Trimmed, AI edit and Redubbed versions relate, which one plays, what replaces what, and how to go back.

# Trimmed, AI-Edited and Redubbed Versions: How They Fit Together

Quick answer: A segment’s Version History can hold six kinds of version: Generated (each time you generate), Reused (a callback clip), and the four fixes — Trimmed, AI edit, Redubbed and Enhanced. Every generation is kept (inactive ones for 90 days). Of each fix, a segment keeps one: a new trim replaces the old trimmed version, and so on. The version marked active is what plays, downloads, shares and goes into the composite; switching between versions is free.

Open Version History on a segment you have trimmed: the trimmed version shows its real length in seconds, next to the generated clip it was cut from.

## The six labels

- Generated — made by the generate or regenerate button, with the tier it was made on.
- Reused — a callback segment that reuses an earlier segment’s clip. Callback clips.
- Trimmed — made by Trim; it shows its actual length after the cut.
- AI edit — made by AI visuals, with a note of what changed, such as “Mara · clothes → red”.
- Redubbed — made by Redub, with Lip sync or Re-perform.
- Enhanced — made by Enhance: the same take one resolution step up. Enhance a clip.

## One of each fix

A segment on SceneMixer keeps at most one Trimmed, one AI edit, one Redubbed and one Enhanced version. Making a new one of the same kind removes the previous one of that kind — that is what keeps the history readable after a dozen tries. Generated versions are never replaced this way; every generation stays until it is cleaned up for age (below).

A fix is always made from the version on screen when you press the button, so fixes can build on each other: AI-edit a clip, then trim the edited version, then redub the trimmed one. Each step is its own version of its own kind, and the history still leads back to the generated clip underneath. One exception: AI visuals and Redub on an enhanced version work from the original clip, so their result is not enhanced; a trimmed enhanced version stays enhanced.

## Which version is active

When a fix finishes, SceneMixer makes it the active version if the clip you started from was still the active one. If you switched versions or regenerated the segment in the meantime, the fix is saved but not switched on; open Version History and click it. Clicking any version makes it active, costs nothing and generates nothing.

The active version is what the player plays, what Download saves, what a share link shows and what the composite joins. Download: the original vs the play version.

## Going back

Switch to the generated clip in Version History. Because fixes point back to the clip they came from, that clip is kept as long as a fix made from it exists — the way back does not expire while you still have the fix. SceneMixer cleans up other versions that are not active after 90 days, as before.

## Length and the composite

A trimmed version is shorter, and an episode composite is priced by the actual length of each active version, 1 credit per second. AI edit, Redubbed and Enhanced versions keep the length of the clip they came from. Regenerating the segment ignores all of this: a new generated clip comes out at the segment’s planned length. Which segments are joined.

## What removes versions

Deleting the segment removes all its versions. Regenerating the whole episode’s storyboard rebuilds the segments, so the old versions are no longer attached; composite or download what you want to keep first. Re-parse an episode: what you lose.

| Kind | Made by | How many kept | Length |
| Generated | Generate / Regenerate | All (inactive ones cleaned after 90 days) | The planned length |
| Reused | A callback segment | — | The earlier segment's clip |
| Trimmed | Trim | One | What you kept |
| AI edit | AI visuals | One | Same as its source |
| Redubbed | Redub | One | Same as its source |
| Enhanced | Enhance | One | Same as its source |

## FAQ

### I trimmed twice. Where is the first trim?

It was replaced by the second; SceneMixer keeps one trimmed version per segment. The generated clip is still there.

### Can I keep two different AI edits?

SceneMixer keeps only one AI edit per segment. Download the first before running the second if you want to keep both files.

### Does switching versions cost credits?

No. Only making a new version on SceneMixer costs.

### Which version is in my share link?

SceneMixer shows the segment's active version when the page is opened.

Updated November 3, 2026


---

# Manga or Webtoon to Anime with AI: Turn Your Comic into a Series

URL: https://scenemixer.com/guides/turn-a-manga-or-webtoon-into-an-ai-anime-series/
Turn a comic you created into an animated series: write the chapters as a script, keep your character designs, pick the comic's look, and let the panels become shots.

# How to Turn Your Own Manga or Webtoon into an AI Anime Series

Quick answer: SceneMixer does not animate comic pages directly; it makes a series from a written script. That turns out to be the better route for a comic you own: you write each chapter as scenes with the dialogue from your speech bubbles, upload your own character designs as the reference sheets, and pick a visual style that matches the comic — Korean Webcomic, Anime, Painted Manhua or American Comic. Every episode is then storyboarded, rendered shot by shot against your designs, and joined into one video with the characters speaking their lines. Only adapt comics you have the rights to: characters from known series are refused.

Brief for AI Create, if you start from an idea instead of pages: A webtoon-style action romance: a delivery rider in a neon city discovers she can see 10 seconds into the future; 8 episodes, Korean Webcomic look, each episode ends on a vision.

## What goes in, what comes out

| From your comic | Goes into SceneMixer as |
| A chapter | An episode of the script |
| Panels | Scene descriptions: where, who, what happens |
| Speech bubbles | Dialogue lines with the speaker’s name |
| Narration boxes | Voice-over lines from an off-screen narrator |
| Character designs | Uploaded as each character’s reference image |
| The art style | The visual style chosen in step 1 |
Comic pages themselves are not an input: paste the text or upload .txt, .docx or .md. What you can feed in · Which script files work.

## Write the chapter as a script

Go panel by panel and write what the reader sees and hears. One page of a webtoon is usually two to four shots. Keep the dialogue exactly as in the bubbles; turn sound effects into the ambient line. SceneMixer reads screenplay format well, but plain prose works too. Screenplay headings and cues · Write dialogue so it is picked up.

| Panel | As a script line |
| Wide panel of a rainy rooftop, the hero alone | Night. A rooftop in the rain. MIRA stands alone at the edge, her delivery jacket soaked. |
| Close-up, speech bubble | MIRA: Ten seconds. That’s all I ever get. |
| Sound effect “KRAK”, door bursting open | The rooftop door bursts open behind her. |
| Narration box | NARRATOR (V.O.): She had already seen this part. |

## Keep your character designs

After the script is read, open each character in SceneMixer's step 2 and press Upload & Replace. Choose Optimize into a reference sheet to have your design redrawn as a clean full-length sheet in the project’s style, or Use this image as is if you drew a proper character sheet already. From then on every shot of that character looks at your design. Uploading is a membership feature on the current plan settings. Use your own image · What changes when you upload.

Keep the card’s written description in line with your design — it still steers how every shot is written. Costume changes from later arcs are states of the same character.

Character sheets from the sample Rust City Knockout: one reference image per character, used in every shot they appear in. A design you upload takes the same place as one of these generated sheets.

## Match the comic’s look

| Your comic looks like | Visual Style |
| A full-colour vertical webtoon | Korean Webcomic |
| Japanese manga, anime adaptation look | Anime, Painterly Anime, Watercolor Anime |
| Chinese manhua, painted | Painted Manhua, Chinese Painted Style |
| Western comic book | American Comic |
| 3D-style manhua | Cel-shaded 3D |
For a vertical webtoon, also choose 9:16 Portrait in SceneMixer's step 1: the comic’s reading direction becomes the phone screen. A consistent anime look · Stylised looks across tiers.

Visual Style: Anime

Visual Style: Painterly Anime

Visual Style: Watercolor Anime

Visual Style: Painted Manhua

Visual Style: Chinese Painted Style

Visual Style: American Comic

Visual Style: Cel-shaded 3D
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Rights and the IP check

Adapt only work you created or have the rights to adapt. An IP check refuses characters, names and iconic outfits from known films, games and comics, in the script and in uploaded images; a refused attempt is refunded. Fan versions of famous series will not get through. IP check: celebrities, brands, franchises.

## Make it in SceneMixer, start to finish

- Script: write chapter 1 (and as many as you like) as scenes and dialogue; paste it on Upload Script.
- Step 1: Type TV Series for a long-running comic or Short Drama for a short arc; the visual style that matches the comic; 9:16 Portrait for a webtoon; the dialogue language. Read the script.
- Step 2: upload your character designs with Upload & Replace, or generate sheets (the first 15 on an account are free).
- Storyboard: episode 1 first (free on your first project); check every shot names who is in frame.
- Video: generate on the default tier at 10 credits a second; composite; post the episode with your chapter.
For a painterly anime sample, watch Rust City Knockout. From a web novel · Making a multi-episode series.

## FAQ

### Can AI convert my manga into anime?

In SceneMixer, yes, from the story rather than the pages: write each chapter as a script with the bubbles' dialogue, upload your character designs, pick an anime style, and each chapter becomes an animated episode.

### Can I upload my comic pages?

No. SceneMixer's drama projects take a written script. Your character designs can be uploaded as reference images in step 2.

### Will my characters look like my designs?

On SceneMixer, uploaded designs become each character's reference, and every shot is matched to it. Optimize redraws the design as a clean sheet in the project's style; Use as is keeps it untouched.

### Which style for a webtoon?

Korean Webcomic, with 9:16 Portrait for phone viewing.

### Can the characters speak?

Yes. SceneMixer performs dialogue lines on screen in the language you pick in step 1, one of 15; narration boxes become voice-over.

### Can I adapt a famous manga as a fan project?

No. SceneMixer's IP check refuses characters from known series. Adapt your own work or work you have the rights to.

### How much of a chapter fits in one episode?

Roughly one webtoon chapter per one-to-three-minute episode; long chapters can be split across two.

### Do I need to know how to animate?

No. You write the scenes and the dialogue; SceneMixer generates the storyboard and the video, and you can edit any shot's text and regenerate it.

Updated November 3, 2026


---

# Turn a Podcast Episode Into an AI Drama

URL: https://scenemixer.com/guides/turn-a-podcast-episode-into-an-ai-drama/
Turn a podcast story into an AI drama: transcript to script, the host as narrator, the story's people acted with their own voices, and a vertical cut for clips.

# How to turn a podcast episode into an AI drama

Quick answer: A story told on a podcast can be acted on screen. Take the transcript of your episode, rewrite the story parts as scenes with dialogue, keep the host's best lines as a narrator's voice-over, and upload it as a script. SceneMixer casts the people in the story, draws the places once, and renders the episode, with a vertical cut for clips. Use your own podcast, or one you have permission to adapt.

Brief for AI Create: From our podcast: the night a small-town radio host received a call from a listener claiming to be calling from 1987; the host narrates, the call is acted; four minutes.

## From transcript to script

| Transcript part | Becomes |
| The host introducing the story | Narrator voice-over over an establishing shot |
| The story itself | Scenes with dialogue for the people in it |
| Guest commentary | Cut, or a short voice-over between scenes |
| Ads and housekeeping | Cut |
Upload as TXT or DOCX. A raw transcript works, but a rewrite into scenes gives a much better read. Which files work.

## Voices

The host as narrator can use a preset voice, or, for members, an uploaded clip of the host's own voice (with their consent). The people in the story get their own voices from SceneMixer's library. Pick character voices · Voice-over lines.

## A clip for every episode

Podcasts grow through short clips. Make the full episode in 16:9 for YouTube, then a second project in 9:16 with the most dramatic minute, reusing the same cast from your library. Reuse characters · Posting short clips.

| Beat | Shot to adapt |
| The host | [Shot 1] 4s: Wide shot, a small radio studio at night. @Host (voice-over): This call came in at 2:14 a.m. |
| The call | [Shot 2] 3s: Close-up on @Dale, the night DJ, lifting the receiver. |
| The voice | [Shot 3] 4s: @Caller (on the phone, crackling): Is the bridge still closed? The one on Route 9? |
| The reaction | [Shot 4] 3s: @Dale looks at a framed photo of the bridge on the wall: a 1987 newspaper. |

## Whose story

True stories about real, identifiable people need care: invented faces, changed names, nothing defamatory. For listener stories, have the listener's permission. Content rules.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style American Small Town; Costume Style Soft Everyday; Visual Style Shot on Film, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the host as a narrator character; the people in the story with specific faces; the studio and the story's places as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Podcast | Drama format |
| True story or mystery | Narrated drama, 3–8 minutes |
| Fiction podcast | Full drama series, one episode per episode |
| Interview show | One acted anecdote per episode |
| Comedy podcast | A sketch built from the best bit |

## FAQ

### Can I turn a podcast into a video drama with AI?

Yes. Rewrite the story parts of the transcript as scenes, keep the host as narrator, and upload it to SceneMixer as a script.

### Can I use the raw transcript?

SceneMixer accepts it, but rewriting into scenes with dialogue gives better results.

### Can the host's own voice narrate?

On SceneMixer, members can upload a clip of the host's voice with their consent; otherwise pick a preset.

### Can I make vertical clips?

Yes. Make a 9:16 project of the best minute with the same cast.

### Can I adapt someone else's podcast?

Only with permission.

### Do I need to change real names?

For real, identifiable people, yes: invented faces and changed names.

### How long should it be?

Three to eight minutes for one story.

### Is there music?

SceneMixer adds a score under the voices, and you can switch it off.

Updated November 3, 2026


---

# Turn a Poem Into an AI Short Film: Stanzas to Shots

URL: https://scenemixer.com/guides/turn-a-poem-into-an-ai-short-film/
Turn a poem into an AI short film: why a short poem needs a paragraph around it, how stanzas become segments, voice-over or silence, with an original poem worked through.

# Turn a poem into an AI short film

Quick answer: A poem becomes a short film in three moves: add a short paragraph that says who, where and when (the site needs about a paragraph to read, and a poem rarely says), give each stanza a segment of its own, and let a narrator read the lines as voice-over while the pictures show what the words don't say. A 12-line poem is about three 15-second segments, 45 seconds of film. Use your own poem or one you have the right to use.

A poem to try (an original, written for this page) is below, with its three stanzas mapped to three segments. Paste the poem with the setting paragraph above it and read the storyboard.

## The first problem: a poem is shorter than the floor

The site needs about a paragraph to break into scenes: roughly 330 characters of English. A haiku, or a short lyric, is under that, and a poem rarely says who is speaking, where they are or when. So put a short setting paragraph above the poem. What you can feed in · Upload a script and read it.

Setting. A woman in her sixties stands at the window of a rented room by the sea, folding a white sheet. She remembers her mother hanging washing in a yard when she was five. Morning wind. The poem is her thought, read aloud by a narrator.

## The poem

The Washing Line
Mother pegs the morning out,
one white sheet, then another;
wind unfolds them like a sail
and the yard becomes a harbour.

I am small enough to hide
in the cold, clean smell of linen,
hear the house go on without me,
count the clothes-pegs, count the sky.

Years on, I fold a stranger's sheet
in a rented room by the sea.
The wind comes in, and for a moment
I am small, and she is calling me.

## Stanzas to segments

| Stanza | Words | Speech at 3 words a second | Segment | The picture |
| 1 | 22 | about 7 s | 15 s | A yard at morning; sheets filling like sails on a line; a small child, seen from behind |
| 2 | 26 | about 9 s | 15 s | Inside the sheet: white light, a hand counting pegs, a house sound far off |
| 3 | 30 | about 10 s | 15 s | The rented room by the sea; the old woman folds a sheet; the wind lifts the curtain; she turns toward a voice |
Each stanza's speech takes 7 to 10 seconds, which fits a 15-second SceneMixer segment and leaves room to breathe. For a longer film, give each stanza two segments and let the second be silent: just the picture and the ambience. Shot seconds and timing.

## Voice-over, not dialogue

A narrator's lines are off-screen audio: heard, while the people in frame keep their mouths closed. Write each line as a shot's voice-over:

[Shot 1] 5s: Wide shot, a yard at morning, white sheets lifting on a line. @Narrator (voice-over, soft): Mother pegs the morning out, one white sheet, then another. (voice: Narrator)
[Shot 2] 5s: Medium shot from behind: a small child watches the sheets fill with wind. @Narrator (voice-over, soft): Wind unfolds them like a sail, and the yard becomes a harbour. (voice: Narrator)
[Shot 3] 5s: Wide shot, the line against the sky. Ambient: wind in cloth, a gate. No dialogue.

Lines, voice-over and captions · How voices work.

## Do not illustrate every line

If the picture says what the words say, the film is a caption. Show what the poem leaves out: the sheet as a sail is a metaphor, so put the sail in the picture only once, and let the rest be the actual sheet. Metaphors work well as props that stay the same object in every shot: the white sheet is the same sheet in the yard and in the rented room. Props and locations.

Visual Style: Painterly Anime

Visual Style: Classical Oil Painting
The cards as they appear in step 1, Set Up Your Project. Pick one in each row before the script is read; the visual style is fixed once you continue.

## Whose poem is it

Use a poem you wrote, or one you have the right to use. The Terms prohibit generating protected works without permission, and the confirmation window before an upload asks you to confirm that you have the right to use the material. Terms · What you can feed in.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Pastoral Healing, American Small Town; Costume Style Soft Everyday; Visual Style Watercolor Storybook, Soft Glow; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw the woman and the child she was, the yard and the rented room; the white sheet is the one prop that crosses both. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Poem length | Setting paragraph | Segments (15 s) | Film |
| A haiku (3 lines) | Needed: the poem alone is under the floor | 1 | 15 s |
| A short lyric (8 lines) | Needed | 2 | 30 s |
| A 12-line poem, 3 stanzas | Recommended | 3 (or 6 with silent pairs) | 45 s (or 90 s) |
| A longer poem | Recommended | One per stanza or two | As long as you choose |

## FAQ

### Can AI turn a poem into a video?

Yes. Add a short paragraph that says who, where and when, paste the poem, and SceneMixer turns the stanzas into segments read by a narrator over pictures.

### What if my poem is very short?

SceneMixer needs about a paragraph, roughly 330 characters of English. Write a setting paragraph above it.

### Who reads the poem?

A narrator, as voice-over; people in frame keep their mouths closed. A SceneMixer member can upload a voice recording they have the right to use.

### Should the pictures show every line?

No. Show what the words leave out, and use a prop for a metaphor.

### Can the poem appear as text on screen?

SceneMixer does not draw readable on-screen text reliably. Add it in your editor, or use the subtitles feature, which shows the lines your characters speak.

### Can I use a famous poem?

Use your own, or one you have the right to use; SceneMixer's Terms prohibit generating protected works without permission.

### Which look suits a poem?

Watercolor Storybook, Painterly Anime or Classical Oil Painting, depending on the mood.

### What does it cost?

A 45-second film on SceneMixer is 45 seconds of video at 10 credits a second on the default tier plus 1 credit a second to join, before retakes.

Updated November 3, 2026


---

# Turn a Reddit Story into an AI Drama: From Post to Episodes

URL: https://scenemixer.com/guides/turn-a-reddit-story-into-an-ai-drama/
Turn a Reddit-style story into an AI short drama: your own story or one you may adapt, rewritten as scenes with dialogue, real people kept out.

# How to turn a Reddit-style story into an AI drama

Quick answer: Reddit-style stories (am I wrong, revenge served cold, my sister stole my wedding) already have a hook, a villain and a twist, which is why they work as short dramas. In SceneMixer, rewrite the post as scenes with dialogue instead of narration, upload it as the script, and let the read split it into episodes. Use your own story or one you have permission to adapt, change names and details, and never depict the real people involved.

Brief for AI Create: A drama in the style of an online confession: a woman finds out her fiancé's family has been holding a second engagement party for him and her cousin; she shows up with the receipts; 6 short episodes, each ends with a reveal at a family event.

## Rights and real people

- Use your own story or one you have the author's permission to adapt; uploads ask you to confirm you have the rights. What you can feed in.
- Change names, places and details so nobody can be identified.
- Never depict real people, including the people in the original story; defaming or exposing real people is not allowed. Content rules.

## From post to script

- Turn narration into scenes. “My sister told everyone” becomes a scene at dinner where she tells everyone.
- Keep the narrator's voice as an inner voice or voice-over for the best lines. Voice-over.
- Put the hook in the first shot: the most shocking line of the post. Hooks.
- Upload the script to SceneMixer and let the read split it into episodes. Upload a script.

## The first shot, shot by shot

| Beat | Shot to adapt |
| The hook | [Shot 1] 4s: Close-up on @Jess holding her phone. @Jess (voice-over): Two weeks before my wedding, I found the second invitation. (voice: Jess) |
| The invitation | [Shot 2] 3s: Close-up on the phone screen: a party invitation with her fiancé's name and her cousin's. |
| The room | [Shot 3] 4s: Wide shot, a restaurant private room, balloons, family in the background. @Jess stands in the doorway. |
| The reaction | [Shot 4] 3s: Close-up on @Aunt Linda dropping her glass. |

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Short Drama; Scene Style Urban Opulence, Suburban Mansion; Costume Style Smart Casual; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: invented characters with one mark each; the phone and the invitation as props; the flat and the restaurant as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Episode | Where | The ending beat |
| 1 | Her flat. | The second invitation. |
| 2 | The restaurant. | She walks into the party. |
| 3 | The family group chat. | Everyone knew. |
| 4 | The wedding venue. | She changes the guest list. |
| 5 | The wedding day. | The cousin arrives in white. |

## FAQ

### How do I turn a Reddit story into an AI drama?

Use your own story or one you have permission for, rewrite it as scenes with dialogue, upload it to SceneMixer as the script and render episode by episode.

### Can I use any Reddit post?

Only posts you wrote or have the author's permission to adapt; uploads on SceneMixer ask you to confirm your rights.

### Can I show the real people from the story?

No. Change names and details and invent the faces.

### How do I keep the narrator's voice?

Use voice-over or inner voice for the narrator's best lines.

### Vertical or landscape?

9:16 Portrait; these stories live on short-video feeds.

### How long should each episode be?

One to two minutes, each ending on a reveal.

### Which visual style suits it?

Digital HD for realism; Korean Webcomic for a drawn look.

### Can AI write it from a short summary?

Yes. Put a one-line brief into SceneMixer's AI Create.

Updated November 3, 2026


---

# Turn a Stage Play Into an AI Short Film

URL: https://scenemixer.com/guides/turn-a-stage-play-into-an-ai-short-film/
Turn a stage play into an AI short film: open the single set into real locations, split long speeches across shots, keep the dialogue, and render it in Film mode.

# How to turn a stage play into an AI short film

Quick answer: A play is dialogue on one stage; a film can go anywhere. Upload the play (yours, or one in the public domain), choose Film, and SceneMixer reads the characters and scenes. Then open it out: give each scene a real location, split long speeches across several shots, and let a look replace a line where the camera can say it. The dialogue stays yours.

Paste one act of your play, choose Film and 16:9, and compare the shot list with your stage directions before rendering.

## Opening out the set

On stage, a living room may stand for a whole house; on screen, it can be the kitchen, the porch, the car outside. Add locations where the action moves, but keep the core room as the main set, drawn once. Your scene image becomes the set.

| Play | Scene Style | Costume Style |
| Family drama | American Small Town, Suburban Mansion | Soft Everyday |
| Victorian piece | English Manor, Gothic Dark | Victorian |
| Classical tragedy | Greco-Roman | Greco-Roman Attire |
| Absurdist | A bare location with one object | Simple, timeless clothes |

## Long speeches, many shots

A two-minute monologue in one take is not possible, and would not hold on screen anyway. Split it: the speaker, the listener's reaction, an object that matters, back to the speaker. SceneMixer keeps the words exactly as written. Write dialogue in the storyboard.

| Beat | Shot to adapt |
| The speech begins | [Shot 1] 5s: Medium shot. @Ada at the kitchen table. @Ada: You never asked why I stayed. |
| The listener | [Shot 2] 3s: Close-up on @Tom by the sink, not turning around. |
| The object | [Shot 3] 3s: Close-up. A suitcase by the door, packed. |
| The speech ends | [Shot 4] 5s: Close-up on @Ada. @Ada: I stayed because you would have let me go. |

## Stage directions

Directions such as She crosses to the window are read as action. Directions about the theatre itself (lights fade, curtain) can be dropped; the storyboard plans its own cuts. Character names in capitals and the name-colon-line layout both work. Scene headings and character cues.

## Rights and length

Adapt your own play or a public-domain text. A full-length play is long; adapt one act first and watch it. The Film type keeps the pace slow and the silences intact. Short film budgets · Series or film.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style American Small Town, English Manor; Costume Style Soft Everyday, Victorian; Visual Style Shot on Film, Black & White Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: the play's characters with specific faces; the main room plus two or three opened-out locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Stage element | On screen |
| One set | A main location plus a few others |
| Monologue | Several shots with reactions |
| Aside to the audience | A look to camera or voice-over |
| Blackout between scenes | A cut; time and weather set on the next segment |
| Offstage voice | A line marked (off-screen) |

## FAQ

### Can AI turn a play into a film?

Yes. Upload the play to SceneMixer, choose Film, open the set into locations and split long speeches across shots.

### Will my dialogue be kept?

Yes; SceneMixer keeps lines word for word when written with the speaker's name.

### Can I adapt a famous play?

Public-domain texts, yes; recent plays only with the rights.

### How do I handle monologues?

Split them across shots with reaction shots and objects in between.

### Can characters speak to the camera?

Yes, for asides; write the look to camera into the shot.

### Which look suits a stage adaptation?

Shot on Film or Black & White Film.

### How long can it be?

Start with one act; a full play is a long, costly film.

### Can the play stay on one set?

Yes; a single location drawn once keeps it theatrical.

Updated November 3, 2026


---

# Turn a Tabletop RPG Campaign Into an AI Series

URL: https://scenemixer.com/guides/turn-a-tabletop-rpg-campaign-into-an-ai-series/
Turn your tabletop campaign into an AI series: session notes rewritten as scenes, the party drawn once, fantasy presets for every realm, and one episode per session.

# How to turn a tabletop RPG campaign into an AI series

Quick answer: A campaign already has a cast, a world and a season's worth of episodes; it just lives in session notes. Rewrite each session as scenes with dialogue, upload them as one script, and SceneMixer draws the party, the villains and the places once with fantasy presets such as Fantasy Adventurer, Dwarven Forge and Elven Forest, then makes one episode per session. Use your own world and characters; published settings and their named characters are someone else's.

Brief for AI Create: A fantasy series from our campaign: a dwarf cleric, an elf thief and a human bard who cannot sing are hired to escort a crate that keeps knocking from the inside; eight episodes.

## From notes to a script

Session notes say what happened; a script says what is seen and heard. For each session, write three to six scenes: where, who, what they do, the best lines. Dice rolls become moments (the arrow misses by an inch). Keep one spelling per character name across sessions. Write dialogue so it is picked up.

## The party, drawn once

| Character type | Costume Style | Where they shine |
| Fighter, rogue, ranger | Fantasy Adventurer | Medieval Castle, Dungeon |
| Dwarf | Dwarven Forge | Mines and forges |
| Elf | Elven Attire | Elven Forest |
| Orc | Orc Tribal | Wasteland, Frontier Vastness |
| Wizard | Academy Robes | Magic Academy |
Level-ups, new armour and curses are states. Players can each have a voice matched from SceneMixer's library. States · Voices.

## One session, one episode

| Beat | Shot to adapt |
| The tavern | [Shot 1] 4s: Wide shot, a crowded tavern. @Brom, @Lyra and @Pip at a corner table around a wooden crate. |
| The knock | [Shot 2] 2s: Close-up. The crate lid jumps once. Ambient: a knock from inside. |
| The bard | [Shot 3] 3s: @Pip (whispering): I vote we do not open it. |
| The cleric | [Shot 4] 3s: @Brom picks up a crowbar. @Brom: Noted. |
A narrator in the style of a game master can open each episode with a voice-over recap. Voice-over lines.

## Your world, not a published one

Homebrew worlds and characters are yours. Named characters, monsters unique to a publisher and well-known settings belong to their owners and trigger the IP check; generic fantasy (dwarves, elves, dragons, taverns) is fine. IP check.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type TV Series; Scene Style Medieval Castle, Elven Forest, Dungeon; Costume Style Fantasy Adventurer, Dwarven Forge; Visual Style Painterly Anime, Digital HD; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: one sheet per player character and recurring villain; level-up states; the tavern, the road and the dungeon as locations. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Session | Episode | Ending beat |
| 1 | The job | The crate knocks back |
| 2 | The road | Bandits want the crate |
| 3 | The forest | The elf's old crew |
| 4 | The dungeon | The crate opens |
| 5 | The reveal | It is a child dragon |

## FAQ

### Can I turn a D&D-style campaign into an AI video?

Yes, with your own world and characters: rewrite the sessions as scenes, upload them to SceneMixer, and render one episode per session.

### Do the players' characters stay consistent?

Yes. SceneMixer draws each once and uses it in every episode; level-ups are states.

### Can I use official settings and monsters?

Generic fantasy is fine; SceneMixer refuses publishers' named characters, monsters and worlds.

### Can there be a game-master narrator?

Yes, SceneMixer runs it as a voice-over character.

### Which look suits a campaign?

Painterly Anime, Classical Oil Painting, Realistic 3D or Digital HD.

### Can I show combat?

Yes, one move per shot on SceneMixer, without wounds.

### How long should an episode be?

Three to five minutes per session works well.

### Can each player voice their character?

On SceneMixer, members can upload a clip they have the right to use; otherwise match a preset voice.

Updated November 3, 2026


---

# Turn a Web Novel Into an AI Drama Series: A Practical Guide

URL: https://scenemixer.com/guides/turn-a-web-novel-into-an-ai-drama-series/
Adapt a web novel into an AI short drama series: rights, which chapters to start with, how much to paste, how chapters become episodes, and what to set first.

# Turning a Web Novel Into an AI Drama Series

Quick answer: Paste the chapters you want to adapt into the script box, or upload them as a TXT, DOCX or Markdown file, and SceneMixer reads them into a cast, locations and a run of episodes. Start with the first arc rather than the whole book: one project reads one stretch of text, and a very long novel is adapted volume by volume, each a project that reuses the same cast. Choose the drama type, the target region and the dialogue language before the reading, because they shape how the chapters are cut and rewritten. Adapt only work you have the rights to.

Paste the first three chapters of your novel into a new project and read the summary in step 1 before generating anything else.

## Rights first

Adapting a novel you did not write needs the author’s or the publisher’s permission, even for a free web novel, and even if every frame is newly generated. Your own novel, a licensed one, or a public-domain work are safe starting points. What you get to keep · AI short drama compliance.

## How much to paste

A web novel is usually far longer than one project reads at once. If you paste more than the limit, SceneMixer says so in the box and asks you to split it into volumes. A good first project is the opening arc: the chapters that introduce the leads and end on the first big turn. Longer stretches are read in stages, and later episodes become ready a little after the first. Length caps and how characters count.

## How chapters become episodes

The reading does not keep your chapter breaks. It splits the text into episodes by story, giving each one a turn of its own, and lists them in step 3. Chapters full of inner monologue get condensed; scenes with conflict and dialogue carry over best. For lines you want spoken word for word, put them on their own line as Name: words before pasting. Write dialogue so it is picked up.

## Settings to choose before the reading

- Drama type. Short Drama for a fast vertical series; TV Series for a slower ensemble; Film for one long story. Short drama, series or film.
- Target region. Keep the novel’s setting, or move the story to another region with local names and places. Target region.
- Dialogue language. The language the characters speak in the videos. Choose a dialogue language.
- Visual style. Realistic, anime and other looks; SceneMixer locks it once you continue. Visual style presets.

## The next volume

When the first project is done, paste the next arc into a new project with the same settings, and bring the cast across with Pick existing in step 2 so faces and voices stay the same. Make the next episode.

## Cost, roughly

Reading is priced by the length of the text, each episode’s storyboard by its share of the text, and video by the second. Your first project includes some free items. Read one project’s prices on the pricing page before pasting a whole volume. What you get for free.

| Step | What to do with a novel |
| Pick the text | The opening arc, not the whole book |
| Check rights | Your own work, licensed, or public domain |
| Settings | Drama type, region, dialogue language, visual style before the reading |
| Read | Check the summary and the episode list in step 1 and step 3 |
| Next arc | A new project, cast reused with Pick existing |

## FAQ

### Can I paste the whole novel?

Only up to SceneMixer's reading limit for one project. Split a long novel into arcs and make each a project.

### Will my chapter titles become episode titles?

No. SceneMixer cuts episodes by story and titles them itself; you can read and rename them afterwards.

### Can I link to the novel's web page?

No. Paste the text itself; SceneMixer does not open links.

### Does the AI change the plot?

SceneMixer condenses and rewrites for the screen, keeping the events and the characters. Read the summary in step 1 before generating anything.

### Can I adapt a translated novel?

Yes, if you have the rights. In SceneMixer, pick the dialogue language you want the characters to speak.

Updated November 3, 2026


---

# Unusual Rows in Your Credit History, Explained

URL: https://scenemixer.com/guides/unusual-rows-in-your-credit-history-explained/
Most rows are freeze, consume and refund. The rarer ones: a charge that lands after the clip did, a release from review, a reclaim after a refund, a debt offset.

# The Unusual Rows in Your Credit History: Late Charges, Releases, Reclaims, Adjustments

Quick answer: Since October 2026 your credit history shows one line per task, tagged by where it stands: no tag once it is spent, In progress, Refunded, Partly refunded or Under review. The rarer events change that same line rather than adding new ones. A late charge: a clip was delivered but its charge could not be recorded at that moment; the line waits as Under review and turns into a normal spent line once the charge is recorded. A release: credits held for review were returned after checking; the line becomes Refunded. A reclaim: after a refund or chargeback, a Purchase reversed line takes back the credits that order had granted. A debt offset: what a membership reclaim could not take is settled from the next monthly grant, so that Membership credits line is smaller. Added by support and Adjusted by support: a manual change by support.

Click your balance, then View all: the full history on the billing page lists every line by day, with the tag that says where each one stands.

## Where the history is

Click your balance in the top bar: the latest entries are in the panel, and View all opens the full Credit History on the billing page, grouped by day, with Load more at the bottom. Each line shows the time, what it was, where it was, and the amount; there is no running balance, no type filter and no note on the line. Reading the everyday lines: Reading Your Credit History.

## A late charge

A generation delivered its clip but the charge failed to record at that second (a rare storage hiccup). The clip stays; the line waits as Under review; once the charge is recorded later, the tag goes and it reads as a normal spent line. The amount is the price SceneMixer shows on the card when you generated; nothing extra. The credits were already held when you generated, so the late charge only turns that hold into a charge; your available balance does not change and never goes below zero.

## Release from review

Some failures hold credits until a person checks whether the provider did the work; the task’s line says Under review meanwhile. When the check finds nothing was delivered, the hold is released and the same line turns into Refunded, the amount struck through. Credits Held for Review.

## Reclaim after a refund or chargeback

Refunding an order or disputing its charge takes back the credits that order granted, in proportion to the refund. If the balance cannot cover it, SceneMixer cancels queued generations first and uses their credits; any remaining shortfall on a membership becomes a debt. It shows as a Purchase reversed line marked After a payment refund or dispute. Refund Rules.

## Debt offset on a grant

A monthly membership grant first settles any debt from an earlier reclaim; the Membership credits line shows only what arrived after the offset, and when the debt swallows a whole grant there is no line that month. Once the debt is cleared, grants land in full again. Yearly Plan Credits Arrive Monthly.

## Adjustments, bonuses, referrals

Added by support and Adjusted by support: a manual credit or debit by support (a compensated failure, a corrected charge); SceneMixer does not print the reason on the line. First subscription bonus and Promotional bonus: first-subscription or offer extras. Membership credits: a plan’s monthly allowance. Referral reward: reward from the referral program, taken back in proportion as Referral reward reversed if the referred order is refunded.

| What happened | What you see |
| A job started | Its line, minus the price, In progress |
| The job delivered | The same line, no tag |
| The job failed without cost | The same line, amount struck through, Refunded |
| Charge recorded late | Under review, then a normal spent line |
| Hold released after review | Under review, then Refunded |
| Credits taken back after a refund or chargeback | A Purchase reversed line |
| A grant settled an earlier reclaim | A smaller Membership credits line |
| Manual change by support | Added by support / Adjusted by support |

## FAQ

### Why was I charged days after the clip?

The charge failed to record on delivery and was recorded later; until then the line said Under review. The amount is the card's price.

### Can a reclaim make my balance negative?

No. SceneMixer reclaims only what is available; the rest is recorded, and for a membership it is deducted from later membership grants.

### Who makes adjustments?

Support makes adjustments from SceneMixer's back office, with an audit record. The reason is not printed on the line; ask support if you need it.

### Can I dispute a line?

Write to support with the line's date, time and what it says.

### Do frozen amounts show in the balance?

SceneMixer's balance shows available credits; held amounts are shown separately as Frozen in the panel, and their lines say In progress or Under review.

Updated November 3, 2026


---

# Upload a Script and Parse It: From Paste to Cast List | SceneMixer

URL: https://scenemixer.com/guides/upload-a-script-and-parse-it/
The exact clicks from a script on your disk to a parsed project: paste or upload, the price on the button, the settings form, Let AI choose, Next, and the reading stages.

# Uploading a Script and Parsing It

Quick answer: On the home page, Upload Script tab: paste the text or Upload File (TXT, DOCX or Markdown). The button reads Get Started ✦ N, the read price for that length, 4 credits per 1,000 characters. Press it; the project opens at step 1, Set Up Your Project: name it, pick the three styles (or Let AI choose), the aspect ratio and, if you like, a target region and dialogue language, then Next. Reading the script runs for a few minutes through its stages; when Script Parsed appears, the summary, the episode list and the cast are on screen and Next takes you to step 2.

What you can feed in for formats and length; the script format for how to write one.

## Paste or upload

Paste directly into the box, or Upload File: TXT, DOCX and Markdown are read; the note under the box reminds you to hold the rights. A novel chapter, a screenplay, a treatment — any prose or script works; the read splits it into episodes itself. The character count and the price update as you type. A text longer than the limit is refused with a message before anything is charged.

## The price on the button

Get Started ✦ N. SceneMixer prices the read by length: 4 credits per 1,000 characters, where 3,333 characters of an alphabetic language count as 1,000. That is the whole first step; sheets and storyboards are priced later, and the first drama’s sheets and first storyboard are free. Where credits go.

## Step 1: Set Up Your Project

Project Name (a name is suggested after the read if you leave it blank). Visual Style, Scene Style, Costume Style: pick from the cards, or press Next with some unset and answer Let AI choose the remaining styles? with Let AI choose. The visual style cannot be changed once you continue — it shapes every sheet and video. Aspect Ratio: 16:9 or 9:16. Target Region and dialogue language if the story should move or speak another language. The settings.

## Reading the script

Next saves the settings and starts the read. The card shows Reading the script with stages — Finding characters, Finding scenes, Finding props, Splitting episodes, Shaping visual style, Preparing episodes — and counts as they come in. You can browse elsewhere; SceneMixer's corner capsule keeps the progress and you move on automatically when it is done. A long script takes longer and may prepare later episodes in the background while you work on the first. While it reads.

## Script Parsed

Script Parsed — AI has finished reading the script. On screen: the title, the protagonist, the era, the adaptation strategy, the episode list with Show more, and the counts of characters, scenes and items. Read the summary; if the split is wrong, this is the moment to fix the text and read again. Checking the summary. Next goes to step 2, where the cast has descriptions and no images yet.

## If it fails

Parsing failed. Credits have been refunded. Please try again: the read did not complete and nothing was charged. Press Next again. A text that is not a story — a table, a list — can fail repeatedly; give it prose.

| Click | What happens |
| Upload Script tab, paste or Upload File | Text in the box; price on the button updates |
| Get Started ✦ N | Project created; step 1 opens |
| Styles, aspect ratio, region | Saved when you press Next |
| Next | The read starts; progress card and corner capsule |
| Script Parsed, then Next | Step 2 with the cast, sets and props |

## FAQ

### What files can I upload?

SceneMixer accepts TXT, DOCX and Markdown, or pasted text. Any prose or script; it is split into episodes for you.

### What does the read cost?

By length: the rate per 1,000 characters is on the button before you press it.

### Can I change the visual style later?

No. It shapes every sheet and video, so SceneMixer fixes it once you continue. Scene and costume styles apply at the read too.

### Can I leave while it reads?

Yes. SceneMixer's corner capsule keeps the progress and the project moves on by itself.

### The read failed. Was I charged?

No. SceneMixer's message says the credits were refunded. Press Next again.

Updated November 3, 2026


---

# Use a Character State in One Scene: Segment, Episode and Default Picks

URL: https://scenemixer.com/guides/use-a-character-state-in-one-scene/
States are extra reference sheets for the same character, scene or prop. Where to pick which state a segment or episode uses, the priority order, states without an image.

# How to Make One Scene Use a Character’s State (Costume, Damage, Drawn Blade)

Quick answer: A state is a second sheet for the same character, scene or prop — the wedding dress, the burnt kitchen, the sword out of its sheath. Which sheet a shot uses is decided in four layers, highest first: the pick on that segment, a custom cover image set for the episode, the pick for the episode, then the base sheet. After a storyboard is generated the site assigns states to segments automatically from the story; you can override any segment in the editor, and set an episode-wide pick in the episode’s asset panel. A state without a generated image is not used; the shot falls back to the base sheet plus a line of text.

In the editor, open the asset panel on the left, click a character, and pick a state for this segment; the chip on the timeline card updates.

## The four levels, in order of priority

1. Segment pick. In the editor, select the segment, open the character (or scene, prop) in the left panel and choose a state. This wins over everything else, for that segment only. 2. Episode cover image. A custom image set on the episode’s asset entry replaces the sheet for the whole episode; it cannot be combined with an episode state pick. 3. Episode pick. The same panel, applied to every segment of the episode that has no segment pick. 4. Base sheet. What every shot uses when nothing else is set. How states are created and edited: Costume Changes and Scene States.

## What the site picks for you

When an episode’s storyboard is generated, SceneMixer reads which segments call for which state — the flashback outfit only in the flashback, the drawn blade only in the fight — and writes a pick on those segments. Your own segment picks replace these. Regenerating the storyboard rebuilds the segments and redoes the automatic picks, so hand picks made before are gone; episode picks survive.

## A state that has no image yet

Only a state with a generated image is ever used. If the state you picked has no image, or its image is out of date because the base sheet was regenerated, the shot uses the base sheet and the state is described in text instead (“wedding dress” goes into the character line). Generate the state’s image from its card in Step 2 — Generate All queues state images along with the sheets and prices them together.

## Who can do what

Using the states SceneMixer created, and choosing which segment or episode uses them, is free of any plan. Creating a state by hand, renaming it, editing its description or deleting it is a member feature. Generating a state’s image is priced like a sheet, with its own small free allowance in the first project. What a Paid Plan Changes.

## Checking before you generate

The timeline card shows chips for the references a segment will use; a state chip carries the state’s name. Hover or tap to confirm, then generate. If the video shows the wrong outfit, check the segment pick first, then whether the state image exists. Timeline Cards: Every Icon.

| Level | Where to set it | Applies to |
| Segment pick | Editor → left panel → the character/scene/prop → state | That segment |
| Episode cover image | Episode asset panel → custom image | Every segment of the episode |
| Episode pick | Episode asset panel → state | Segments without a segment pick |
| Base sheet | Nothing to set | Everything else |

## FAQ

### I created states but the video ignores them.

Check that the state has a generated image, then that the segment's pick points to it. A state without an image falls back to text on SceneMixer.

### Can I set a state for the whole project?

No; SceneMixer's picks are per episode or per segment. Set the episode pick in each episode that needs it.

### Why did my picks disappear?

You regenerated the storyboard. SceneMixer rebuilds segment picks automatically; set them again if the automatic ones are wrong.

### Do states cost credits?

Generating a state's image costs credits on SceneMixer, like a sheet. Choosing which segment uses it does not.

### Can a scene have a state too?

Yes: the burnt kitchen, the flooded street. Same picks, same priority.

Updated November 3, 2026


---

# Use a Screenplay as Input: Scene Headings and Character Cues

URL: https://scenemixer.com/guides/use-a-screenplay-scene-headings-and-character-cues/
Bring a screenplay into SceneMixer: which file to export, how scene headings, character cues and parentheticals come through, and the one change that keeps dialogue safe.

# Using a Standard Screenplay as Input: Scene Headings, Character Cues and Parentheticals

Quick answer: A screenplay works as input like any other script. Export it from your screenwriting app as plain text or a Word file (Fountain files are plain text: rename them to .txt), or paste the pages in. Scene headings such as INT. KITCHEN – NIGHT help the reading find places and times; keep them. For dialogue you want spoken word for word, the safest form is the name and the line on one line, MARA: Read it before you say no., with a parenthetical in brackets before the colon. Transitions like CUT TO: can stay or go; the storyboard decides the cuts itself.

Paste one scene of your screenplay into a new project and compare the step 1 outline with the page.

## Getting the file in

The script box takes pasted text and .txt, .docx and .md files. PDFs are refused when you pick them, so export from your screenwriting app as plain text or as a Word document instead. A Fountain file is already plain text; rename it from .fountain to .txt. Which script files work.

## Scene headings

Keep them. INT. or EXT., the place, and the time of day tell SceneMixer where each scene happens and whether it is day or night, and they help it tell two rooms apart. Use the same name for the same place every time: MARA’S KITCHEN in one scene and KITCHEN in the next can become two locations.

## Character cues and dialogue

Screenplay layout puts the name above the line. That is often understood, but for lines that must be spoken exactly as written, put the name and the line on one line: MARA (quietly): Read it before you say no. Use one spelling of each name throughout; a character called MARA in one scene and MRS. CHEN in another becomes two characters on SceneMixer. Write dialogue so it is picked up.

## Parentheticals, V.O. and O.S.

A parenthetical becomes the delivery note: (quietly), (furious, under her breath). Voices heard but not seen are written as a note before the line: (voice-over), (on the phone), (over the radio). Dialogue lines, voice-over and captions.

## Action lines

Write what can be seen and heard, as you would in a screenplay anyway. Camera directions in the action are fine but not needed: the storyboard plans its own shots, and you can edit them after the reading. Long stretches of description are condensed.

## Length

A feature screenplay is longer than one project reads at once; adapt it in parts. SceneMixer's script box says when a paste is over the limit. Length caps.

| Screenplay element | Keep it? | Tip |
| Scene heading (INT./EXT., place, time) | Yes | One name per place throughout |
| Character cue | Yes | Safest as NAME: line on one line |
| Parenthetical | Yes | Becomes the delivery note |
| V.O. / O.S. | Yes | Write as (voice-over) or (on the phone) |
| CUT TO:, FADE IN: | Optional | The storyboard decides cuts |

## FAQ

### Can I upload a Final Draft file directly?

No. Export it as plain text or a Word document first, then upload it to SceneMixer.

### Can I upload a PDF of my screenplay?

No. SceneMixer does not take PDFs; copy the text out of it, or export from your screenwriting app as text or Word.

### Will my scene numbers be kept?

They come through as text; SceneMixer's reading builds episodes from the story.

### Do I need to rewrite the whole screenplay?

No. Keep it as it is; change only dialogue you need word for word to NAME: line form.

Updated November 3, 2026


---

# Use AI Drama Clips as Previs and Animatics Before a Real Shoot

URL: https://scenemixer.com/guides/use-ai-drama-clips-as-previs-and-animatics/
Turn a script into a timed shot list and moving rough cuts before a shoot: check pacing, pitch the scene, brief the crew. What it is good for, and where it stops.

# Using SceneMixer for Previs and Animatics Before a Live-Action Shoot

Quick answer: Paste a scene or a short script and SceneMixer gives you a storyboard — numbered shots with framing and seconds — and then a moving version of each segment, which you can join into a rough cut of the scene. For a live-action shoot that is a quick animatic: you see the pacing, hear the lines at speaking pace, and have something to show a client or a crew. Draft on a low tier, copy the storyboard text out as a shot list, and share links for review. Treat it as a sketch: the clips show intent, not exact lenses, positions or performances.

Paste one scene of your shooting script, generate its storyboard, and copy the first segment into your shot-list template.

## What you get

- A shot list: each segment’s storyboard numbers its shots and gives framing, action, dialogue and seconds.
- Moving sketches: a clip per segment, with the lines spoken.
- A rough cut: segments joined into one file per episode.

Using a screenplay as input.

## Pacing and timing

The storyboard gives every shot a length in seconds and makes room for each line to be said. Reading the totals tells you early whether a scene runs long. Change the seconds in the text before generating, for free. Shot seconds and timing.

## Cheap drafts

For previs, a low tier is usually enough: it is cheaper per second and still shows blocking, framing and rhythm. Pick the aspect ratio of the real shoot in SceneMixer’s step 1. Draft at a low tier · Aspect ratio.

## Sharing with a client or a crew

Share a segment or a joined episode with a link; viewers do not need an account. Copy the storyboard text into your own documents for the shooting script or the call sheet. Share links · Copying storyboard text.

## Where it stops

The clips are generated, not photographed: lenses, exact distances and performances are suggestions. Faces come from character sheets, not your cast, unless you upload their photos with their permission. Use the previs to agree on story, rhythm and coverage, then plan the real camera on set.

| Need | In SceneMixer |
| Shot list | The storyboard, copied out |
| Pacing check | Shot seconds and the segment totals |
| Moving rough cut | Segments joined into one episode file |
| Client review | Share links, no account needed |
| Low cost | Draft on a low tier |

## FAQ

### Can I use photos of my real actors?

Only with their permission, and uploading on SceneMixer needs a paid plan.

### Will the camera moves match what we shoot?

Not exactly. Treat them as intent; plan the real lenses on set.

### Can I export the storyboard?

Copy the text out of the SceneMixer editor; it pastes as plain text into any document.

### Is it suitable for a commercial board?

Yes, for pacing and pitching. For product ads made start to finish here, see the ad guides.

Updated November 3, 2026


---

# Use Your Own Photo for a Character or Scene | SceneMixer

URL: https://scenemixer.com/guides/use-your-own-photo/
Upload your own photo instead of generating a sheet: optimise it or use it as is. What each does, which photos work best, and why old segments need regenerating.

# Using Your Own Photo

Quick answer: Open a card in step 2 and press Upload & Replace. A dialog asks How should we use this image? Optimize into a reference sheet reads the face, hair and clothing from your photo and regenerates it in the project’s style, on a clean background; Use this image as is puts the photo on the card untouched. Optimise for snapshots — a snapshot’s background, lighting and pose would otherwise follow the character into every shot. Use as is for artwork you made to be a sheet. JPG or PNG; uploading is a membership feature on the current plan settings.

Open a finished episode and click a character: a reference sheet is a full figure on a plain background, which is what optimise produces.

## Where

Step 2, click the entry, Upload & Replace. Or drag a file onto the card. JPG or PNG; other common formats are converted, SVG is not accepted, and an oversized file is refused with a message to compress it. On the current plan settings uploading is a membership feature and the button says so; photos for an ad are the exception. Paid plans.

## The two choices

Optimize into a reference sheet. The dialog explains why it is the default: a reference sheet is the baseline every later shot inherits, and a snapshot brings its background, lighting and pose into the whole drama. SceneMixer reads the face, hair and clothing from your image and regenerates it together with the script’s setting — the identity details from your photo, the style and staging from the project. This is priced like an image.

Use this image as is. The photo goes on the card exactly as uploaded, marked Uploaded. Choose this for artwork already made to be a sheet: a full figure, plain background, even light. It costs nothing.

## What happens next

The card shows the new image. For characters, a headshot and a half-body view are taken from it when the face is clear; when it is not, the whole photo is used. From then on, videos follow the photo: the face, skin tone and clothes come from it rather than from the description written when the script was read. Segments generated before the upload keep the old look until you regenerate them. Generate All never overwrites an uploaded image. If the photo’s proportions are unusual the card says so (Unusual aspect ratio — consider regenerating). Pressing Regenerate later replaces the upload with a generated image; upload again to go back.

## Which photos work best

One person, face clearly visible and turned toward the camera, even light, nothing covering the face. A full-length photo works, but the smaller the face, the less of it the video can hold on to; a waist-up photo usually does better than a distant one. Avoid group photos, heavy filters, strong coloured light and sunglasses.

## AI Describe: a description from the image

An uploaded photo has no description behind it, and the description is what the storyboard reads. SceneMixer's AI Describe looks at the image and writes the appearance and clothing fields for you; Describe all Characters does it for every card with an image. Read the result and correct it — it is the text every later shot will trust. Writing a description.

## Photos of real people

Use only images you have the right to use. A recognisable public figure or a franchise character will be refused by the IP check when you generate video. Rights · Blocks.

The first time you upload a file, a window titled Before you upload asks you to confirm that you have the right to use the material (scripts, images, video, audio) and that every real person whose face or voice appears in it has given consent. Press Confirm and upload to go on; you only need to confirm once, several files in one go ask once, and cancelling stops the upload.

## Scenes and props

The same button on a scene or prop card in SceneMixer. A photo of a real place is a good scene image if it shows the space as a whole; optimise it if it has people in it or a strong time of day. Locations.

## Voices

Your own voice recording goes on a character with Upload Voice, separately from the image. Character voices.

| Your image is | Choose | Cost |
| A snapshot, a selfie, a frame from a video | Optimize into a reference sheet | Priced like an image |
| Artwork made to be a sheet: full figure, plain background | Use this image as is | Free |
| A photo of a real place | As is, or optimise if people or strong light are in it | Free / priced |
| A voice recording | Upload Voice on the character | Free |

## FAQ

### Which option should I pick for a photo of a person?

Optimize into a reference sheet. SceneMixer keeps the face and clothing and drops the background, lighting and pose that would otherwise leak into every shot.

### I uploaded a photo. Why do old segments still show the previous face?

They were generated before the upload. Regenerate those segments; SceneMixer uses the photo for new ones.

### Will Generate All replace my upload?

No. SceneMixer never overwrites uploaded images. Only pressing Regenerate on that card replaces it.

### Does the uploaded photo need a description?

Yes, for the storyboard's sake. Press SceneMixer's AI Describe to write one from the image, then correct it.

### Why is Upload greyed out?

Uploading on SceneMixer is a membership feature on the current plan settings; the button says so.

### Can I upload a photo of a celebrity?

No. Use images you have the right to use; SceneMixer refuses recognisable figures at video time.

Updated November 3, 2026


---

# Using SceneMixer as a Team: Review, Hand-offs and Accounts

URL: https://scenemixer.com/guides/using-scenemixer-as-a-team/
There is no shared team workspace: each project belongs to one account. How teams still work together: one owner per project, share links for review, files for hand-off.

# Using SceneMixer as a Team

Quick answer: SceneMixer has no team workspace or shared editing: a project belongs to the account that created it. Teams usually work this way: one person owns and generates each project; others review through share links, which open without an account; scripts and storyboard text travel as text; finished episodes travel as downloaded files. Activity under an account is that account holder’s responsibility, so keep your sign-in to yourself. For studio or company needs, write to us through the contact page.

Share one finished segment and open the link in a private browser window: that is exactly what a reviewer without an account sees.

## One owner per project

A project, its cast, its storyboards and its videos live in one account, and only that account can edit or generate. Pick an owner for each project, usually the person who will press Generate and watch the credits. Several projects can have different owners.

## Review without an account

Share a segment or a composited episode and send the link. Reviewers watch it in a browser without signing in; they cannot edit anything. Collect notes by segment number so the owner can fix the right shot. Share links: who can see them.

## Writers and editors

- Script: the writer delivers a TXT, DOCX or Markdown file; the owner starts the project with it.
- Storyboard notes: copy a segment’s text into a shared document for comments; the owner pastes changes back and checks the links after saving. Copying and pasting storyboard text.
- Post-production: the owner downloads the episodes; an editor adds titles and captions in their own software. SceneMixer gives every newly made episode its own score; episodes that had clips before scoring rolled out, in early October 2026, are not scored. To lay a different track under one, turn the Music switch off in the editor — the switch shows only on clips that have a scored version — and download the version without music. The built-in score.

## Keeping a cast across owners

Characters can be copied only between projects in the same account. If two people make parts of the same series in different accounts, agree on the sheets first: the second owner can upload the same character images to their project, which needs a paid plan. Use your own photo or image.

## Accounts and sign-in

Everything done under an account is that account holder’s responsibility. Do not hand out your sign-in; send links and files instead. Terms.

## Company or studio use

For larger volumes, invoicing questions or a studio setup, write to us; there is a card for business enquiries on the contact page.

| Team task | How |
| Edit and generate | The project owner's account |
| Review cuts | Share links, no account needed |
| Comment on storyboard text | Copy it into a shared document |
| Final edit and captions | Downloaded files in your editing software |
| Your own music instead | Turn the Music switch off, download, then lay your track in the editor |
| Business terms | The contact page |

## FAQ

### Can two accounts edit the same project?

No. A project on SceneMixer belongs to one account.

### Can I move a project to someone else's account?

No. Hand off the script, the sheets as images, and the downloaded videos instead.

### Do reviewers need an account?

No. SceneMixer's share links open without signing in.

### Can several of us share one login?

Activity under an account is the holder's responsibility; we recommend each person keeps their own account and you share links and files.

Updated November 3, 2026


---

# Using SceneMixer on a Phone: What Is Different From the Desktop

URL: https://scenemixer.com/guides/using-scenemixer-on-a-phone-what-is-different/
The whole site works in a phone browser on the same account. What changes on a small screen: a stacked editor, the model picker in a bottom sheet, the share menu.

# Using the Site on a Phone: Same Account, Same Projects, a Stacked Editor, and a Few Things Easier on a Desktop

Quick answer: Sign in on your phone and everything is there: projects, credits, generations in progress. The pages adapt to the screen; the biggest change is the episode editor, which on a phone stacks into views — the clip preview, the timeline of segments, and the storyboard text — instead of three columns side by side. The model picker and the AI-badge switch live in a bottom sheet; sharing adds the phone’s own share menu; uploads come from your photos. Generations run on the server, so starting one on the phone and checking on the desktop, or the reverse, is normal. Pasting a long script and editing storyboard text at length are easier with a keyboard and a wide screen; nothing is phone-only or desktop-only.

Open an episode on your phone: swipe between preview, timeline and text; tap the model name to open the tier sheet.

## The editor on a phone

Three views instead of three columns: preview (the clip, play controls, download, share), timeline (segment cards with their prices and states), and the storyboard text of the selected segment with its edit button. The cast and scene rail is behind a button. Everything the desktop editor does is reachable; it is arranged for one hand. The Episode Editor: A Tour.

## Model picker and badge switch

Tap the model name to open the tier sheet; the AI-badge control sits at the bottom of that sheet. Choices are saved to your SceneMixer account and apply on the desktop too.

## Sharing and downloading

Share shows the same platform list as on the desktop, plus a Share… entry that opens the phone’s own share menu with the link. Download saves the original-quality file; the preview plays the lighter play version. Download: Original vs Play Version.

## Uploads

Sheets and reference photos (a paid-plan feature, as on the desktop) and ad photos upload from your camera roll or files app; the size and format rules are the same as on the desktop. Scripts can be pasted from any app; .docx and .txt files from your files app work too.

## Easier on a desktop

Pasting and reading a long script, editing several storyboard segments in a row, comparing versions of a clip side by side, and reading the Step 1 summary. SceneMixer blocks nothing on the phone; these are just faster with a keyboard and space. What follows you between devices: What Follows Your Account.

| Task | Phone | Desktop |
| Watch, download, share clips | Preview view | Right column |
| Start a generation | Timeline card | Timeline card |
| Pick a tier, badge switch | Bottom sheet | Selector in the top bar |
| Edit storyboard text | Text view (fine for a line or two) | Middle column (better for long edits) |
| Paste a long script | Works | Easier |

## FAQ

### Is there anything I cannot do on the phone?

No SceneMixer feature is phone-only or desktop-only.

### Where is the badge switch on the phone?

At the bottom of SceneMixer's model sheet.

### Can I start a generation on the phone and download on the desktop?

Yes; the clip is in your SceneMixer account.

### Why does share look different?

SceneMixer adds a Share… entry for the phone's own share menu; the platform list is the same.

### Does the phone use more credits?

No. SceneMixer's prices are the same everywhere.

Updated November 3, 2026


---

# Vertical or Landscape: Aspect Ratio for an AI Drama | SceneMixer

URL: https://scenemixer.com/guides/vertical-or-landscape-aspect-ratio/
Where the aspect ratio is set, what it changes (the video) and what it does not (the images), why it is fixed after step 1, and when to pick 9:16 for a vertical drama.

# Vertical 9:16 or Landscape 16:9

Quick answer: In step 1, under Aspect Ratio, pick 16:9 Landscape (the default, marked recommended) or 9:16 Portrait. It sets the shape of every video segment in the project. It does not change the character or scene images: character sheets are always portrait and scene images always widescreen, whichever ratio the videos use. Pick 9:16 for a phone-first vertical drama, 16:9 for anything watched on a screen wider than it is tall. The choice is per project and is fixed once you continue past step 1.

Open a finished landscape episode; the frame you see is what 16:9 gives you.

## Where it is

Step 1, Set Up Your Project, the Aspect Ratio row: two options, 16:9 Landscape — Film / Large screen and 9:16 Portrait — Short drama / Mobile. Landscape is pre-selected and carries the Recommended badge.

## What it changes

The frame of every video segment in the project, and therefore of the episode composite. Shots are framed for the ratio you chose: a 9:16 drama gets close framing that fills a phone screen; a 16:9 drama gets wider staging with more of the room in view.

## What it does not change

The images in step 2. Character sheets are always portrait, because a full-body character sheet is tall. Scene and prop images are always widescreen, because they describe a place or an object rather than a frame. Changing the aspect ratio on SceneMixer never regenerates an image. Why scene images are widescreen.

## Which to pick

9:16 when the drama will live on phones: vertical-first platforms, feeds, the ReelShort-style format. 16:9 when it will be watched on a laptop, a TV or a site with a wide player, and for anything with more than two people in a shot, which vertical framing crowds. Making ReelShort-style dramas.

## Can it change later

The ratio is chosen on the step 1 settings screen, together with the visual style, and both are fixed once you continue: after the script is read, step 1 shows the summary and the ratio as read-only. To make the same story in the other shape, start a new project with the same script and pick the other ratio; you can bring the cast across from the asset library. What can still change after the read.

## Ads

An ad picks its aspect ratio in SceneMixer's ad composer, per ad, with the same two options and the same consequences. Ad options.

| Ratio | Default | Good for | Images in step 2 |
| 16:9 Landscape | Yes (recommended) | Laptop, TV, wide players, group shots | Unchanged |
| 9:16 Portrait | No | Phone-first vertical drama, feeds | Unchanged |

## FAQ

### Which aspect ratio is the default?

16:9 landscape, marked Recommended in SceneMixer's step 1. Choose 9:16 for a vertical drama.

### Will choosing 9:16 make my character images vertical?

Character sheets are always portrait and scene images always widescreen, whatever ratio the videos use. The aspect ratio setting on SceneMixer affects video only.

### Can I change the ratio after generating videos?

No. SceneMixer fixes the ratio once you continue past step 1. For the other shape, start a new project with the same script.

### Is 9:16 more expensive?

No. Video is priced by the second, not by the shape of the frame.

### Can an episode mix both?

No. Every segment in a SceneMixer project uses the ratio chosen in step 1.

Updated November 3, 2026


---

# Micro Drama Script Format, Template and Example (2026) | SceneMixer

URL: https://scenemixer.com/guides/vertical-short-drama-script-format/
Micro drama (vertical short drama) script format: episode length, words per second, a copy-ready scene template, hook structure, and what AI generation needs.

Guide · Writing

# Vertical Short Drama Script Format (2026): Episode Length, Word Counts, Hook Structure and AI-Generation Constraints

Published September 5, 2026 · Updated September 20, 2026

Quick answer: A vertical short drama script is organised by episode, not by act: each episode runs 1–2 minutes, carries 180–360 spoken words in 2–4 scenes, opens on conflict inside 5 seconds, turns roughly every 30 seconds and ends on a hook that forces the next tap. The format the apps in the ReelShort and DramaBox mould have settled on, as publicly visible in their catalogues as of 2026-09-05, is 60–100 episodes of 60–120 seconds, with a paywall somewhere between episode 8 and 12. Below: a length and word-count table, a copy-ready scene layout, the hook structure with a first-three-episodes beat sheet, and the extra constraints that apply when the series is generated by AI rather than shot.

## Three ways a vertical script differs from a feature or TV script

- Scene density. A feature can spend three pages on one scene; a 90-second episode holds 2–4 scenes and keeps only the one or two beats that move the conflict.

- Line length. Single lines of dialogue stay under about 20 words. A phone caption line holds 6–8 words; a long line wraps to three rows and the viewer reads captions instead of watching the face.

- The hook is per episode. TV relies on one cliffhanger per hour; vertical drama needs one per episode, and the three hook types (identity, emotion, realisation) have to rotate. Two identical hooks in a row feel like a trick.

## 2026 length table

| Format | Episode length | Episode count | Spoken words per episode | Season total | Basis (checked 2026-09-05) |

| Live-action vertical drama (ReelShort / DramaBox style) | 60–120 s, some to 3 min | 60–100 | 180–360 | 10,800–36,000 | Publicly visible catalogue norms on the major apps; word counts derived at 3 words/s |

| AI-generated vertical drama | 60–120 s | 20–60 | 180–360 | 3,600–21,600 | Shorter seasons are the norm while per-second generation cost dominates the budget |

| Mid-form (3–20 min) | 3–20 min | 10–30 | 540–3,600 | — | Same pacing formula, longer arcs |

The conversion rests on one number: a typical spoken pace of about 3 words per second. A 90-second episode is roughly 270 words of dialogue and voice-over; action lines do not count. Work the formula backwards after every draft instead of guessing the running time.

## Scene layout: an example you can copy

There is no single mandated template, but the layout below, organised by episode and scene, is what studios read fastest and what script-analysis tools split most reliably:

EPISODE 3

SCENE 1 · INT. VALE FAMILY CHAPEL – NIGHT
[ACTION] Claire kneels on the cushion, a half-burned marriage contract clenched in her hand. The door opens; the draft bends the candle flames.
MR. VALE (low): Did you burn the contract?
CLAIRE (not looking up): I did.
MR. VALE: Then you can explain it to the Carters yourself.
[ACTION] She lifts her head. No tears.

SCENE 2 · EXT. VALE HOUSE, FRONT GATE – NIGHT
[ACTION] Ethan Carter's car idles at the gate, headlights on her.
ETHAN (through the window): Get in. The wedding goes ahead.
[HOOK · realisation] On the passenger seat: the other half of the contract, intact.

- Scene heading in a fixed order: scene number, INT./EXT., location, time. Keep one spelling of every location for the whole season ("Vale Family Chapel", never "the chapel" one page and "Vale chapel" the next); it is what scene reference images and consistency are built on later.

- Action lines marked [ACTION] or indented, only what can be seen or heard, no interior monologue, under about 40 words each.

- Dialogue as CHARACTER (tone): line. Tone goes in parentheses; no "she said".

- Hook labelled with its type at the end of the episode so you can check the rotation yourself.

## Hook structure: 5 seconds, 30 seconds, episode end

### Three beats

- Conflict in the first 5 seconds. The first line or the first action is the conflict itself, not the setting. "Did you burn the contract?" beats "Morning at the Vale house."

- A turn about every 30 seconds. A 90-second episode turns at least twice: new information, a shifted allegiance, or escalated danger.

- The episode-end hook, rotating. Identity (someone is not who they said), emotion (a confession, refusal or rupture left hanging), realisation (viewer and lead discover an earlier judgement was wrong).

### First-three-episodes beat sheet

| Episode | What must happen | End hook |

| 1 | Conflict inside 5 s; the lead's situation and first trap; what she wants is clear. | Identity |

| 2 | The antagonist enters; the stakes go on the table (money, a marriage, a life, a reputation, pick one); the lead is cornered. | Emotion |

| 3 | First reversal: the apparent weakling shows a card, or an ally turns. | Realisation |

| 4–5 | The lead's first real win (a face-slap, a turnaround, a piece of evidence) and a bigger problem behind it. | Identity → Emotion |

The logic: free-to-watch feeds decide on completion of the first three episodes, paid apps convert at the paywall around episodes 8–12, and both need the first win no later than episode 5.

## Submitting to a studio or app

- ☐ Episode count and length match the studio's current call; most requests specify a full-season outline plus the first 10–20 episodes in full.

- ☐ Every episode carries a labelled hook; the three types rotate.

- ☐ One spelling per location and per character name across the whole season.

- ☐ Single dialogue lines under about 20 words; action lines under about 40.

- ☐ Genre fits what the app is buying this quarter (billionaire romance, revenge, werewolf and second-chance stories have dominated the English-language catalogues; check the front page of the app you are pitching on the day you submit).

- ☐ Rights are clean: adaptation rights for any source novel, no real-person likenesses.

## Extra constraints when the series is AI-generated

When the script feeds an AI video model instead of a crew, format is not the only test; the other is "can the model produce this picture". Patterns confirmed across a large volume of finished clips:

| Write freely | Write carefully | Avoid |

| Two-person dialogue close-ups, single-character emotional beats, hard light and backlight, rain and snow, in-car scenes | More than four people in frame, fights that need continuous choreography, fine hand work | Battles and crowds, complex machinery, readable text close-ups (contracts, phone screens) |

- Keep each episode to three or four physical locations: every extra set multiplies the cross-shot consistency problem.

- Convert dialogue at 3 words per second: a 15-second segment holds at most about 45 words; more than that means another segment.

- Write causes, not effects, in action lines: "a pendant lamp throws warm-white light from upper right", not "light brightens her forehead". It is a prompt-level rule, but building the habit at script stage saves storyboard rewrites.

A script in this layout goes into SceneMixer's analysis without reformatting: the "short drama" content type applies a per-episode running-time cap, rotates the end hooks between identity, emotion and realisation, and splits out characters, sets, props and the episode structure directly. How a scene then becomes a 15-second segment is covered in the continuity section of Character Consistency.

## FAQ

### How many words is one episode of a vertical short drama?

About 180–360 words of dialogue and voice-over per episode. At a typical spoken pace of 3 words per second that is 60–120 seconds of screen time; action lines are not counted because they are not spoken. A 60-episode season is roughly 10,800–21,600 words.

### What has to happen in the first three episodes?

Episode 1 opens on the conflict within 5 seconds, establishes the lead's situation and what she wants, and ends on an identity hook. Episode 2 puts the antagonist and the stakes on the table. Episode 3 delivers the first reversal so the viewer trusts the series will keep turning. The lead's first real win must land by episode 5 or completion rates collapse before the paywall.

### How is an AI-generated series script different from one for a live shoot?

Same structure, different picture constraints. AI video is reliable on two-person dialogue close-ups, single-character emotional beats, hard light and weather, and unreliable on more than four people in frame, continuous fight choreography, crowds and complex machinery. So write fewer ensemble scenes and long fights, keep each episode to three or four physical locations, and convert dialogue to seconds at 3 words per second.

### Is there a standard script format for ReelShort or DramaBox submissions?

The apps do not publish a single mandated template; producers and studios generally accept a standard screenplay layout organised by episode: scene heading with INT./EXT., location and time, action lines, character name and dialogue, and a labelled hook at the end of each episode. Use the same location names throughout and keep single lines of dialogue under about 20 words. Check the current submission guidelines of whichever studio or platform you pitch to.


---

# Video AI: Five Jobs Inside One AI Video, and Which You Can Touch

URL: https://scenemixer.com/guides/video-ai/
Video AI is not one model. Reading, drawing, shooting, scoring and joining are five jobs: what goes in and comes out of each, which you can see and which you can change.

# Video AI: The Five Jobs Inside One AI-Made Video

Quick answer: “Video AI” sounds like one thing that turns a request into a film. Inside one finished episode five different jobs happen, and only three of them are models you can pick. A text model reads your words into a cast and shots. An image model draws the cast and the places. A video model shoots each segment, picture, voices, lip movement and ambience made together. A scoring step chooses music for the episode. And a plain video tool joins the clips, with no model in it at all. Knowing which job produced a problem tells you where to fix it and what the fix costs, so this page lists each job’s input, output, price and where you can see and change it.

Try it: open any sample project and name the job behind five things: the summary in Step 1, a character picture in Step 2, a shot list in the editor, the Music switch beside the player, and the Composite Episode button.

## Job 1: Reading, done by a text model

In: your script, or the one line that SceneMixer’s AI Create expands. Out: a summary, characters with an appearance and a voice described in words, locations, props, an episode split, then for each episode a storyboard of numbered shots with seconds, names and dialogue lines. See it: Step 1 shows the summary; the storyboard is in Step 3 and in the editor as a Shots list or plain Text. Change it: edit any of it by hand for free, ask an AI edit in plain words at 1 credit per revision, or restore an earlier version from text history. Cost: the read is 4 credits per 3,333 English characters, once; a storyboard is 1 credit per 100 counted characters of its episode.

The text model draws nothing and shoots nothing. A storyboard is a format the rest of the system reads literally, shot markers, seconds and references to the cast, so its quality is the ceiling for everything after it. What is inside a project.

## Job 2: Drawing, done by an image model

In: each description plus the look you chose. Out: a sheet per character (always 9:16), per location and per prop (always 16:9), optional state images, and cover posters. See it: Step 2, one card per entry. Change it: edit the description and regenerate, or replace the sheet with your own image on a paid plan. Cost: the first 15 on an account are free, then 7 credits each on the default image tier.

These pictures are not the video. They are references that every later shot is made from, which is why sheets are neutral on purpose: a plain background and plain light, because the model copies the light and mood of a reference as faithfully as the face. Mood is written per shot in the storyboard. Why sheets look plain.

## Job 3: Shooting, done by a video model

In: the sheets of everyone and everything in the shot, the shot text with its seconds, and each speaker’s voice. Out: one segment of up to 15 seconds, or 30 on the long-take tiers, with the picture, the spoken lines, the speaker’s lip movement and the ambience of the place generated together as one clip. See it: the preview player, with every take kept in Version History. Change it: edit the shot text and regenerate that segment, or use Edit this clip for a trim, a picture change, a re-voice or an enhance. Cost: : 10 credits a second on the default tier, more on higher tiers; the price is on the button.

This is the job that costs real money and takes real time, typically ten to twenty minutes per segment. It is also probabilistic: two identical requests can give two different clips. The tier picker in the editor’s top bar sets the resolution, from 480P to 2K, and the longest segment it takes. Dialogue and sound in one generation.

## Job 4: Scoring

In: the finished storyboard, its scenes, the mood of each stretch and where the lines fall. Out: a plan for the episode’s music, pieces of at least half a minute that change at scene boundaries with a short crossfade, mixed in under the voices. See it: the Music switch beside the player. Change it: only on or off; there is no track picker, volume control or button to compose again. Cost: included in the video price. The built-in score.

## Job 5: Joining, which is not a model

In: the active take of every segment, in order. Out: the episode file. See it: SceneMixer’s Composite Episode button, available once every segment is done. Change it: switch a segment to another take or regenerate it, then join again; the earlier file stays until the new one replaces it. Cost: : 1 credit a second.

The join is plain video processing. It joins without re-encoding wherever it can, so the picture you approved is the picture that ships, and it cannot repair a clip that is wrong. Clips are joined as they are, so a mark that was on some segments and not others stays that way. Composite and share an episode.

## Which job caused this?

| What you see | Which job | Where to fix it |
| A shot shows the wrong person or action | Reading | Edit the shot text, then regenerate that segment |
| The same face is wrong in every shot | Drawing | Edit the description in Step 2 and regenerate the sheet |
| One shot is odd, the others are fine | Shooting | Regenerate that segment, or Edit this clip |
| A line is spoken in the wrong voice | Shooting, from the voice tag | Check the speaker and voice tags in the shot, then the character’s voice |
| The music is not wanted | Scoring | Turn the Music switch off and download |
| Some parts of the episode carry the badge, some do not | Joining | Set the badge switch, regenerate the odd segments, join again |

## Which of these you can choose a model for

In SceneMixer, text, image and video each work through a tier: a row of configuration that says which service, which model, what resolution and how long a segment may be. The video tier is the one worth choosing, because that is where the money and the visible difference are; its picker is in the editor’s top bar. The text and image tiers live in Settings under Model Preferences, and the defaults are the ones the pipeline was tuned around. This page names no models because tiers are configuration and change; the picker and the pricing page show the live list. What runs underneath.

| Job | Input | Output | You can change | Cost |
| Reading | Your words | Cast, locations, props, shots | Everything, by hand or by AI edit | By length; hand edits free |
| Drawing | Descriptions and the look | Sheets and state images | Description, regenerate, upload (paid plan) | 15 free, then per image |
| Shooting | Sheets, shot text, voices | One clip with sound | Shot text, tier, clip fixes | Per second |
| Scoring | The storyboard | Music under the voices | On or off | Included |
| Joining | Finished segments | The episode file | Which take, then join again | 1 credit a second |

## FAQ

### What is video AI?

A set of models and tools that read a text, draw reference pictures, shoot clips with sound, score them and join them. On SceneMixer it is five jobs, and three of them use models you can pick.

### Is one AI model making the whole video?

No. SceneMixer uses a text model, an image model and a video model for different jobs, and the join is ordinary video processing. The music is planned and mixed by the site.

### Which part costs the most?

Shooting. SceneMixer charges video per second, and every other step is small by comparison.

### Can I choose the model?

You can choose the video tier in SceneMixer’s editor and the text and image tiers in Settings. The pricing page shows the live list; defaults are what the pipeline was tuned on.

### Why can I not edit the sound separately?

SceneMixer generates voices, lip movement and ambience with the picture as one clip. You can re-voice a finished clip, turn the score off, or regenerate the segment.

### Does the AI read my script literally?

SceneMixer reads it into a cast and shots, adding what a storyboard needs. You can read and edit the result before any video is paid for.

### Is there a human in the loop?

You are. The two points where you check are the cast in SceneMixer’s Step 2 and the storyboard before video. Both are free to edit.

### Why do two identical requests give different clips?

Video generation is probabilistic. That is why you read the storyboard first, shoot one segment before the rest, and regenerate single shots.

Updated November 2, 2026


---

# Video Editor With AI: What the AI Fixes and What Each Fix Costs

URL: https://scenemixer.com/guides/video-editor-with-ai/
What an AI video editor can fix in a finished clip: new wording, one of eight picture changes, a new line, a trim. What each costs against regenerating, and the limits.

# Video Editor With AI: Fixes, Prices and Limits

Quick answer: In SceneMixer the AI does not cut footage on a timeline. It fixes a finished clip in a handful of ways, and each has a price you can set beside the price of regenerating the segment. Trim cuts the start, the end or a moment for a flat fee. AI visuals changes one thing from a menu of eight at 2 times the per-second generation price. Redub changes a line or a voice, by lip sync at 1 credit a second or by re-performing at the AI visuals price. Enhance raises the resolution one step. And AI edit rewrites the storyboard text from one sentence for 1 credit a revision. This guide puts them in the order to try them, cheapest first, and says what none of them can change.

## The repair ladder: try the cheapest that can work

| Try | It fixes | It costs | The rest of the clip |
| Switch to an earlier take | A better take already exists | Free, instant | Whole clip swaps |
| Edit the shot text by hand, then regenerate | Anything the text said wrongly | Editing is free; regenerating is the tier’s price per second | Made again from scratch |
| AI edit on the text | Wording, mood, camera, who is in a shot, even the number of shots | 1 credit per revision shown | Text only; regenerate to see it |
| Trim | Starts early, runs on, a stumble in the middle | 5 credits per trim, whatever the length | Kept exactly; nothing new is generated |
| Redub: lip sync | A line says the wrong thing, or the voice is wrong | 1 credit per second of clip | Picture kept; all lines re-voiced; ambience not kept |
| AI visuals, or redub: re-perform | One thing in the picture, or a line acted again (re-perform: clips of 15 seconds or less) | 2 times the tier’s per-second price, per second of clip | Kept apart from the one change; small details can differ |
| Enhance | Right, but soft | Per second, by option and target resolution | Same take, one resolution step up |

SceneMixer saves every fix as a new version of the segment, and the clip you started from is kept, so switching back is free. None of these has a free allowance, but the credits are held when you press and come back if the fix fails. Edit this clip: the four tabs · What each fix costs.

## A worked example: one 10-second clip that is nearly right

Take a 10-second clip made on SceneMixer’s default tier. What each way of fixing it costs, before any rounding:

| Way | Price | Compared with regenerating |
| Regenerate the segment | 10 seconds at 10 credits a second | The baseline |
| AI visuals (one change) | 10 seconds at 20 credits a second | 2 times the baseline |
| Re-perform | The same as AI visuals | 2 times the baseline |
| Lip sync | 10 seconds at 1 credit a second | A small fraction of it |
| Trim | 5 credits | A small fraction of it |
The pattern holds on every tier: a picture fix costs 2 times a regeneration of the same length, and a voice-only fix costs a small fraction of either. So AI visuals is worth it when the clip is right apart from one thing and a regeneration would roll the dice on everything else: the acting, the framing, the timing you already like. Regenerating is cheaper when the storyboard is wrong, or when the clip was made on an expensive tier and a cheaper one would do.

## What the AI decides, and what you decide

| Tool | You provide | The AI decides | The check |
| AI visuals | One change from a fixed menu, its option, and who it applies to | How the clip looks with only that changed | No free-text box, on purpose: these are the changes that come out reliably |
| AI edit (text) | A request in your own words, up to 500 characters | Which words, shots and seconds change | Shots, seconds, names and language must match, or the revision is dropped at no charge; you press Save |
| Redub | The new words and the voice for each line | Nothing about the picture | Lines are checked to fit the clip before you are charged |
| Trim | Where the clip starts and ends | Nothing; no AI | Cut at the exact frame, picture and sound together |

SceneMixer’s eight picture changes are clothes colour, time of day, weather, background people, lights, a scar, an expression, and swapping one character for another. On a 30-second clip you choose the first or second half, and only that half is edited and charged. AI visuals: what it can change · Change a line after the video is made · The eight AI edit starters.

## What none of the fixes can change

All of them start from the clip you have, so they keep how it was shot. They cannot change the framing (close-up versus wide), the camera angle, where people stand or which way they face, and they cannot remove a person who should not be there or fix words that appeared in the picture. Those come from the storyboard: edit the shot text and generate the segment again. When a shot comes out wrong.

And there is no timeline for colour grading, titles over footage, mixing with real footage or frame-accurate cuts between clips. For those, download the original-quality file and use an editor made for that. AI video editor covers where the line between the two sits.

## The one button that costs the most

In SceneMixer, regenerating a segment is safe. Regenerating an episode’s storyboard is not: it rebuilds every segment’s text and discards the finished videos with them, with no refund. If the text is the problem, fix it with AI edit or by hand, which keeps the videos, and use History to go back if you regret a change. Regenerate a storyboard: what you lose · Text history.

## FAQ

### Can AI edit a video after it is made?

Yes, in set ways: trim it, change one thing in the picture from a menu of eight, change a line or voice, or raise the resolution one step. Each saves a new version and keeps the clip you started from.

### Is there an AI video editor with a timeline?

No. SceneMixer’s timeline lists the segments of an episode. Changes are made to one segment at a time by editing its text or using the fixes. For frame-accurate edits, export the file.

### What does an AI fix cost compared with regenerating?

A picture fix is 2 times the per-second price of regenerating the same length. A voice-only fix and a trim cost a small fraction of either.

### Can I type exactly what I want changed in the picture?

No. SceneMixer’s AI visuals offers eight changes only. Anything else is a storyboard edit and a regeneration.

### What if the fix fails?

SceneMixer holds the credits when you press, and they come back if it fails. A fix blocked by content review is refunded too.

### Do I pay for an AI edit I discard?

Yes, 1 credit: the revision was made and shown to you. Failed, refused or unusable attempts cost nothing.

### Can I undo a fix?

Yes. Every SceneMixer fix is a new version; switching back to the one you started from is free. Text changes can be restored from History.

### Can it change the camera angle?

Not on a finished clip. Change the camera in the shot text, by hand or with AI edit, then regenerate the segment.

Updated November 2, 2026


---

# Video Generation Paused After Repeated Moderation Blocks: The Rules

URL: https://scenemixer.com/guides/video-generation-paused-after-repeated-moderation-blocks/
When the provider's content check blocks your video generations repeatedly in a day, new submissions pause for a while. The counts, what resets, what is refunded.

# ‘Content Review Was Triggered Several Times, Video Generation Is Paused’: How the Brake Works

Quick answer: Each time a video generation is blocked by the provider’s content check, it counts in a 24-hour window that starts at your first block. Once the count goes past 12, video submissions from your account are refused for 1 hour. The count does not reset when the pause ends: within the same 24 hours, one more block starts another pause. Only video submissions are paused; sheets, storyboards and everything else continue. Blocks within the daily number are refunded automatically once the provider confirms them; the block that goes over the number, and every later block in the same 24 hours, is held for review instead. The way out is to change what is being submitted — the storyboard text, the sheets, the props — not to resubmit the same content.

Open the segment cards that failed: each shows the block reason from the provider; those are the segments to edit before the pause ends.

## What triggers the pause

Only confirmed content blocks from the video provider count — the failures whose card says the content was refused. Timeouts, busy providers, invalid inputs and your own cancellations do not count. The window runs for 24 hours from your first block and then starts over; the pause starts at the moment the count is exceeded. What a single block means: Blocked by Moderation.

## During the pause

SceneMixer refuses a video generation submission with the message and a remaining time; nothing is charged. Generations already running finish normally. Sheets, storyboards, edits, composites of existing clips, ads and remakes that are not at their video stage all work. Use the hour to edit: the cards name what was refused.

## After the pause

Submissions work again, but the 24-hour count is still there. If the next submission is blocked too, a new pause starts at once. The count starts over only when the 24 hours from the first block have passed. So the second pause is the signal to stop and change approach, not to try variations.

## What was charged

SceneMixer refunds a generation the provider blocked automatically while you are within the daily number, because the provider confirmed it produced nothing. From the block that goes over the number, the credits are held for review instead; support decides them. Submissions refused during the pause never hold anything. Refund Rules.

## Repeated pauses and the account

The brake is automatic and does not itself affect your account standing. Content that breaks the rules does: repeated attempts to generate prohibited material are reviewed by people and can lead to restrictions or a closed account. What is prohibited and what happens: Content Rules and Moderation. If you believe the blocks are wrong — a lookalike name, an innocent prop — see IP Suspected and write to support with the segment.

| Event | Effect |
| Provider blocks a video generation | Counts once; refunded within the daily number, held for review past it |
| Count exceeds the daily number | Video submissions paused for the stated hours |
| Pause ends | Submissions allowed; count unchanged |
| Next block after the pause | New pause starts immediately |
| 24 hours after the first block | Count starts over |

## FAQ

### Does the pause stop my running generations?

No. Only new video submissions on SceneMixer are refused.

### Was I charged for blocked generations?

SceneMixer refunds blocks within the daily number automatically; from the one that went over, the credits are held for review.

### Can support lift the pause?

SceneMixer's pause is short and automatic; support looks at wrong blocks, not at shortening the wait.

### Do sheet or storyboard blocks count?

No. SceneMixer counts only video blocks toward this brake; text and image checks have their own messages.

### Why not reset the count when the pause ends?

So that resubmitting the same content cannot loop through SceneMixer's brake once an hour.

Updated November 3, 2026


---

# Video Generator AI: What a Second Costs, Tier by Tier

URL: https://scenemixer.com/guides/video-generator-ai/
The price list behind an AI video generator: every tier's cost per second, longest segment and resolution, what never changes the price, and what makes a bill grow.

# Video Generator AI: What One Second Costs, and What Moves It

Quick answer: A video generator AI does not charge for a video; it charges for the seconds it generates, tier by tier. On SceneMixer the unit is one second on one tier: 10 credits a second on the default tier, up to several times that on the top one, and the shortest thing you can generate is 4 seconds. This page gives the whole price list in one table, then the settings that do not change the price (frame, voices, language, type, segment length, plan), and the one thing that does make a bill grow: generating seconds you do not keep.

## The unit is a second, not a video

Every other step of making a video on SceneMixer is priced by something small: a read by the thousand characters, a setup image by the picture, a storyboard by the hundred characters. Video is the only step priced by duration, because the model works on every frame of every second. Video is also where nearly everything goes: in a typical episode the video lines are more than nine-tenths of the bill. Why video costs more than everything else.

Two numbers frame a purchase. The shortest segment is 4 seconds, so the smallest video you can pay for is four times the per-second price of your tier. The longest is set by the tier: 15 seconds on some, 30 on others. A segment longer than its tier's cap is refused when you press generate, before anything is charged. The price is shown on the segment's button before you press it, and the credits are held then, charged when the clip is delivered and returned if it fails.

## The price list: every tier, one second at a time

These are the figures SceneMixer charges, read live from its price list. Money is shown at the lowest credit price (the largest packs or an annual plan); small packs and monthly plans run a little higher.

| Tier | Credits a second | About, per second | Longest segment (s) | Resolution |
| Wan 3.0 · 480P | 10 | $0.06 | 30 | 480P |
| Wan 3.0 · 720P | 19 | $0.12 | 30 | 720P |
| Wan 3.0 · 1080P | 38 | $0.24 | 30 | 1080P |
| Seedance 2.5 · 480P | 19 | $0.12 | 30 | 480P |
| Seedance 2.5 Lite | 38 | $0.24 | 30 | 720P |
| Seedance 2.5 Pro | 75 | $0.47 | 30 | 1080P |
| Seedance 2.0 · 480P | 13 | $0.08 | 15 | 480P |
| Seedance 2.0 Fast | 25 | $0.16 | 15 | 720P |
| MiniMax H3 Fast | 16 | $0.10 | 15 | 768P |
| MiniMax H3 | 25 | $0.16 | 15 | 2K |

Read the table as three rules. The same family climbs with resolution: within one model family each resolution step up costs roughly double. The longest segment is a property of the model, not a price line. And the default is deliberately the cheapest tier, so the first thing you find out is whether a shot reads. The full list sits on the pricing page, and which model to pick by purpose says what each family is good for.

## What does not change the price of a second

People often expect a surcharge somewhere on SceneMixer that is not there. Each of these is a choice you can make freely, and none changes what a second costs.

| Setting | Changes the price? | What it does change |
| Frame: 9:16 or 16:9 | No | How shots are framed; video is priced by the second, not by the shape of the frame |
| Voices, dialogue and sound | No | They come with the seconds on every tier used here |
| Dialogue language | No | Who speaks what; a segment costs the same in any language |
| Type: Short Drama, TV Series, Film | No | Where episodes break, and the cap on episode length |
| Segment length: 15 or 30 seconds | Not per second | The size of each retake: a long take costs twice a standard segment to redo |
| Your plan | No | How many generations run at once and how many credits arrive each month; a generation costs the same credits for everyone |
| Music | No | It is mixed in at no extra charge |

Two of these are worth a second look. Segment length leaves the per-second price alone but changes how much one mistake costs. And a plan buys parallelism and a monthly allowance, not a lower rate and not a faster clip. What a paid plan changes.

## What does make a bill grow: seconds you generate and do not keep

The seconds in your finished episode are a small part of the seconds you pay SceneMixer for. A segment regenerated three times is three segments, and each delivered take is charged, however good or bad it is. A failed take is not: tasks that fail before reaching the model are refunded automatically, and ambiguous failures are held for review rather than silently charged. Refunds: what comes back automatically.

Fixing a clip you already have is priced differently from making a new take, and is worth comparing before you press:

| Fix | Price | Use it when |
| Trim | 5 credits each, whatever the length | The clip is right apart from a bad start or end |
| Lip sync (Redub) | 1 credit a second | Only the mouth is wrong |
| AI visuals | 2 the tier's per-second price, per second of clip | One thing in the picture is wrong and you like the rest |
| A new take | The tier's per-second price, per second | Framing, positions or action are wrong |

Joining the finished segments into one episode adds 1 credit a second of the finished length. A trimmed segment costs less to join. What each clip fix costs.

## One episode as a bill

Put the pieces together for a 60-second episode on SceneMixer’s default tier: 10 credits a second for 60 seconds, plus 1 credit a second for 60 seconds of joining. That is 660 credits, about $4.11 at the lowest credit price, before any retake. Reading the script, the first episode's storyboard and the first 15 setup images are small or free beside it. The way to spend less is not a cheaper button but fewer regenerated seconds: fix the text and the sheets first, draft on the cheapest tier, and finish once. Draft low, finish high and the cost breakdown with a worked series go further.

## FAQ

### How much does AI video cost per second?

On SceneMixer’s default tier it is 10 credits a second, about $0.06 at the lowest credit price. The top tier costs several times that. Every tier's price is in the table above and on the segment's button before you generate.

### What is the shortest video I can generate?

A SceneMixer segment can be as short as 4 seconds, so the smallest video is four times the per-second price of the tier you choose.

### Is there a flat price per video?

No. SceneMixer prices video per second, so a longer or retaken episode costs more. The price on each card adds up in front of you.

### Does vertical video cost more than landscape?

No. The frame does not change SceneMixer’s per-second price.

### Is dialogue or music an extra?

No. Voices and sound come with the seconds on every tier SceneMixer uses, and the score is mixed in at no charge.

### Does a membership make each second cheaper?

No. A SceneMixer generation costs the same credits on every plan. A plan adds monthly credits and lets more of your jobs run at once.

### Do I pay for a failed generation?

Not when it fails before reaching the model: those are refunded automatically. If a result cannot be confirmed either way, the credits are held for review instead of charged.

### Why does a retake cost the same as the first take?

Because it is another generation of the same seconds on the same tier. That is why fixing the text or the sheet first is cheaper than trying again.

Updated November 2, 2026


---

# The Video Says a Line You Never Wrote: Why, and How to Fix It

URL: https://scenemixer.com/guides/video-says-a-line-you-never-wrote/
An extra sentence at the start, a voice in a No dialogue shot, a thought read aloud: why AI video adds speech, how to write shots that prevent it, and the fixes.

# The Video Says a Line You Never Wrote, or Someone Talks in a Silent Shot

Quick answer: The video is generated with its sound in one go, so the model can add speech the storyboard does not contain. It happens most when a shot with no line still describes someone talking (“he speaks to the nobles”) — the model invents the words — when the picture description of a dialogue shot contains a whole sentence of what a character thinks, which can be read out, and when a segment opens with a long silence before the first line. Write silent shots as closed-mouth action, keep thoughts out of the description and show them instead, and regenerate. For a clip you otherwise like, cut the extra words with Trim.

Open the segment, press Edit, and read every shot marked No dialogue: if one of them says someone talks, explains, asks or tells, that is where the invented line comes from.

## What it looks like

- A sentence nobody wrote, often at the very start of a segment, sometimes in words that do not quite make sense.
- A shot marked No dialogue in which a character speaks, or a narrator talks over it.
- One more sentence after the written line has finished.
- A thought from the description spoken aloud: “She is going to bluff her way out of this.”
- A written line said twice.
A wrong word inside a written line, or the right line in the wrong voice, is a different problem: Text in the picture and wrong dialogue.

## Why it happens

Dialogue on SceneMixer is generated with the picture: the model produces the image and the sound together from the segment’s text. That is what makes lips move with the words, and it is also why the model can produce words of its own when the text invites them. Three invitations account for most of it:

- A silent shot that describes talking. “He addresses the crowd”, “they argue in low voices”, “she explains the plan” — with no line attached. The model sees a person who should be speaking and supplies the words.
- A thought written as a sentence in a dialogue shot. In a shot where someone speaks, a full sentence of what they think or intend (“She is weighing whether to tell him the truth.”) can be blended into the spoken line.
- A long silence before the first line. When a segment opens with several silent seconds and the dialogue comes later, the model sometimes fills the gap with improvised speech.
Storyboards written by the site avoid the first two, so they show up mostly in text you edit or paste yourself. And a small share of clips adds a word or two even from clean text — a regeneration usually clears that.

## Write silent shots so nobody speaks

In a shot with no line, describe what people do, never that they talk. If someone in the shot is a listener, say their lips are closed.

| Invites a made-up line | Write instead |
| [Shot 2] 4s: Medium shot. @Aldric speaks to the nobles gathered around him. No dialogue. | [Shot 2] 4s: Medium shot. @Aldric raises a hand and the nobles around him fall silent; he looks from face to face, lips closed, no lip movement. No dialogue. Ambient: a murmur dying away. |
| [Shot 5] 3s: Close-up. In frame only: @Mara. She tells him she is leaving. No dialogue. | [Shot 5] 3s: Close-up. In frame only: @Mara. She sets the key on the table and pushes it toward him, lips closed, no lip movement. No dialogue. |
| The two of them argue in the hallway. | Give them the lines, or write the argument as action: he blocks the door, she steps around him. |
If you want speech there, write the line. On SceneMixer, every spoken word belongs in a line with a speaker and a voice chip. How to write one shot.

## Keep thoughts out of dialogue shots

In a shot where someone speaks, everything before the line should be what the camera sees. A sentence that states a thought, a plan or a judgement can come out of a mouth. Show it instead:

| Can be read aloud | Write instead |
| [Shot 3] 5s: Medium shot. @Mara faces the guard. She is going to bluff her way through this. @Mara (bright, too casual): I’m on the list. (voice: 🔊Mara) | [Shot 3] 5s: Medium shot. @Mara faces the guard, chin up, a smile that does not reach her eyes. @Mara (bright, too casual): I’m on the list. (voice: 🔊Mara) |
| @Jules looks at the letter. He realises she lied to him about everything. | @Jules looks at the letter; his jaw tightens and he folds it once, slowly. |
Delivery notes in brackets say how a line is said — (low, steady), (barely a whisper) — not what the character means by it. Describe the cause, not the result.

## Watch the start of a segment

Most invented speech lands in the first seconds. When a segment begins with a long silent shot and the first line comes several seconds later, either give the opening shot a clear action that fills its time (a door opening, a car pulling up), keep it short, or let the first line come earlier. Shot seconds and timing.

## A narrator you did not ask for

If no character is meant to narrate, check that no shot says “a voice explains” or “we hear her thoughts”, and that every shot without a line is marked No dialogue. If you do want narration, make it a character in SceneMixer: a narrator set to Voice only (off-screen), with lines written as @Narrator (voice-over): … (voice: 🔊Narrator). Then it is a voice you chose, not one the model made up. Lines, voice-over and captions.

## Fixing a clip that already has it

| Where the extra speech is | Fix | Cost |
| At the very start or end | Edit this clip → Trim: cut it off | 5 credits per trim |
| In the middle, in a pause between lines | Trim with Remove selected: cut that moment out | The same |
| Mixed into a line you want to keep | Fix the shot text, then regenerate the segment | The segment’s video price |
| Everywhere, or at random | Regenerate; one clip in a few can add a word even from clean text | The segment’s video price |
Editing the storyboard text on SceneMixer is free; only the regeneration is charged, and the earlier take stays in Version History. If you ask AI edit to remove the extra words, it may answer The storyboard doesn’t say that. It appeared while the video was being generated, and regenerating this segment usually clears it. — which is the point: the words were never in the text. Trim a clip · Remove a moment from the middle.

## A quick check before you generate

- Every shot without a line ends with No dialogue.
- No silent shot says anyone talks, explains, asks, tells, argues or calls out.
- Listeners in dialogue shots are written with closed lips.
- No sentence before a line states what someone thinks or plans.
- The first line does not come after a long silent opening.
- Any narration is a voice-only character with real lines.
Checklist before you generate videos.

| Symptom | Most likely cause | Change |
| Extra sentence at the start | Long silence before the first line | Shorter or busier opening shot, earlier first line |
| Someone speaks in a No dialogue shot | The shot says they talk | Rewrite as closed-mouth action |
| A thought is spoken | A thought written as a sentence in a dialogue shot | Show it as a look or a gesture |
| Narrator out of nowhere | A shot says a voice explains | Remove it, or make a real voice-only narrator |
| One stray word now and then | Occasional glitch | Trim, or regenerate |

## FAQ

### Why does the video say a line I never wrote?

Picture and sound are generated together on SceneMixer, so the model can add speech when the text invites it: a silent shot that says someone talks, a thought written as a sentence in a dialogue shot, or a long silence before the first line. Fix the text and regenerate.

### The shot says No dialogue, but someone still talks. Why?

Usually the same shot also describes them speaking, explaining or arguing. Rewrite that part as action with closed lips, keep No dialogue, and regenerate.

### Can I remove the extra words without regenerating?

If they sit at the start, the end or in a pause, yes: use SceneMixer's Trim under Edit this clip to cut them off or cut that moment out. If they are mixed into a line you want to keep, regenerate.

### Can AI edit remove it?

SceneMixer's AI edit changes the storyboard text, and the extra words are not in the text, so it will tell you to regenerate. Fix the shot that invited them, then regenerate.

### Does it cost anything to fix?

Editing the storyboard on SceneMixer is free. Regenerating the segment is charged at its video price, and a trim has its own small price on the button.

### Will a higher video tier stop it?

Not reliably. The cause is in the text; fix the text first. A regeneration on any tier can clear an occasional stray word.

### How do I add a narrator on purpose?

Create a narrator character in SceneMixer set to Voice only (off-screen) and write the narration as lines with a voice-over note and the narrator's voice chip.

### Is this a bug that will be fixed?

It is a known behaviour of generating speech and picture together. SceneMixer already cleans the most common causes out of its own storyboards; text you write or paste yourself is worth checking with the list on this page.

Updated November 3, 2026


---

# Video Won't Play or Download: What to Check | SceneMixer

URL: https://scenemixer.com/guides/video-wont-play-or-download/
A clip that will not play in the editor, a download that does not start, a shared link that shows a poster instead of a player: the usual causes, in order.

# The Video Won’t Play, or Won’t Download

Quick answer: First check the segment’s status: only a Done segment has a clip. Then refresh the page — the player streams a light copy and a stale page can hold an old address. If it still will not play, press Download: the full-quality file arrives even when the in-page player is unhappy. A shared link that shows a still picture instead of a player has reached its daily viewing limit and plays again the next day.

Open a finished episode and play a segment. If that plays and yours does not, the difference is in the status of your segment.

## Nothing to play yet

Select the segment and read its card. Pending, Generating or Failed means there is no clip to play. Done means there is. For an episode composite, the record must say Done too. Statuses.

## The player will not start

Refresh the page. The player streams a lighter copy of the clip made for smooth playback, and a page that has been open a long time can be holding an address that has since changed. If the pane shows the clip’s thumbnail but pressing play does nothing, the same refresh fixes it. On a phone, check that the browser is allowed to play video and that you are not on a data-saver mode that blocks it.

If the player still fails, press Download. The full-quality file does not depend on the in-page player. Original and play versions.

## The download does not start

Browsers block downloads that open in a new window when pop-ups are blocked; allow them for this site and press again. On a phone, the file goes wherever your browser puts downloads — Files on iPhone, the Downloads folder on Android — not straight into Photos. A composite of a long episode is a large file; give it time on a slow connection.

## You downloaded the wrong version

On SceneMixer, Download gives you the version that is active in the episode. To download an earlier take, open Version History on the segment, choose it with Use, then download. Versions.

## A shared link shows a picture, not a player

Shared pages have a daily viewing allowance per clip, and the site as a whole has a daily bandwidth allowance. When either is reached, the page still opens but shows the poster frame instead of the player, and plays again the next day. Your own copy in the editor is unaffected. Sharing.

## The file plays but looks soft

The in-page player is the light copy. Download for the full-quality file. If the download itself looks soft, the segment was generated on a 480P tier; regenerate it on a higher tier. Tiers.

## Still stuck

Send feedback from SceneMixer's sidebar with the segment number, your browser, and a screenshot of what you see. How to get help.

| Symptom | Check | Then |
| Nothing in the preview pane | Segment status | Generate, or wait for Done |
| Thumbnail shown, play does nothing | Page age | Refresh |
| Player errors | Download instead | The file is independent of the player |
| Download does not start | Pop-up blocking; phone download folder | Allow, press again; look in Files / Downloads |
| Wrong version downloaded | Active version | Choose the version, then download |
| Shared link shows a still | Daily viewing allowance | It plays again the next day |
| Looks soft | Play copy vs download; tier | Download; regenerate on a higher tier |

## FAQ

### The player shows a thumbnail but will not play.

Refresh the page. If it still fails, press Download; the file does not depend on SceneMixer's player.

### Where did my download go on my phone?

Wherever your browser saves downloads: Files on iPhone, Downloads on Android. SceneMixer does not add it to Photos automatically.

### Why does my shared link show a picture instead of the video?

Right after sharing, the shared copy may still be being prepared; the page says so and refreshes itself. Otherwise the clip reached its daily viewing allowance on SceneMixer and plays again the next day; your own copy is unaffected.

### Is the download the same quality as the player?

On SceneMixer, the download is the full-quality file; the player streams a lighter copy.

### Can I download an older version of a segment?

Yes. Choose it in SceneMixer's Version History with Use, then Download.

### The downloaded file will not open on my computer.

Open it with another player, such as VLC or your browser. If it still fails, check the file finished downloading (its size should be several megabytes), download it again, and if that fails too, send feedback with the segment number.

Updated November 3, 2026


---

# Visual Style Presets, One by One: Live Action, 3D, Anime | SceneMixer

URL: https://scenemixer.com/guides/visual-style-presets-one-by-one/
All 21 visual styles in Step 1: live action, 3D and 2D animation including claymation. What each looks like, what it suits, and why it is locked after Step 1.

# Visual Style Presets, One by One

Quick answer: Visual Style in Step 1 sets how everything is drawn — the character sheets, the scene images and every video. There are 21 presets in three groups: Live-action (Digital HD, Shot on Film, Soft Glow, Black & White Film), 3D Animation (Chinese 3D Animation, Realistic 3D, Cel-shaded 3D, 3D Toy Cartoon, Family 3D Animation, Game Engine Realism) and 2D Animation (Chinese Painted Style, Ink Wash Animation, Anime, Korean Webcomic, American Comic, Watercolor Anime, Painterly Anime, Painted Manhua, Watercolor Storybook, Classical Oil Painting, Claymation). Each card has a sample and a one-line description. Once you continue past Step 1 the style is fixed for the project.

A finished sample in Digital HD, and animated ones on the samples page. All samples. For the sets and wardrobe that go with it: scene and costume styles, one by one.

## Where it is, and why it is locked

Step 1, Set Up Your Project, the Visual Style row: a grid of cards, each with a sample image and a one-line description, grouped by medium. The card you select plays a short loop of that look. The notice beside the row is the important part: it can’t be changed once you continue, because it shapes every character sheet, scene and video. To switch styles, start over with a new script. Members can add their own style text with + Custom Style; how to write one that holds up.

The visual style also decides which version of the scene and costume cards you see: pick a 3D or 2D visual style and those two rows switch to 3D or 2D pictures of the same presets.

## Live-action

- Digital HD — crisp and true-to-life, like today’s streaming dramas. The default choice for drama meant to pass as filmed.
- Shot on Film — film grain, warm colour and soft blacks. A cinema grade on the same real people and rooms.
- Soft Glow — hazy highlights, low contrast, pale and dreamy. Good for period romance.
- Black & White Film — rich blacks, whites and grays, like a classic film. Good for flashbacks and noir.

## 3D Animation

- Chinese 3D Animation — detailed 3D characters, flowing hair, glossy jade and metal. The look of Chinese fantasy and xianxia series.
- Realistic 3D — near-human 3D characters and skin, like film CG.
- Cel-shaded 3D — anime characters built in 3D, flat colour bands.
- 3D Toy Cartoon — rounded, matte toy-like figures in bright colours.
- Family 3D Animation — big-eyed, rounded characters with glowing skin, like an animated feature. Suits family stories.
- Game Engine Realism — next-generation game graphics: sharp materials, no grain.

## 2D Animation

- Chinese Painted Style — fine brush lines, muted ink tones, silk texture.
- Ink Wash Animation — shapes painted in ink washes on rice paper, sparse.
- Anime — clean lines and flat colours, like TV anime.
- Korean Webcomic — slender, polished characters with soft airbrushed shading. Suits modern romance.
- American Comic — bold ink lines and heavy black shadows. Suits action and crime.
- Watercolor Anime — a watercolour-on-paper look with fine lines and natural colours.
- Painterly Anime — lavish, saturated scenery with simple-lined characters.
- Painted Manhua — no line art; volume built from brushwork, rich contrast.
- Watercolor Storybook — transparent watercolour and pencil, like a children’s book.
- Classical Oil Painting — visible brushwork over warm brown, like a museum painting.
- Claymation — hand-pressed clay figures and miniature sets, fingerprints left in.

## How to pick

Match the medium to the audience first — live action for drama feeds, the anime and comic styles for anime audiences, Family 3D Animation for family stories — then pick within the group by the sample cards. If you are unsure, Digital HD is the safe default. The scene style and costume style rows on the same SceneMixer screen work with whichever visual style you choose. How the style settings work together.

One caution for the IP check: a style named after a studio or a franchise is not offered here for a reason. Stay with the presets. What gets blocked.

## Renamed on October 17, 2026

All 21 presets got plain names on October 17, 2026, and the former stop-motion group was folded into 2D Animation. If you know a style by its old name, or an older project or article uses it, this is the list.

| Old name | Now |
| Realistic HD | Digital HD |
| Filmic Live-Action | Shot on Film |
| Soft-Mist Live Action | Soft Glow |
| 2D Cel Shading | Anime |
| Hand-painted Anime Film | Watercolor Anime |
| 2D Chinese Style | Chinese Painted Style |
| Manhua Full-render | Painted Manhua |
| American Graphic Novel | American Comic |
| Oil Painting | Classical Oil Painting |
| Realistic 3D CG | Realistic 3D |
| Feature 3D Animation | Family 3D Animation |
| 3D CG Xianxia Fantasy | Chinese 3D Animation |

The other nine presets kept their names.

| Group | Presets | Pick for |
| Live-action | Digital HD · Shot on Film · Soft Glow · Black & White Film | Drama meant to pass as filmed |
| 3D Animation | Chinese 3D Animation · Realistic 3D · Cel-shaded 3D · 3D Toy Cartoon · Family 3D Animation · Game Engine Realism | Family films, fantasy, games |
| 2D Animation | Chinese Painted Style · Ink Wash Animation · Anime · Korean Webcomic · American Comic · Watercolor Anime · Painterly Anime · Painted Manhua · Watercolor Storybook · Classical Oil Painting · Claymation | Anime, comic and handmade looks |

## FAQ

### How many visual styles are there?

SceneMixer offers 21, in three groups: live-action (4), 3D animation (6) and 2D animation (11, including Claymation).

### Can I change the visual style after Step 1?

No. SceneMixer's visual style shapes every image and video in the project. Start a new project with a new script to switch.

### Which style is the default?

Digital HD is the safe choice for most drama. SceneMixer's picker pre-fills an AI suggestion from your script; change it if you like.

### Where did Realistic HD, Filmic Live-Action and the other old names go?

They were renamed on October 17, 2026; see the table above. SceneMixer now calls Realistic HD Digital HD.

### Can I write my own style?

SceneMixer members can, with + Custom Style. Everyone else picks from the presets.

### Do the scene style and costume style depend on the visual style?

No. They are separate rows on SceneMixer and combine with any visual style. Their cards change to 3D or 2D pictures to match the group you pick.

### Why is there no preset named after a famous studio?

Names of studios and franchises trigger SceneMixer's IP checks on the video side. The presets describe the look without the name.

Updated November 3, 2026


---

# Wan 3.0 Price per Second: Three Tiers, What Each Is For | SceneMixer

URL: https://scenemixer.com/guides/wan-3/
The Wan 3.0 family on this site: what each of its three tiers costs per second, how long one generation runs, and the reference slots every tier takes.

# Wan 3.0 on SceneMixer: three tiers, billed by the second

Quick answer: On this site Wan 3.0 is three tiers, one per resolution. Wan 3.0 · 480P is the entry step and the one a new project starts on: 10 credits per second (about $0.06 at the lowest credit price). Wan 3.0 · 720P carries the Recommended badge at 19 (about $0.12), and Wan 3.0 · 1080P is the top step at 38 (about $0.24). A single generation runs up to 30 seconds on any of the three, with dialogue and ambient sound and the same reference slots — so the step you pick sets the picture, not the length.

## One setting, three steps

In SceneMixer, a step is picked in one place: the tier button in the episode editor’s top bar, or Settings → Model Preferences → Video Model. The setting governs what you render next — already-made segments stay as they are, and each version remembers the step it was rendered on. The list marks Wan 3.0 · 720P as Recommended — advice for finished segments, not a switch: you stay where you are until you pick another step.

Each step has its own page for the detail behind it: the entry step, the recommended middle, and the step above.

## The price steps, side by side

SceneMixer puts three steps on one price scale: every charge is per second of rendered picture.

| Step | Credits, per second | In money, at the lowest credit price | Longest generation | Resolution |
| Wan 3.0 · 480P | 10 | $0.06 | 30 seconds | 480P |
| Wan 3.0 · 720P | 19 | $0.12 | 30 seconds | 720P |
| Wan 3.0 · 1080P | 38 | $0.24 | 30 seconds | 1080P |

The ladder below only moves the picture. The take ceiling does not move with it, and neither do the reference slots: up to ten images and five voice clips go into every generation, whichever step renders it.

## What you can make with it

The episode above is Crimson Vow: this site carried it from the first draft of the script to the final composite on Wan 3.0 · 720P, and swapping in either of the other two steps re-renders the same storyboard — the tier is only a setting.

What the family suits: the take that must not be cut — a walk-and-talk, a slow push-in, an argument that plays out in one frame; a crowded shot that wants many sheets attached, since up to ten images and five voice clips go in per generation; and a whole episode blown through at the entry rate before the keepers are chosen.

A project is not one film but a whole series if you want it: episodes stack in order, each one a chain of its own segments, and the composite tools work episode by episode.

## Draft on one step, finish on another

The pattern this family is built for: keep the drafts on Wan 3.0 · 480P while the storyboard is still moving, then regenerate only the segments you are keeping on Wan 3.0 · 720P or Wan 3.0 · 1080P. Both versions stay listed on the segment, SceneMixer's composite uses whichever one is active, and switching steps re-renders nothing — draft cheap, finish where it counts walks through it.

And a second step is not needed at all if the draft is already good: The Veil at Dawn went from pasted script to finished composite entirely on Wan 3.0 · 480P — every segment of it — which is the other way to settle the draft-then-finish question.

Voices do not move with the step: dialogue and ambient sound come out of every generation on every tier, fed by the same voice clips. The one pronunciation quirk documented for this family — the odd German or Ukrainian line sliding off — is fixed by shortening the line and regenerating, not by moving to another step.

## How to start

A new project needs no detour through the model settings first: it renders on Wan 3.0 · 480P from the very first segment, at 10 credits a second, with the two higher steps one click away whenever a keeper wants them. Two costs do sit ahead of that first render: the script read is a paid analysis, priced by the script’s length, and the character, location and prop sheets draw on 15 free images per account before further sheets are billed. SceneMixer's storyboard for episode 1 of a first drama comes without a charge; video itself is billed by the delivered second.

Planning a scene longer than one take, or weighing this family against the rest? How a long scene is cut into segments covers the rhythm, and the purpose-by-purpose ranking lines every step up side by side.

## FAQ

### What does Wan 3.0 cost per second?

On SceneMixer: 10 credits per second on Wan 3.0 · 480P, 19 on Wan 3.0 · 720P, and 38 on Wan 3.0 · 1080P — about $0.06, $0.12 and $0.24 at the lowest credit price. Billing happens on delivery; a render that fails comes back, or waits in a hold while support looks at it.

### How long can one Wan 3.0 generation run?

A single SceneMixer generation runs to 30 seconds at most, on any of the three steps. A scene that needs more time on screen is simply more segments, and the storyboard can be planned that way from the start.

### Which of the three steps should I draft on?

The entry step, for anything whose blocking is not settled yet. Move the keepers to the recommended middle, and go to the step above only for what will be watched large. Regenerating on SceneMixer is a new charge at any step, and both versions stay on the segment.

### Wan 3.0 or MiniMax H3?

Leave the Wan steps for MiniMax H3 when the picture, not the take, is the point: MiniMax H3 renders at 2K for 25 credits a second and MiniMax H3 Fast at 768P for 16, and the family cuts several shots inside one generation, so an exchange and the reaction to it can arrive as a single render. Its segments close at 15 seconds — enough for a landing beat, a held glance, a line. What holds a segment on Wan is time or a crowded frame: up to 30 seconds in one take at every step, and ten images plus five voice clips where the H3 family takes nine and three. The rates on SceneMixer to weigh against that picture: 10 credits a second to start, 38 at the top. The side-by-side guide sorts the two by the scene in front of you.

### Do all three tiers take the same references?

Yes: up to ten images and five voice clips per generation, the same voice clips at every step, and the same content checks and refunds on SceneMixer. Only the picture and the price move with the step.

Updated November 3, 2026


---

# Wan 3.0 · 1080P Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/wan-3-1080p-tier/
What the highest Wan 3.0 step costs per second on SceneMixer, how long one take runs and when it is worth the rate.

# Wan 3.0 · 1080P: price per second, limits and when to use it

Quick answer: : Wan 3.0 · 1080P costs 38 credits per second (about $0.24): takes up to 30 seconds at 1080P, the step for the version you publish or put on a large screen.

The still above is from The Glass Floor, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

Original vs play version explains which file you download after a 1080P render.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

## What 1080P buys

Room. A wide shot keeps faces in the background readable; a slow push-in does not fall apart as it gets close; embroidery, rain, dust and skin texture survive. It matters most on a television, a projector, a desktop player, and in any edit where you will crop or reframe later. On a phone held at arm’s length the difference from 720P is small, which is the honest reason this tier is not the default.

Everything else is the family standard on SceneMixer: up to 30 seconds per generation, 16:9 or 9:16, dialogue and ambient sound in the same render, every version kept.

## References it accepts

Same slots as the rest of the family:

## What it costs

38 credits per second, against 10 at 480P and 19 at 720P. A 30-second take is 30 × 38. The arithmetic argues for discipline: get the segment right cheaply, then pay this rate once. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## Downloads and playback at this tier

A finished 1080P segment has two files. The original is what you download on SceneMixer and what the episode composite is built from. The play version is a lighter copy the in-page player and share links use so a phone on a slow connection does not stall; it is the same resolution, compressed harder. If a 1080P clip looks softer in the browser than you expect, download the original before judging it.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When 1080P is not worth it

While the storyboard is still moving; for a vertical clip that will only ever be watched on phones; for a segment whose problem is blocking or dialogue rather than detail. In those cases 480P shows you the same thing for a fraction of the credits on SceneMixer, and 720P is the sharing tier. Come here for the final version of a segment that already works.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Is 1080P the highest resolution on the site?

Among SceneMixer's Wan tiers, yes. MiniMax H3 offers a 2K render with its own reference rules and 15-second segments.

### Does 1080P take longer to generate?

Usually somewhat longer than 480P for the same seconds on SceneMixer.

### Can I download the 1080P file without the AI badge?

On a paid SceneMixer plan, new renders come without the badge by default, and you can choose to keep it. The download is the full-quality file.

### Do I have to regenerate the whole episode at 1080P?

No. Regenerate only the segments you keep; SceneMixer's composite uses each segment’s active version whatever its tier.

### Is a 30-second take at 1080P the most a single generation can be?

Yes, on SceneMixer's Wan family: up to 30 seconds per segment.

Updated November 3, 2026


---

# Wan 3.0 · 480P Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/wan-3-480p-tier/
What the default video tier on SceneMixer costs per second, how long one take runs, which references it reads and when to step up.

# Wan 3.0 · 480P: price per second, limits and when to use it

Quick answer: : Wan 3.0 · 480P costs 10 credits per second (about $0.06). It is the default video tier, where every new project starts: takes up to 30 seconds at 480P with dialogue and ambient sound in the render.

The still above is from The Eagle's Vow, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

480P, 720P or 1080P compares the three Wan tiers side by side.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

If you have never changed the video tier in SceneMixer, this is the one you are on.

## What a segment on this tier can be

Length: up to 30 seconds in one generation. That is the long-take mode; with the 15-second segment length a segment is at most 15 seconds. Resolution: : 480P, in the aspect ratio of the project (16:9 or 9:16). Sound: spoken lines and ambient sound come out of the same generation as the picture; each character speaks in their own voice — the preset or uploaded voice if you set one, otherwise the written voice description. Cuts: a segment can contain several shots; the storyboard decides where they fall.

## References it accepts

SceneMixer reads the character sheets, location images and prop images for the assets named in the segment, plus voices and any state images you assigned. The limits are per generation:

## What it costs

10 credits per second of finished video. A 15-second segment is 15 × 10; a 30-second long take is twice that. For comparison, the same seconds at 720P cost 19 per second and at 1080P 38 per second. Regenerating a segment is a new charge each time; every version is kept, so nothing is lost by trying. Credits are only taken when the video is delivered.

## What it is good at, and what to watch

Good at: drafting. Blocking a scene, checking that the right people are in frame, hearing how the lines land, finding out whether a 30-second take holds together. Most of what makes a shot right or wrong — who is in it, what they do, where the camera is — is visible at 480P.

Watch: fine detail (a pattern on a prop, embroidery, distant faces) is soft at 480P; that is what 720P and 1080P are for, not a reason to rewrite the storyboard. On SceneMixer, dialogue in German and Ukrainian occasionally mispronounces a word on this tier; shorter sentences help, and a regeneration is cheap here.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you are most likely to meet on this tier, what they mean and what to do:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When to switch tier

Up to 720P when a segment is right and you want it crisp for sharing. Up to 1080P for the version you will publish or cut into a longer film. Sideways to MiniMax H3 for a 2K picture, with its own reference rules; note it renders 15-second segments. Switching on SceneMixer changes nothing in the storyboard: pick the tier, regenerate the segment, keep the version you like.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Is 480P really the default for every new project?

Yes. Until you pick something else in SceneMixer's editor or in Settings, every segment renders on this tier.

### Can I upgrade a finished 480P segment to 1080P without regenerating?

No. SceneMixer decides resolution at generation. Pick the 1080P tier and regenerate the segment; the storyboard, references and voices stay exactly as they were.

### Does a 30-second segment cost more per second?

No. SceneMixer's price is per second at the same rate; a 30-second take simply has twice the seconds of a 15-second one.

### Why did my segment fail with “Invalid input”?

Usually one of the reference files for the segment could not be read. Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support.

### Is the sound generated too, or added later?

SceneMixer generates it with the picture. Lines are spoken in each character’s voice; ambient sound matches the scene.

Updated November 3, 2026


---

# Wan 3.0 · 720P Price, Limits and When to Use It | SceneMixer

URL: https://scenemixer.com/guides/wan-3-720p-tier/
What the Recommended Wan 3.0 step costs per second on SceneMixer, why it carries the badge, how long one take runs and when to pick it.

# Wan 3.0 · 720P: price per second, limits and when to use it

Quick answer: : Wan 3.0 · 720P costs 19 credits per second (about $0.12) and carries the Recommended badge: takes up to 30 seconds at 720P, the step most people deliver clips on.

The still above is from Rocket Pup, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

The default 480P tier is where most people draft; this page is about the step up.

## Where you pick it

Two places, one setting. In the episode editor: the button in the top bar that shows the current tier’s name, next to Composite. In Settings: Model Preferences → Video Model. Both change the same thing — the tier your next segments render on, in this project and in new ones. Segments already made keep the tier they were made with, and each version records it. The price on a segment’s generate button always shows the tier you are about to use.

In the Settings list on SceneMixer it carries the Recommended badge. The badge is advice, not a default: you stay on 480P until you pick another tier.

## What changes against 480P

Nothing in the storyboard, the references or the voices: a segment regenerated on this tier uses the same inputs. What changes is on screen: faces in a two-shot read cleanly, fabric and hair keep their texture instead of turning to soft blocks. On a phone at full screen, 720P is usually the point where a viewer stops noticing the resolution and starts watching the story.

The rules are the family’s: up to 30 seconds in one generation, 16:9 or 9:16 as the project sets, dialogue and ambient sound generated with the picture.

## References it accepts

Identical to the other Wan tiers. Per generation:

## What it costs

19 credits per second on SceneMixer. Against the other two Wan tiers: 480P is 10 per second, 1080P is 38. A common way to spend: draft every segment at 480P, then regenerate only the segments you keep at 720P — the versions panel keeps both, and the composite uses whichever version is active. Credits are taken on delivery; if a generation fails, the credits come back or are held until support checks it.

## When 720P is the right stop

Sharing a segment link, posting a vertical clip, sending a cut to a client for notes, or building a composite you will watch on a phone or tablet. It is also the sensible tier for a long take you already know works at 480P: same take, cleaner picture, without paying the 1080P rate for seconds that will be watched small.

Dialogue on SceneMixer behaves the same as on 480P. German and Ukrainian lines occasionally mispronounce a word on the Wan tiers; the fix is shorter sentences and a regeneration, not a different resolution.

## Content checks

Text and references are checked before rendering. A rejection is not charged: the credits come back and the credit history says so. The count is kept per account: after 12 rejections within 24 hours, the next one is held for review instead of refunded, and video generation pauses for 1 hour. So edit the flagged material rather than retrying the same segment. Content rules and checks explains what gets flagged.

## Common errors

The messages you may meet on this tier:

| What you see | Usual cause | What to do |
| “Invalid input — check the reference images and duration, then retry.” | Usually one of the reference files for this segment could not be read | Generate again once; if it keeps failing, check that the character, location and prop images open, then contact support |
| “Blocked by content safety rules — edit the script or description and retry.” | The script or a reference tripped the content check | Edit the flagged part; see the content-check section for when a rejection is refunded |
| “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” | The service was saturated; the site already tried again once for you | Wait a few minutes and generate again; nothing was deducted |
| Video generation paused | More than 12 rejected generations within 24 hours; the next one is held for review | Wait out the pause (1 hour); fix the material before retrying |

## When to switch tier

Down to 480P while you are still changing the storyboard: every regeneration is a charge, and the cheaper tier shows you the same blocking. Up to 1080P for a version you will publish on a large screen or cut into a longer film. SceneMixer keeps the tier each segment was made with, so a mixed episode — 480P drafts, 720P keepers — is normal.

| Reference type | Limit on this tier | Note |
| Images | Up to 10 | Character sheets, location and prop images, states, your own uploads |
| Voice clips | Up to 5, about 14 s in total | Character voices; a clip that does not fit is left out and that character speaks from the written voice description |

## FAQ

### Does Recommended mean 720P is the default?

No. Recommended is a badge on the tier SceneMixer suggests for most finished segments. You stay on the default, 480P, until you pick another tier.

### Will a segment look different between 480P and 720P beyond sharpness?

The inputs on SceneMixer are identical, but every generation is a fresh render; expect the same scene and blocking with small differences in motion, not a pixel-for-pixel upscale.

### Can I mix 480P and 720P segments in one episode?

Yes. SceneMixer remembers each segment's tier. The composite stitches the active version of every segment whatever tier made it.

### Is 720P enough for a vertical short?

For phone playback, usually yes. Go to 1080P when the clip will be shown large or cut into a longer film.

### Do the segment length and the reference limits change?

No. The Wan tiers on SceneMixer share them: up to 30 seconds per segment and the same reference slots.

Updated November 3, 2026


---

# Wan 3.0 Prompt Guide: Shots That Survive a Re-render | SceneMixer

URL: https://scenemixer.com/guides/wan-3-prompt-guide/
How to write shots for Wan 3.0 on this site: one storyboard text for all three steps, what to judge on the entry step, and how to spend the reference slots.

# Wan 3.0 prompt guide: shots that survive a re-render

Quick answer: Write the shots for Wan 3.0 once and render them on any of its three steps: Wan 3.0 · 480P at 10 credits a second, Wan 3.0 · 720P at 19 and Wan 3.0 · 1080P at 38. The text does not change between steps, only the picture does, so judge staging, timing and delivery on the entry step and re-render only the shots you keep. One generation runs up to 30 seconds.

## Write once, render on any step

The episode above, The Veil at Dawn, went from a pasted script to a finished composite entirely on Wan 3.0 · 480P: every segment, no second pass. That is the plainest proof of the habit this page describes — the shots were written to be understood at the entry step, and nothing in them needed a different wording to look good on the next one.

The family is three steps that differ only in resolution and price. A segment can run up to 30 seconds on any of them and takes the same references, so there is no wording that works on one step and fails on another. Write the shot the way the shot grammar asks — marker and seconds, framing and camera, who is in the picture, one thing that happens, the line, the ambience — and the same text re-renders higher whenever you want it. How to write one shot has the grammar.

## What to judge on the entry step

Treat the entry step as a rehearsal. What it shows you reliably is the part that is decided in the text: whether the framing is the one you meant, whether the camera moves where the shot says it does, whether the seconds are enough for the line, whether two people end up where the storyboard put them. Those are the problems that wording can fix, and they are the cheapest to fix here.

What it does not settle is fine detail in the picture. If a face or a prop only matters at full size, judge it on the step above, once the staging is right. Doing it in this order is the whole of the draft-then-finish pattern: the text gets its corrections at the low rate, and the high rate is spent once.

## Fix the shot, not the tier

When a render misses, the temptation is to try another step. Try another sentence first. Change one thing in the shot — the framing word, the action, the seconds, the line — and regenerate that segment; every version is kept, so the failed one is still there to compare. Editing one shot and regenerating shows the loop. Changing two things at once makes the next result impossible to read, and moving to a more expensive step to fix a staging problem buys a sharper version of the same mistake.

## Spend the reference slots on the camera's people

On SceneMixer, a generation takes up to ten images and five voice clips, with about fourteen seconds of voice in total. That is generous, and it is the main reason an ensemble scene belongs on this family. Use the slots deliberately: attach the sheets for the people the camera will actually find, keep the speaking characters' voices, and describe the rest of a crowd in words. When a segment asks for more than the budget, the site leaves out the least important references first, which shows up as a voice that drifts rather than as an error.

## Lines in other languages

If the dialogue is German or Ukrainian, keep each line short: on the default step an occasional word may slide off, and a short line leaves less room for it. SceneMixer marks no step as better for those two languages, so the fix is shorter sentences and a regenerate, not a move up in price. Dialogue language and video tier covers the details.

## When this family is the wrong place to write

If a beat wants a very short, dense exchange with cuts inside the generation, or you want to see how another family reads the same shots, the other tiers take the same text. The Wan 3.0 family page has the prices; the comparison with MiniMax H3 is the one for choosing between a longer take and a denser one.

## FAQ

### Do I write prompts differently for each of the three Wan 3.0 steps?

No. The three steps on SceneMixer share one storyboard text, one segment ceiling of 30 seconds and the same reference slots. Only the picture and the price change.

### What should I check on the cheapest Wan 3.0 step?

Framing, camera movement, seconds against the line, and where people end up. Fine detail is better judged on a higher step once the staging is right.

### How many references does a Wan 3.0 generation take?

Up to ten images and five voice clips, with about fourteen seconds of voice in total. If a segment needs more, SceneMixer leaves out the least important references first.

### A shot rendered wrong. Should I move up a step?

Edit the shot first: change one thing and regenerate that segment. A higher step gives you a sharper version of the same render, not a different staging.

### Are German and Ukrainian lines handled differently?

They may slip a word on the default step, and SceneMixer marks no step as better for them. Shorter lines and a regenerate are the fix.

Updated November 3, 2026


---

# Wan 3.0 for AI Short Drama: Parameters & Limits | SceneMixer

URL: https://scenemixer.com/guides/wan-3-short-drama/
Wan 3.0 for short drama: wan3.0-video vs prime, 480P–1080P, 2–30 s clips, up to 10 reference images, moderation rules and how the Wan tier works.

Guide · Tools

# Wan 3.0 for AI Short Drama (2026): Parameters, Reference Limits, Moderation Rules and the 30-Second Take

Published September 5, 2026 · Updated November 3, 2026

Quick answer: Wan 3.0 is Alibaba Cloud's video generation model on the Model Studio platform, called through an asynchronous submit-then-poll API, in two tiers: the standard wan3.0-video and the faster wan3.0-video-prime. The four numbers that matter for series production: clips of 2–30 seconds, three resolutions (480P, 720P, 1080P), up to 10 reference images plus 5 reference videos and 5 reference audios per call, and prompts of up to 20,000 characters. Billing is per second of successfully delivered video; failures are free. It is one of the two video paths in SceneMixer, the "Wan 3.0 · 720P" tier is the one marked recommended, and it is one of the models available in the 30-second take mode. Below: a parameter table, how to feed references, the two moderation gates, prompt rules and how the tier behaves inside SceneMixer.

## Parameter table (checked 2026-09-05)

| Item | Wan 3.0 | Why it matters for short drama |

| Tiers | wan3.0-video (standard) / wan3.0-video-prime (fast lane) | Prime is a speed lane, not a quality upgrade; list price about 1.5× standard |

| Call shape | Asynchronous: submit returns a task id, poll until SUCCEEDED or FAILED; submissions need the X-DashScope-Async: enable header | Ten segments of an episode queue in parallel and come back one by one |

| Resolution | 480P / 720P / 1080P | 720P is enough for vertical drama; SceneMixer defaults to 480P and offers 720P and 1080P |

| Clip length | 2–30 s, integer | Twice the usual 4–15 s ceiling; a full two-hander fits in one segment |

| Reference images | Up to 10 | Lead full-body sheet, face crop and scene plate together; cross-shot consistency rides on these |

| Reference video / audio | Up to 5 each, 15 s total each | Audio locks a character's voice; video can continue the previous segment's motion |

| Prompt | ≤ 20,000 characters; optional server-side prompt extension | Room for a complete storyboard; keep extension off (see below) |

| Other | seed for reproducibility; aspect ratio, including adaptive to the reference | Use seed in experiments, not in production |

| Billing | Per second of output; only successfully generated video is charged, failures and moderation blocks are free | Retries never cost you for a failure |

Source: Alibaba Cloud Model Studio, Wan 3.0 video generation API reference, help.aliyun.com, accessed 2026-09-05; the figures match SceneMixer's integration code. Parameters change; the official page wins.

## Feeding references: ten images is a ceiling, not a target

- Characters: one 9:16 full-body sheet on white per character in the shot, plus a face crop for leads. Why white and directional light: see Character Consistency. Light baked into a reference is inherited by every shot.

- Scene: one empty plate with the light source stated, no fog, beams or silhouettes.

- Props: only plot-critical props, white-background close-ups.

- Audio: reference audio locks the character's voice; SceneMixer matches voices from a preset library and attaches them automatically.

- Aspect ratio: pass 9:16 explicitly for vertical drama. With adaptive ratio the model follows the first reference image, and a mixed bag of portrait and landscape references produces surprise framing.

## Two moderation gates: free, but they stall the schedule

| Error code | Meaning | Typical trigger | How to avoid |

| DataInspectionFailed | Content safety check failed | Gore, sexual content, sensitive figures or slogans | Check the script against your platform's rules first (see the compliance checklist); keep borderline description out of the storyboard |

| IPInfringementSuspect (Input / Output) | Suspected intellectual-property infringement in the input or the output | Studio names, trademarks or famous artist names in the prompt or the references; reference images generated in a studio's style can be rejected on the output side even after the text is fixed | Describe the look ("feature-film-grade American family animation 3D rendering"), never "in the style of studio X"; regenerate references that were made under a trademarked style |

A real case from 2026-09-04: a project set to a "Pixar-style 3D animation" look had six of ten segments rejected on input and the remaining four rejected on output. Removing the studio name from the style descriptor fixed it; the same account's next project passed eleven of eleven. Neither gate charges anything on SceneMixer, but blocked segments have to be regenerated with a different style or character design.

## Prompting: natural language, and keep auto-extension off

- Turn off prompt extension. Server-side rewriting reorders the cut points, timings and character orientation that the storyboard already fixed. The storyboard text you send is the finished product.

- Write dialogue in the official speech format: He says: "line", relaxed tone, medium pace. Quotation marks delimit the spoken line; emotion and pace go in the plain clause after it. Never wrap performance notes inside the quotes or the model will read them aloud.

- Causes, not effects: write where the light comes from ("a pendant throws warm-white light from upper right"), not where it lands; no 8K / masterpiece / volumetric trigger words.

- Long segments, short shots: fill a segment to 15 or 30 seconds with 2–4 shots and control pace through shot length.

- Only name who is in frame: never name absent people; nouns inside negations act as attractors.

## Inside SceneMixer

- Choosing the tier: new projects use "Wan 3.0 · 480P" by default; "Wan 3.0 · 720P" (marked recommended) and "Wan 3.0 · 1080P" are in the model menu. SceneMixer runs its Wan tiers on the standard wan3.0-video lane, not prime. Current credit rates are on the pricing page.

- 30-second take: at the top of the episode list in step 3, switch the segment mode to "30s take". That episode's storyboard is then generated in 30-second segments and its model list keeps only models that support 30 seconds, including Wan 3.0. Use it for dialogue-heavy episodes you do not want cut at 15 seconds.

- References are automatic: character sheets, face crops, scene plates and matched voices are attached per segment by the pipeline; nothing to upload by hand.

- Turnaround: over the last 14 days the median Wan task on the international site took about 8 minutes per segment; an episode's ten segments queue in parallel, so edit the next episode's storyboard while you wait.

- Refunds: tasks blocked by moderation, rejected for parameters or timed out upstream release their credits automatically; credits settle only when a clip is delivered, and regenerating one segment charges only that segment.

To see full episodes produced with Wan, open the sample dramas; each project lists the model and resolution used.

## FAQ

### How long a clip can Wan 3.0 generate in one call?

The API accepts an integer duration from 2 to 30 seconds. SceneMixer splits storyboards into 15-second segments by default; switching an episode to the 30-second take mode in step 3 limits that episode's model list to models that support 30 seconds, which includes Wan 3.0.

### Do I pay credits when a Wan generation fails?

No. Alibaba Cloud Model Studio bills only successfully generated output video; failed calls and processing errors are not charged. SceneMixer mirrors that: when a task fails or is blocked by moderation, the frozen credits are released automatically, and credits are only settled once a clip is delivered.

### Why does a prompt mentioning a studio's animation style get rejected?

Model Studio runs an intellectual-property check on both input and output; the error code IPInfringementSuspect means suspected infringement. Studio names, trademarks and well-known artist names trigger it, and reference images generated in that style can still be rejected on the output side even after the text is changed. Describe the look instead, for example "feature-film-grade American family animation 3D rendering".

### Standard or prime: which Wan tier should I use?

Prime is a faster lane for the same model, not a higher-quality tier, and its list price is about 1.5 times the standard tier. Use prime for batch output and deadlines, standard for polishing a single episode on a budget. SceneMixer's Wan tiers all run on the standard lane and are billed per second of output.


---

# Wan 3.0 vs Seedance 2.5: Same Rules, Different Price | SceneMixer

URL: https://scenemixer.com/guides/wan-3-vs-seedance-2-5-on-this-site/
How the Wan 3.0 and Seedance 2.5 families compare on SceneMixer: the price per second at each resolution, the longest take, and the references they share.

# Wan 3.0 vs Seedance 2.5 on this site

Quick answer: Wan 3.0 and Seedance 2.5 take the same references and both run takes up to 30 seconds with sound; what differs is the price at each resolution: 10 to 38 credits per second on Wan 3.0 against 19 to 75 on Seedance 2.5.

The still above is from Rust City Knockout, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

Wan 3.0 · 480P and Seedance 2.5 · 480P are the two entry tiers.

## What is identical

Segment length (up to 30 seconds), native sound, aspect ratio handling, the reference slots, the content checks and the refund on rejection, the way versions are kept and composited. A segment set up for one family runs on the other on SceneMixer with nothing changed but the tier.

## What differs

The price per second at each resolution:

| Resolution | Wan 3.0 (credits per second) | Seedance 2.5 (credits per second) |
| 480P | 10 | 19 |
| 720P | 19 | 38 |
| 1080P | 38 | 75 |

The labels also differ, which matters if you keep notes per family or share settings with collaborators who know one name.

## How to choose

By price: compare the two columns above at the resolution you need. By habit: if a team has standardised on the Seedance 2.5 names and its drafts already sit on that family, staying there costs nothing but credits. There is no quality reason on SceneMixer to prefer one family over the other at the same resolution.

## What this page does not claim

Nothing about what runs underneath either name; this page describes what each tier does on this site — segment length, sound, references, checks, price — which is what your credits buy.

## FAQ

### Is Seedance 2.5 higher quality than Wan 3.0 here?

At the same resolution the two behave the same on SceneMixer; the difference you will notice is price.

### Can I move a segment from one family to the other?

Yes. Pick the tier in SceneMixer and regenerate; references, voices and the storyboard are unchanged.

### Do both support 30-second takes?

Yes, all six tiers.

### Do both take the same references?

Yes: SceneMixer takes up to ten images and five voice clips per generation.

### Which is the site default?

Wan 3.0 · 480P.

Updated November 3, 2026


---

# What a Paid Plan Actually Changes | SceneMixer

URL: https://scenemixer.com/guides/what-a-paid-plan-changes/
Monthly credits, more generations at once, more upload room, unlocked features and a priority queue on the top tiers. It does not make one generation faster.

# What a Paid Plan Actually Changes

Quick answer: Four things change: credits arrive at the start of each billing period, more of your jobs run at the same time, you get more room for uploads, and a handful of features unlock — plus a priority queue on the top three tiers. Three things do not: a single generation does not get faster, a generation costs the same number of credits for everyone, and nothing about the result is different.

Plans start at $7.99 a month; the pricing page has every tier's price and monthly allowance. How many jobs each runs at once is on this page, in the table below.

## The allowance is the main thing

Each plan grants a number of credits at the start of every billing period. They land in the same balance as credits you buy and are spent the same way, in the same order, on the same things. They are not a separate wallet with its own rules, and they do not expire at the end of the month.

That is the economic shape of a plan: it is a standing order for credits at a better rate than buying them one pack at a time, not a licence that changes what the product does.

## More running at once

Every account runs a fixed number of generations in parallel; anything past that waits and is promoted automatically as slots free. On SceneMixer, a free account runs three at once. The paid tiers run five, eight, ten, fifteen and twenty, in order.

Read that precisely: it is how many at once, not how fast each one is. A single fifteen-second segment takes exactly as long on every plan, because it is the same model doing the same work. What changes is whether your other nine segments are waiting behind it or running alongside it, which is the difference that actually shows up on a deadline.

The top three tiers also carry a priority queue: your jobs come off our queue before other users'. It does not reorder anything at the model service itself and it does not make a single generation faster — it only decides who waits when we are busy.

## What unlocks

The line SceneMixer draws is: correcting what the AI produced is free; injecting your own material or your own creative direction is paid.

So editing a character's description, rewriting a storyboard, switching takes and asking for a plain regeneration are available on every account, including a brand-new one. Uploading your own images or audio, writing a visual style in free text, adding an asset that the script never mentioned, and retouching an image with your own prompt need a plan.

Available on every account is not the same as costless: any generation, including a re-generation, spends credits at the same rate whatever plan you are on. Editing text and switching between takes you already have are the ones that genuinely cost nothing.

Separately, video on a paid plan is delivered without the visible AI badge by default, and you can switch it back on. What is permitted there is governed by the terms.

Upload storage also goes up with the tier. Only material you upload counts against it — everything generated for you is stored at our expense on every plan.

## What does not change

A generation on SceneMixer costs the same number of credits whichever plan you are on. The tier picker offers the same tiers. The models are the same models. Nothing about the finished clip is different because of what you pay monthly.

There is currently no member discount on credit packs on either site. If that changes, it will be on the pricing page, which reads live configuration — not here.

## Changing or stopping a plan

Upgrading takes effect immediately and the new tier's allowance is granted at once. Time you had already paid for on the lower tier is not thrown away — it is pushed back and resumes after the higher tier finishes.

Downgrading mid-subscription is not possible. You can subscribe to a lower tier once the current one ends. This is a deliberate refusal rather than a missing feature: a downgrade in the middle of a period either has to claw back credits already granted and spent, or hand out the difference for free.

Cancelling stops the renewal. You keep the plan and its benefits until the end of the period you have already paid for. There is no pro-rata refund of the current period, and credits are not cashable or transferable — the ones already in your balance stay there and stay spendable.

Cancelling only means something on a payment method that renews by itself. If you paid in a way that does not auto-renew, there is nothing to cancel; the plan simply ends when the period does.

## Is it worth it

An honest test: look at how many credits you spent last month and how often you were waiting on your own queue. If you spent less than the smallest allowance and were never blocked by your own slot count, a plan is not saving you anything — buy a pack when you need one.

The case for a plan is volume and parallelism: you are generating most weeks, and you want ten segments in flight rather than three. Not speed, and not quality.

| Changes with a plan | Does not change with a plan |
| Credits granted every period | What one generation costs in credits |
| How many jobs run at once | How long one generation takes |
| Upload storage | Which model tiers are offered |
| Your own uploads, free-text style, manual assets, prompt retouch | The models underneath |
| A priority queue on the top three tiers | How the model service orders its own work |
| Video delivered without the AI mark by default | The credit price of anything |

## FAQ

### Is there an unlimited plan?

No. Every SceneMixer plan works in credits: a paid plan adds credits each period and the extra features listed above, and each generation spends credits as usual. The pricing page shows what each plan includes.

### Does a plan remove the AI label from shared links?

No. Your downloads can be label-free on a paid plan, but a video shared through a SceneMixer link always shows the label.

### Does a plan make generation faster?

No. A plan on SceneMixer raises how many of your jobs run at the same time. One segment takes the same time on every plan, because it is the same model doing the same work.

### Do plan credits expire at the end of the month?

No. SceneMixer grants them into the same balance as credits you buy, and they stay there.

### Do members pay less per generation?

No. A generation costs the same number of credits on every plan. There is currently no member discount on SceneMixer's credit packs either; the pricing page is the live source.

### Can I downgrade?

Not in the middle of a subscription. On SceneMixer, you can subscribe to a lower tier once the current period ends.

### What happens to my credits if I cancel?

They stay in your SceneMixer balance and remain spendable. Cancelling stops the renewal; you keep the plan until the period you paid for ends, and there is no pro-rata refund of it.

### Do I get priority in the queue?

On the top three tiers, yes: your jobs are taken off SceneMixer's queue before other users'. It does not reorder anything at the model service and it does not speed up a single generation.

Updated November 3, 2026


---

# What Ask AI Can Check in Your Account, and What It Cannot Do

URL: https://scenemixer.com/guides/what-ask-ai-can-check-in-your-account/
Since October 2026 the Ask AI window can read your own credits, tasks, projects, orders and plan to answer you. What it sees, how to ask, what it cannot do.

# What Ask AI Can Look Up in Your Own Account

Quick answer: Since October 2026, the Ask AI help window can look things up in your own account when your question needs it: your balance and held credits, recent credit history, recent generations and why one failed, your projects with their episodes and segments, your top-up and membership orders, your plan, referral rewards and storage. It reads, it never changes anything: it cannot refund, cancel, regenerate or edit. It only sees your account, never anyone else’s. Ask in plain words and name what you mean — “why did segment 5 of episode 2 in Night Market fail?” — and the answer comes with links to the guides it used.

Open Ask AI (the capsule in the bottom-right corner, or the speech-bubble icon in the top bar), tap Where did my credits go? under Common questions, then press Ask this.

## What changed

The help window used to answer from the guides only and could not see your account. Since October 2026 it can also read your own records, and its greeting says so: You can also ask about your own projects, credits, membership and top-ups: I’ll check your account’s records to answer, and I can only see yours.

If the greeting does not include that sentence, looking up accounts is switched off for the moment and the window answers from the guides only. Everything else about the window — where it is, the hourly and daily limits, sending feedback when it does not solve the problem — is unchanged. The Ask AI window and official replies.

## What it can look up

| Area | What it can see | Ask like this |
| Credits | Your balance, credits held right now and which tasks hold them, your recent credit history | “Where did my credits go this week?” “Why are some of my credits frozen?” |
| Generations | Your recent tasks, their status, and in plain words why one failed (blocked by review, a temporary service problem, a reference image issue…) | “My last video failed. Did I get the credits back?” |
| Projects | Your projects, their episodes and segments; for a segment you name, its storyboard text | “In Night Market, episode 2, segment 5: why is the other guy in the shot?” |
| Orders and plan | Your credit top-ups, membership orders and current plan | “Did my last top-up go through?” “Which plan am I on?” |
| Referrals | Your referral rewards | “Have I earned anything from my invite link?” |
| Storage | How much of your file space is used | “How much storage do I have left?” |
SceneMixer looks only at what your question is about: ask about credits and it reads your credits, not your projects. Lists are the most recent entries, not your whole history; when there are more than it was given, it says so instead of pretending the list is complete. If one part could not be read at that moment, it says that part is unavailable and answers the rest.

## How to ask so it finds the right thing

- Name the project. Use the name as it appears on the project card. If two projects have similar names, it asks which one you mean.
- Give the episode and segment number for anything about one clip: “episode 3, segment 12”.
- Ask from the page. When you open the window inside a project or an episode, SceneMixer knows which one you are looking at.
- Say it is about your account. A general question (“how do refunds work?”) gets a general answer from the guides and a pointer to where you can check yourself; “check my credits” or “look at my last task” tells it to read your records.
- Use the two account questions under Common questions: Where did my credits go? and A generation failed — did I get my credits back? Tap one, read the general answer, then press Ask this to have it check your own records.

## What it cannot do

- Change anything. No refunds, no cancelling a plan, no regenerating, no editing a storyboard, no moving credits. SceneMixer explains and points you to the button or page that does it.
- See other people’s accounts. Asking about a sample or a shared project that is not yours does not open it; it can only describe your own.
- See payment details or passwords. Card numbers and passwords are not in what it can read.
- Promise an outcome for credits under review or a payment still processing. It tells you the current state and what happens next.
For anything that needs a person — a refund decision, a charge you do not recognise — use Still not solved? → Send feedback in the same window: your questions and the answers go with it. How to get help.

## Questions it answers well

| Question | What it does |
| Where did my credits go? | Reads your recent credit history and tells you which project and step each charge went to. |
| A generation failed — did I get my credits back? | Finds your recent failed tasks, says in plain words why each failed, and whether the credits came back or are held for review. |
| Why are credits frozen? | Lists the tasks holding them right now: running jobs, or ones held for review. |
| Why does this shot have the wrong person in it? | Reads the storyboard text of the segment you name and points at the line that causes it. |
| Did my payment go through? | Checks your recent orders and their status. |

## Privacy

- SceneMixer reads only when your question needs it, and only your own account.
- It cannot write to your account; every lookup is read-only.
- The text of help conversations is kept for 30 days and then deleted.
- Do not type personal or sensitive details into the window; it does not need them to look things up.
Support staff reading your account when you report a problem is a separate thing, also read-only: What support can see and do in your account.

## Checking it yourself

Everything the window can read, you can also open yourself:

- Credits: click your balance at the top right; the latest entries are listed, and View all opens the full Credit History on the billing page.
- Orders: Order History on the billing page.
- A task: the segment’s card on the timeline and its message. What each task status means.
- Storage: the Asset Library. The asset library and the quota.

| It can | It cannot |
| Read your balance, held credits and recent credit history | Refund or move credits |
| Tell you why a task failed, in plain words | Regenerate or stop a task |
| Read the storyboard text of a segment you name | Edit the storyboard |
| Check your orders and plan | Cancel, upgrade or renew a plan |
| See your own account | See anyone else’s account or project |

## FAQ

### Can Ask AI see my account?

Yes, since October 2026, the Ask AI on SceneMixer can see your account, but only your own and only to read. It looks at your credits, tasks, projects, orders, plan, referral rewards or storage when your question is about them.

### Can it refund me or cancel my plan?

No. SceneMixer cannot change anything. It explains what happened and points you to the right button; for a decision only a person can make, send feedback from the same window.

### Can it see other people's projects?

No. Asking about a sample or someone else's shared project does not give it access; SceneMixer only reads projects in your account.

### How do I make it check my records instead of answering in general?

Say so, for example check my credits or look at my last task, name the project and the episode and segment, or press Ask this under one of the account questions in SceneMixer's Common questions.

### Why does it say there are more entries than it listed?

SceneMixer reads the most recent entries, not your whole history, and says when there are more. For the full list open Credit History or Order History on the billing page.

### Is it safe? What does it keep?

On SceneMixer, every lookup is read-only and limited to your account. The text of help conversations is kept for 30 days and then deleted. Do not type personal or sensitive details into it.

### Does it cost credits?

No. Asking is free on SceneMixer; there is an hourly and a daily limit on the number of questions.

### The greeting does not mention my account. Why?

Account lookups are switched off for the moment, so SceneMixer answers from the guides only. You can still check everything yourself on the billing page and in your projects.

Updated November 3, 2026


---

# What Comes Out: Aspect, Length, Resolution | SceneMixer

URL: https://scenemixer.com/guides/what-comes-out-video-specs/
The spec of a finished clip: aspect ratio, how long one segment can be, what resolution each tier gives, and which file you download.

# What Comes Out

Quick answer: Projects are 16:9 or 9:16, and a new project is 16:9 unless you change it — that setting governs video only. Segments run up to 15 or 30 seconds, chosen per episode, and the story decides how long each one actually is. Resolution comes from the tier you pick. What you stream on the site is a lighter copy; what you download is the full-quality one.

Every number below is visible on the sample projects.

## Aspect ratio

A project is either landscape or portrait, and you choose once, in the first step. A new project starts as 16:9 landscape. Portrait 9:16 is one click away and is what you want for a vertical short-drama feed.

The aspect governs video, and only video. Setup images do not follow it: character sheets are always 9:16 because a full-length standing figure needs the height, and location and prop plates are always 16:9 whatever the project is. A plate is a reference the video model reads, not a frame it has to fill — the shot gets composed at generation time.

Once you continue past the first step, the aspect ratio is fixed for the project. To make the same story in the other shape, start a new project with the same script; the cast can come across from the asset library.

## How long one segment can be

A segment is one generation. You choose the ceiling when you generate an episode's storyboard — Standard 15s or Long take 30s, per episode — and the storyboard is written within it. It is a ceiling, not a target: a segment with two short shots may run only a few seconds. The tier then has to be able to deliver it: the default and the recommended tiers on SceneMixer both go to thirty seconds, while the older model families stop at fifteen and the picker greys them out for a segment that is too long.

Within a segment, the storyboard divides the time into numbered shots with their own durations, and those durations add up to the segment length. That is the arithmetic the editor is checking when it complains about a segment being over. When the shots add up to more than the episode's limit, the silent ones are tightened first, and if that is not enough the segment is split at a shot boundary. Nothing gets cut off the end.

To lengthen one segment, raise the seconds of its shots in the editor, up to the ceiling; changing segment duration has the details. Which is worth knowing when you plan: writing towards the ceiling gives the model a whole continuous take to work with, and continuity inside one generation is free in a way that continuity across two generations never is.

## Resolution

Resolution is a property of the tier, not a separate setting. Lower tiers are lower resolution and cheaper per second; higher ones are more pixels and cost proportionally more. SceneMixer's picker lists each tier's resolution next to its per-second price, and the pricing page has the full table.

The default is deliberately a low tier. It is the cheapest way to find out whether a shot reads before you spend a higher one on it.

## Two files, not one

Each finished segment exists as more than one file: the player on the site streams a lighter copy, and the download button gives you the full-quality one. So if a clip looks softer in the browser than you expected, download it before drawing conclusions.

When every segment in an episode came out of the same tier, compositing copies the streams rather than re-encoding them — the composite is the same picture as your segments, joined. Mix resolutions inside one episode and the join has to re-encode onto a single canvas, which does cost a generation of quality. If you iterate cheap and re-shoot the keepers expensive, re-shoot all of them. Compositing and sharing covers that step in full.

## The label

SceneMixer adds two marks to delivered video: a small badge burned into the corner of the picture and a line of AI-generated-content metadata, both written in the same pass. Without a paid plan you always get them; on a paid plan we add neither by default and you can switch them on. Because the badge is part of the picture rather than an overlay, when it is there it survives download, re-upload and re-encoding. We do not add a badge to generated images. Anything the model services mark at generation time is a separate layer; section 05 of the terms governs all of it.

What is permitted is governed by the terms, and what you get to keep goes through it in detail.

| Property | Value | Set by |
| Project aspect | 16:9 or 9:16 — new projects are 16:9 | You, in step one |
| Character sheet aspect | Always 9:16 | Fixed |
| Location and prop image aspect | Always 16:9 | Fixed |
| Segment length | Up to 15s or 30s | You, per episode — if the tier can take it |
| Resolution | Listed per tier in the picker | The tier |
| Streamed file | Lighter re-encode for playback | Automatic |
| Downloaded file | Full quality | Automatic |
| Composited episode | Copied when all segments match; re-encoded if they do not | Automatic |

## FAQ

### Is the default vertical or horizontal?

Horizontal. A new SceneMixer project starts as 16:9, and 9:16 is one click away in the first step.

### Can a segment be longer than fifteen seconds?

Yes — pick Long take 30s when you generate that episode's storyboard. The tier has to be able to take it; SceneMixer's default and recommended tiers both go to thirty. If the shots still add up to more than the limit, the silent ones are tightened and the segment is split, never truncated.

### Why does the video look softer in the browser than after downloading?

SceneMixer's player streams a lighter copy for smooth playback. The download is the full-quality file.

### Does compositing re-encode my segments?

Only if you mixed tiers within the episode. When every segment matches, SceneMixer copies the streams and nothing is re-compressed.

### Can I change the aspect ratio after generating?

No. SceneMixer fixes it once you continue past the first step. For the other shape, start a new project with the same script.

### I chose Long take 30s, so why is a segment only a few seconds?

On SceneMixer, thirty seconds is the most a segment can hold, not a length every segment is filled to. Raise the shot seconds in the editor to lengthen one.

### Do generated images carry the AI badge too?

SceneMixer applies its badge to video, not to images. Labelling rules are in section 05 of the terms.

Updated November 3, 2026


---

# What Each AI Video Clip Fix Costs, and When You Are Refunded

URL: https://scenemixer.com/guides/what-each-clip-fix-costs-and-when-you-are-refunded/
Trim, AI visuals, Lip sync, Re-perform and AI edit of the storyboard text, priced side by side, with worked examples, refund rules and a comparison with regenerating.

# What Each Clip Fix Costs, and When the Credits Come Back

Quick answer: Trim: 5 credits per trim. AI visuals and Re-perform: 2× the per-second price of the tier the clip was made on, times the clip’s length. Lip sync: 1 credit per second of clip. Enhance: per second of clip, by option and target resolution. AI edit of the storyboard text: 1 credit per revision. Every price is on the button before you press it. The credits are held when you press and come back if the fix fails; none of these has a free allowance.

Before choosing between a fix and a regeneration, compare the two prices on screen: the fix price is on the panel's button, the regeneration price on the segment's generate button.

## The prices

- Trim — 5 credits per saved trim, whatever the length. If the site ever sets it to free, the button says Free.
- AI visuals — 2× the per-second price of the clip’s tier, per second of clip, rounded up to whole seconds, at least two seconds. On clips longer than 15 seconds only one half is edited, and only that half is charged. Clips made on the MiniMax H3 tiers are priced at the tier they are edited on. Per-tier table.
- Redub → Lip sync — 1 credit per second of the whole clip, rounded up.
- Redub → Re-perform — the same as AI visuals on that clip; clips of 15 seconds or less only.
- Enhance — per second of clip, rounded up, at a rate set by the option (Instant HD, AI Detail, Cinema Restore) and the resolution it goes to; the price for the segment is on each card and on the button. Enhance a clip.
- AI edit (the storyboard text, not the video) — 1 credit for each revision that comes back. AI edit the storyboard text.

## Worked examples

A 10-second clip made on Wan 3.0 · 480P (10 credits a second to generate):

- Regenerate it: 10 × 10 credits.
- AI visuals on it: 10 × 20 credits — twice the regeneration.
- Lip sync: 10 × 1 credit. Re-perform: the same as AI visuals.
- Trim: 5 credits.

The same clip on Wan 3.0 · 720P: regenerate 10 × 19, AI visuals 10 × 38. On SceneMixer, the pattern holds on every tier: a picture fix costs twice a regeneration of the same length, and a voice-only fix costs a small fraction of either.

## When a fix is worth it, and when regenerating is cheaper

A fix is worth it when the clip is right apart from one thing and a regeneration would roll the dice on everything else — the acting, the framing, the timing you already like. Regenerating is cheaper when the storyboard is right and you just need another take, or when the clip was made on an expensive tier and a cheaper one would do. And some problems only a regeneration fixes: framing, camera angle, positions, extra people. Which fix for which problem.

## When the credits come back

Trim, AI visuals, Lip sync, Re-perform, Enhance: the price is held when you press. If the fix does not go through, SceneMixer refunds it. An AI visuals or Re-perform edit blocked by content review is refunded too, and counts toward the same pause as a blocked generation. In the rare case where a result cannot be confirmed either way, the credits wait as held for review instead of being charged. Credits held for review.

AI edit of the storyboard text: charged only when a revision comes back and is shown to you, and the review says so (Charged ✦ 1). If you discard that revision, the credit is not returned — the revision was made. Every failure, refusal or unusable reply costs nothing.

## A trim can save on the composite

SceneMixer prices the episode composite by the actual length of each segment’s active version, 1 credit per second, so a trimmed segment costs less to join. Which segments are joined.

## In your credit history

Each fix is one line in your credit history, named after the fix (Trim, AI visuals, Redub, Enhance clarity) with the segment it was made on: In progress while it runs, then the amount spent, or the amount struck through and Refunded if it fails — the same as a generated segment. Reading your credit history.

| Fix | Price | Refunded when |
| Trim | 5 per trim | It fails |
| AI visuals | 2× tier per second × length | It fails or is blocked by review |
| Redub → Lip sync | 1 per second of clip | It fails |
| Redub → Re-perform | Same as AI visuals | It fails or is blocked by review |
| Enhance | Per second, by option and resolution; on the card | It fails |
| AI edit (text) | 1 per revision shown | It fails, is refused or comes back unusable |

## FAQ

### Do the free sheets or first-project perks cover any of this?

No. SceneMixer always charges these fixes.

### Is AI visuals ever cheaper than regenerating?

Not for the same tier and length on SceneMixer: it is twice the price. It is worth it when you want to keep the rest of the clip exactly as it is.

### Why was I charged for an AI edit I discarded?

The revision was made and shown to you; discarding it does not undo the work. On SceneMixer, failed attempts are free.

### Can I stop a fix and get the credits back?

SceneMixer has no stop button for fixes. If a fix fails, the credits come back by themselves.

Updated November 3, 2026


---

# What Each Task Status Means: Pending, Generating, Failed | SceneMixer

URL: https://scenemixer.com/guides/what-each-task-status-means/
Pending, Waiting, Queued, Generating, Done, Failed, on hold: what every status word on a segment, image or composite means, and whether you need to do anything.

# What Each Status Means

Quick answer: Every job on SceneMixer — a segment’s video, a character’s image, an episode’s composite — shows a one-word status where it lives. Pending: nothing generated yet. Waiting or Queued: submitted, not started. Generating: running, with a usual time shown. taking longer than usual, still running: still running. Done: delivered and charged. Failed: not delivered; read the message. Only Failed ever needs something from you.

Open a finished episode: every card on its timeline is in the Done state.

## Segment videos

Pending. The segment has text but no video. Press Generate Video when it is ready.

Waiting. You pressed generate, but it is behind other jobs of yours. There is a limit to how many run at once for one account (paid plans raise it); the rest wait their turn and start on their own.

Queued, starting shortly. Submitted and about to start. Credits are held.

Generating. Running. The pane shows the usual time for the tier — typically ten to twenty minutes. Leave the page if you like. Why it takes time.

taking longer than usual, still running. Past the usual time, not failed. Wait.

Stopping. You pressed Stop on a parse, storyboard or AI Create job; it is being cancelled. Video clips have no Stop. What stopping refunds.

Done. The clip is in the preview pane; credits charged; a new entry in Version History.

Failed. Not delivered. The message says whether credits came back, are on hold, or the clip was blocked. Failure messages.

## Images in step 2

A character, scene or prop card is Pending (no image yet), Generating, Done (the image is shown), Failed (regenerate or upload), or Uploaded (your own image, which is never overwritten by generation). SceneMixer's progress line at the top of the tab counts how many of the batch are done.

## Composites

Composite Episode is available once every segment is Done; before that the button explains how many are left. While it runs the record says Processing; then Done with a play button, or Failed with a message and Composite Again. Compositing.

## Episodes and projects

An episode whose storyboard is being written says so on its card in step 3 and in the editor; generating a video for it is refused until the storyboard lands. A project being read shows the reading stages in step 1. An ad shows its stage on its progress page; a remake shows its reading stages. None of these need anything from you while they run.

## On hold

One status is not on a card: credits on hold pending a support review, shown in a failure message. A SceneMixer support person checks what happened and either returns the credits or delivers the clip. Do not resubmit.

| Status | Means | You |
| Pending | No video yet | Generate when ready |
| Waiting | Behind your other jobs | Nothing |
| Queued, starting shortly | Submitted, about to start | Nothing |
| Generating | Running; usual time shown | Nothing — leave the page if you like |
| taking longer than usual, still running | Past the usual time, still running | Wait |
| Stopping | Being cancelled | Nothing |
| Done | Delivered, charged | Watch, keep, or regenerate |
| Failed | Not delivered | Read the message |
| On hold (in a message) | Under support review | Do not resubmit |

## FAQ

### What is the difference between Waiting and Queued?

Waiting is behind your own other jobs — the per-account limit. Queued is submitted and about to start.

### It has said Generating for a long time.

If it says taking longer than usual, it is still running. Wait for it; a running clip cannot be stopped.

### Does Done mean I was charged?

Yes. On SceneMixer, credits are held when you confirm and charged when the clip is delivered.

### What does Pending mean on an image card?

No image has been generated for that entry yet. Generate it, or press SceneMixer's Generate All.

### Why is the composite button not available?

Every segment must be Done first. The composite button in SceneMixer says how many are left.

Updated November 3, 2026


---

# What Follows Your Account and What Stays in This Browser

URL: https://scenemixer.com/guides/what-follows-your-account-and-what-stays-in-this-browser/
Everything you make is in your account and shows on any device. A few conveniences are per browser: the language you chose, unsent drafts. Which is which.

# Switching Devices: What Follows Your Account, What Stays in This Browser, and How to Replay the Tour

Quick answer: Projects, sheets, clips, credits, orders, membership, your model tier preference, the badge setting and the record of which guides you have seen are stored with your account and appear on any device once you sign in. A few things are kept only in the browser you set them in: the interface language you picked (a new browser starts from its own language), text you typed into a box but did not submit, panels you collapsed, and the anonymous usage id. Signing in on two devices at once is fine; both see the same projects and task states. To see the onboarding tour again, use the setting for it or add ?tour=1 to a page’s address.

Sign in on a second device: your projects and balance are there; the interface language is that browser’s until you pick one.

## What lives in the account

Everything that costs credits or that you made: projects and their episodes, sheets, states, storyboards, clips, composites, the library, orders and credit history, membership, referrals. Also your settings: preferred tiers, AI-badge choice, watermark preference, and the guides marked seen. Settings Item by Item.

## What stays in the browser

The interface language you chose from the menu (until you choose it on the other device too; a browser you have not chosen on picks its own language), drafts in the script box you have not submitted, which panels and hints you dismissed, the last tab you had open in some views, and the anonymous statistics id. Clearing site data in the browser resets these and nothing else; your SceneMixer account is untouched.

## Two devices at once

Allowed. Generations started on one show on the other within moments; a segment being edited on one device is saved when you save it, and the other device sees the saved version on its next refresh. Avoid editing the same segment on both at the same time; the last save wins. Does Generation Continue?.

## Replaying the onboarding tour

Each page’s tour shows once per account. To see it again, open Settings and choose to show the tour again, or add ?tour=1 to the page address. Sample projects run the same tours; ad projects do not have one.

## Browsers

Any current browser works on SceneMixer: Chrome, Edge, Firefox, Safari and their mobile versions from the last few years. Very old browsers may fail to open the storyboard editor or the live progress connection; update rather than switch settings. Private windows work but forget the per-browser items when closed.

| Thing | Kept in |
| Projects, sheets, clips, library | Account |
| Credits, orders, membership | Account |
| Tier preference, badge setting | Account |
| Guides seen | Account (and browser) |
| Interface language choice | Browser |
| Unsubmitted drafts, collapsed panels | Browser |

## FAQ

### My language reset on my laptop.

Language choice is per browser; pick it there once.

### Can I be signed in on my phone and desktop at once?

Yes.

### The tour keeps showing on a new device.

Guides seen are synced to the account on sign-in; if you dismissed it while offline, dismiss once more.

### Does clearing my browser delete my projects?

No. Projects are in your SceneMixer account.

### Which browser is best?

Any current one; SceneMixer is tested on Chrome, Safari, Firefox and Edge.

Updated November 3, 2026


---

# What Is a World Model in AI Video? How SceneMixer Uses One

URL: https://scenemixer.com/guides/what-is-a-world-model-in-ai-video/
World models explained in plain words: what Genie 3, Marble and Helix World are doing in 2026, and how SceneMixer builds a world for every scene before it shoots.

# What Is a World Model? And Why Short Drama Needs One

Quick answer: A world model is an AI system that keeps track of a place, so what it shows next agrees with what it showed before: when you turn around, the door is still where it was. In 2026 the phrase mostly means explorable worlds you can move through, such as Google DeepMind’s Genie 3 or World Labs’ Marble. SceneMixer uses a narrower, practical version for short drama. Before any shot is made, each scene gets a world: a 360° view of the set, the furniture measured in it, and a place for every character. Then every camera is placed inside that world. It runs automatically on new projects on paid plans, with no extra credits per shot.

Drag the frame in Try the demo set below to look around a set built this way. In your own project the same view appears in the editor’s preview while a segment renders.

## A world model, in one paragraph

Most AI video models paint the next frames from what is already on screen. Nothing in them remembers what is behind the camera. Ask for a reverse shot and the model has to invent the other side of the room, and it invents a new one every time. A world model keeps a running picture of the place itself: the layout, the objects, who is standing where. Each new view is drawn from that picture, so views agree.

A film crew does the same thing without calling it anything. They build the set once, put the actors on their marks, and move the camera. The room does not change between takes because it is a real room. A world model is the AI version of having a set.

## Why everyone is talking about world models in 2026

As of October 2026, three projects show where the idea is going:

- Genie 3, from Google DeepMind. You describe a place and it generates a world you can move through in real time, frame by frame, as you steer. In January 2026 Google opened Project Genie, a research prototype built on it, to Google AI Ultra subscribers.
- Marble, from World Labs. The company co-founded by Fei-Fei Li turns text, images or video into a 3D world you can walk around and export. Version 1.1 came out in April 2026.
- Helix World. A preview released in October 2026 by Noiz AI with the Hong Kong University of Science and Technology: a world model that generates picture and sound together, in real time.
The open problems are the same ones short-drama makers already know. Faces and clothes drift over a long take. Things change when the camera looks away and back. Worlds hold together for a minute or so, not for an episode. And the phrase itself is used loosely: some products that call themselves world models are still drawing shot by shot.

## What a short drama needs from a world

A drama does not need a world you can wander for an hour. It needs one room that holds still for the length of a conversation, and comes back the same in the next episode.

- Shot and reverse shot agree. If she sits with the window behind her, his reverse shot shows the other side of the room, not a fresh wall.
- People keep their places. Whoever sat at the head of the table is still there after the cut.
- The room survives the episode break. The office in episode 1 is the same office in episode 6.
- Cameras make sense. An over-the-shoulder shot is taken from behind a real shoulder, at a real distance.
That is a much smaller job than an explorable world, and that is why it can be done reliably for every scene today. More on the craft side of this: Why characters swap sides between cuts.

## How SceneMixer builds a world for each scene

- The scene image you already have for the location is the starting point.
- A 360° view of the set is made from it, so the room exists all the way round, not only in the direction the image faced.
- The room is measured: walls, ceiling, doors and window openings, and where the furniture stands.
- Every character in the scene gets a place in that room, from what the storyboard says: at the table, by the door, in the passenger seat.
- Every shot’s camera is placed inside the same world, so the wide shot, the close-ups and the reverse all look into one room.
Outdoor scenes get a lighter world: the ground, the objects close by, and the distant view in four directions. A room is measured once and reused across shots and across episodes. How the scene image turns into a set, and how to get a good one: Your scene image becomes the set.

Camera 2

Camera 3

Camera 4Four cameras placed in the same demo set. Each view is worked out from the same 360° image, so the bar, the windows and the tables stay where they are from shot to shot.

## Try the demo set

This is the demo set from the SceneMixer homepage: a restaurant dining room with four characters and four cameras. Drag to look around; the view moves from camera to camera along the timeline at the bottom.

Try it: drag to look around the demo set. Four characters stand in the room, the numbered cameras show where each shot is taken from, and the view moves from camera to camera along the timeline at the bottom. The editor’s preview works the same way, with simpler figures.

In a real project the editor shows your own set this way while a segment is being generated. What every part of it means: Look around the set while a shot renders.

## Who gets it, and what it costs

SceneMixer's world model runs automatically on new projects on paid plans. There is no switch to turn on and no extra credits per shot; you pay for images and video exactly as before. What a paid plan changes.

Projects created before the world model launched, and projects on the free plan, keep the earlier automatic staging, which also holds seats and sides across cuts from the storyboard: Everyone stays where they sat.

## What it does not do

- It is not a game. You can look around the set while a shot renders, and step a little on some sets, but you cannot walk the world freely or play in it.
- You do not move people by dragging. The view is read-only. Positions come from the storyboard; change the text to change them. Write positions the world can read.
- SceneMixer does not set faces. Faces come from character images. Same face in every shot.
- It does not make every clip perfect. The video is still generated, shot by shot, inside the world. When one shot comes out wrong, edit and regenerate that shot: Edit one shot and regenerate.

| Question | Each shot drawn from scratch | With a world for the scene |
| Where is the door in shot 3? | Wherever that shot imagines it | Where it was in shot 1 |
| Who sits on the left? | Can flip at the cut | The same person, every cut |
| What does the reverse shot show? | A newly invented wall | The other side of the same room |
| The same office in episode 6? | Similar, not the same | The same measured room |
| What you set up | Nothing, and nothing holds | Nothing: on for new projects on paid plans |

## FAQ

### What is a world model, in simple terms?

An AI system that keeps track of a place (its layout, its objects, who is where) and draws each new view from that, so views agree. A video model without one invents the room again in every shot.

### Is SceneMixer’s world model the same thing as Genie 3 or Marble?

No. Those build worlds you can explore. SceneMixer builds a set for each scene of your drama, measures it, places the characters and the cameras in it, and generates the shots from there. Same idea, much narrower job.

### Can I walk around my world?

You can look around it. While a segment renders, SceneMixer’s editor preview shows the set in 360° and you can drag, tilt your phone, and on some sets take a small step. You cannot walk it freely, and nothing you do there changes the video.

### Does the world model cost extra?

No. SceneMixer runs it on new projects on paid plans with no extra credits per shot. Images and video are charged as before.

### Why doesn’t my project have it?

Projects created before it launched and projects on the free plan keep the earlier automatic staging. Sample projects and segments that are already finished do not show SceneMixer’s 360° view either.

### Do I have to set anything up?

No. SceneMixer has no switch. Write the storyboard as you normally would; the world is built from the scene image and the people are placed from the storyboard text.

### Does it stop my character’s face from changing?

No. SceneMixer’s world holds the room and the positions. Faces come from each character’s image; that has its own checklist.

### Is the same world used in the next episode?

Yes. SceneMixer measures a room once and reuses it across shots and episodes. If you make a new image for that scene, the room is measured again from the new image.

### Can I export the world or use it somewhere else?

No. The world exists to shoot your drama inside SceneMixer. There is no export.

Updated November 3, 2026


---

# What Is Previs? Previsualization for Filmmakers and AI Video

URL: https://scenemixer.com/guides/what-is-previs/
What is previs (previsualization)? A rough, early version of scenes made before shooting to test pacing, blocking and dialogue, and how AI video is used for it today.

# What is previs?

Quick answer: Previs, short for previsualization, is a rough early version of a film's scenes, made before the real shoot, to test pacing, blocking, camera and dialogue while changes are still cheap. It used to mean storyboards and animatics; now many teams use generated video. In SceneMixer a voiced, rough version of a scene or a pitch trailer is a common use: it shows tone, cast and pacing, not a finished broadcast.

Take the one scene you are most unsure of, write it as a page, and make it as a rough voiced clip before you decide how to shoot it.

## What previs is for

| Question | What previs answers |
| Does the scene have the right pace? | Watch it at speed with sound |
| Where do the actors stand? | Blocking across the cuts |
| Which shots do we need? | A shot list from a real run-through |
| Does the line land? | Hear it spoken |
| Will a buyer get it? | A one-minute pitch trailer |

## Storyboard, animatic, previs

| Stage | What it is |
| Storyboard | Drawings of the shots in order |
| Animatic | Storyboard panels timed to sound |
| Previs | A rough moving version of the scene, with camera and performance |

Using AI drama clips as previs and animatics · AI storyboard generator · Storyboard to video.

## Previs with generated video, honestly

Generated previs is fast and cheap enough to iterate, and it keeps the same cast and sets across shots. It is not a final cut and it does not replace performance: present it as previs. A pilot or pitch trailer · AI film production software.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Neon Nights, Urban Opulence; Costume Style Smart Casual, Streetwear; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: Draw only the people and the place of the one scene you are testing. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post.

| Tool | Output | Time |
| Storyboard | Still drawings | Fast |
| Animatic | Timed stills with sound | Moderate |
| Generated previs | Moving, voiced scene | Moderate, and iterable |

## FAQ

### What is previs?

Previsualization: a rough early version of a scene made before shooting to test pacing, blocking, camera and dialogue.

### How is previs different from a storyboard?

A storyboard shows stills; previs is a rough moving version, often with sound.

### Can AI video be used for previs?

Yes. It is fast to iterate and keeps the same cast across shots.

### Is AI previs a finished film?

No. Present it as previs: tone, cast and pacing.

### How long should a previs clip be?

Long enough to test the scene; 30 to 90 seconds is common.

### How much does it cost?

Video is 10 credits a second on the default tier plus 1 credit a second to join.

### Can I try it free?

Editing the storyboard and a 5-second preview are free on SceneMixer.

Updated November 3, 2026


---

# What Runs Underneath: How Model Tiers Work | SceneMixer

URL: https://scenemixer.com/guides/what-runs-underneath/
All three tiers — text, image and video — are yours to change; video is the one that matters. What a tier is, what the multiplier means, why we do not rank models.

# What Runs Underneath

Quick answer: A tier is a named row of configuration: which service, which model, which resolution, how long a single segment may be, and what it costs relative to a baseline. All three kinds — text, image and video — are yours to change in settings. Video is the one worth thinking about: it is where the money and the visible difference are. The text and image defaults are what the rest of the pipeline is tuned around, and most people should leave them alone.

## A tier is configuration, not a product name

Every generation goes through a tier. The tier says which service to call, which model on it, what resolution to ask for, the longest a single segment may be, and a multiplier that turns a baseline price into that tier's price per second.

All of it is configuration we can change without shipping code. That is deliberate — model services add, rename and retire models on their own schedule, and a platform that needed a release to follow them would always be a month behind.

It is also why this page does not list model names. Anything printed here would be a snapshot. The picker in the editor and the pricing page read the live configuration; they are the answer, and this page is the explanation.

## All three are yours, but video is the one that matters

Video is where the cost is, by an order of magnitude, and where the difference is visible. SceneMixer's picker is in the editor's top bar, next to the progress counter, and lists each tier's resolution and the longest segment it will take; the credit cost of the shot you are about to generate is on the generate button itself. The default is a low-resolution tier, because that is the sane place to find out whether a shot works.

The text and image tiers live in Settings, under Model Preferences, and you can change them too. We would gently suggest not to. The text model breaks your script into episodes and writes the storyboards, and a storyboard is a format the rest of the system parses literally — shot markers, durations, asset references. The image model produces the setup images every later segment references for consistency. Both defaults are the ones the pipeline's rules were written and tested against, and a mismatch there shows up as a project that behaves oddly rather than as an error message.

## What the multiplier means

One tier is the baseline at one hundred. Every other tier's number is its cost relative to that: roughly what the upstream price ratio is, so a tier that costs us twice as much costs you about twice as much. It is a ratio, not a quality score.

In SceneMixer, the number on the button and the amount actually taken come from the same function, so a quoted price is the price. There is a safety cap on the multiplier and one on segment length, so a bad configuration cannot quietly produce a bill ten times what anyone intended.

## What happens when a call fails

If a call fails in a way that looks transient — the service says it is busy, rate-limits us, or returns something retryable — it is retried once on a different account for the same model. Most transient failures are per-account, so this clears them without changing what you asked for.

Two kinds are deliberately not retried. A content-moderation refusal is a judgement, and sending it again is asking the same question twice. A parameter error will fail identically the second time. Both end the job rather than burning another call.

A retry does not change your tier. You get what you picked, or you get a failure and the credits handled as described on the failures page.

## How to choose

Use a low tier for everything until the shot is right, then re-generate the keepers on a higher one. SceneMixer keeps both takes in that segment's history, so the comparison costs you one extra generation and nothing else. One caveat if you do this: re-shoot every segment in the episode on the final tier rather than some of them, because compositing an episode whose segments have different resolutions has to re-encode instead of copying them.

The resolution is the honest headline. A higher tier is more pixels and usually a longer maximum segment; what it is not is a guarantee of a better take, because the thing that most determines whether a shot reads correctly is the text you gave it, not the tier you paid for.

## Why there is no league table here

We do not score or rank other companies' models, and you will not find a “best model” claim anywhere on SceneMixer. Scoring a competitor's product on our own storefront is not a comparison anyone should trust, and in several of the markets we operate in it is also a legal problem.

What we will publish is our own parameters — resolution, segment length, price per second — and those are on the pricing page. The only benchmark that answers your question is your own shot, run twice, which is exactly what the cheap tier is for.

| Tier kind | Where you change it | What it controls |
| Video | Editor top bar, per segment | Resolution, maximum segment length, price per second |
| Text | Settings → Model Preferences | Script breakdown, episode outlines, storyboards |
| Image | Settings → Model Preferences | Character sheets, location plates, prop and state images |

## FAQ

### Can I choose the text or image model?

Yes, in SceneMixer's Settings under Model Preferences. The defaults are what the rest of the pipeline was written and tested against — the storyboard is a format later steps read literally, and the setup images are what every later segment references — so most people are better off leaving them alone.

### Why does the page not name the models?

Tiers are live configuration we change as services add and retire models. Any name printed in a static page would be a snapshot. SceneMixer's picker in the editor and the pricing page read the current configuration.

### Is the most expensive tier the best one?

It is more pixels and often a longer maximum segment. Whether a shot reads correctly is decided far more by the storyboard text than by the tier.

### If a model fails, do you fall back to a different one?

A transient failure is retried once on a different account for the same model. Moderation refusals and parameter errors are not retried. SceneMixer does not silently switch you to something you did not pick.

### Which companies' models are these?

The privacy policy is where the processing disclosure lives. SceneMixer does not publish rankings or comparisons of third-party models on this site.

### Does a higher tier generate faster?

Not reliably either way — run time depends on the model and on how busy it is, not on the price. What a cheap tier reliably saves you is credits, which is the reason to iterate on one.

Updated November 3, 2026


---

# What Support Can See and Do in Your Account: Read-Only, Logged

URL: https://scenemixer.com/guides/what-support-can-see-and-do-in-your-account/
Support can open your account in a read-only view to help with a report. What it shows, what it cannot do (spend, change settings, generate), how it is logged.

# When Support Looks at Your Account: What They Can See, What They Cannot Do

Quick answer: When you report a problem, support may open your account to see what you see: your projects, storyboards, sheets, clips, credit history and orders. That view is read-only by construction: it cannot start a generation, spend or move credits, change any setting, change your language, delete anything, or mark guides as seen. Every request made in that mode is logged with the staff member’s identity. Nothing you have set changes because someone looked. Most reports need no look at all: a screenshot of the card or message, the project name and the time are usually enough.

Use the Ask window’s feedback with a screenshot: the report carries your account id and the page you were on, which is what support needs first.

## What support sees

The same pages you see, as you see them: home, project wizard, episode editors, library, billing, settings. Storyboards and descriptions are readable so a problem in the text can be found; clips can be played. Passwords are never visible to anyone, including staff.

## What the read-only view cannot do

Start or stop generations, buy, refund or move credits, change settings (badge switch, interface language, watermark choice), edit or delete projects, sheets or storyboards, change your dialogue or reading language, or dismiss the onboarding tour. SceneMixer's read-only mode refuses those actions at the server; a banner in the view says read-only. Refunds and credit adjustments are done from the back office with a separate audit record, never by acting inside your account.

## How it is logged

Requests made while viewing your account carry the staff member’s id and are written to the request log; credit changes made from the back office are recorded with a reason; in your credit history they show as Added by support or Adjusted by support, or as the task’s own line turning Refunded, and support can tell you the reason. Reading Your Credit History.

## What to send so nobody needs to look

The exact message or the card showing the state, the project and episode name, the approximate time, and what you expected. The feedback form in SceneMixer’s Ask window attaches screenshots and your account id automatically. How to Get Help.

## Privacy

Your scripts and clips are not used to train models and are not shown to other users; the read-only view exists only to resolve your report. Details in the privacy policy linked in the footer.

| Action | Support in read-only view | Back office |
| Read projects, storyboards, clips | Yes | Yes |
| Start or stop a generation | No | No |
| Spend, buy or move credits | No | Adjust with audit record |
| Change your settings or language | No | No |
| Delete anything | No | No |
| Refund an order | No | Yes, logged |

## FAQ

### Can support generate a clip to test on my account?

No. SceneMixer's read-only view cannot start generations; tests are run on staff accounts.

### Did someone change my language?

Not from the read-only view; it cannot. Check SceneMixer's top-bar language menu on this browser.

### Can support see my password?

No one can; SceneMixer does not store passwords in readable form.

### Will I be told when my account was viewed?

There is no notification; SceneMixer's log exists and is used for audit.

### Can I refuse?

Say so in your report and SceneMixer's support will work from screenshots only.

Updated November 3, 2026


---

# What the System Keeps Track of From One Segment to the Next

URL: https://scenemixer.com/guides/what-the-system-keeps-track-of-between-segments/
In newer storyboards a segment starts from a running record of the story, not the clip before it: positions, posture, what is held, wounds, wet clothes, time and weather.

# What the system keeps track of from one segment to the next

Quick answer: Every segment is still generated on its own, from the character, scene and prop images and its own storyboard text; the clip before it is not shown to the video. What carries the story across the join, in storyboards written or regenerated since late October 2026, is a running record the site keeps while it reads your text: who is in which scene, whether they sit, stand, lie, kneel or crouch, what is in their hands, which look they wear, wounds and mud on their body, whether they are wet or unconscious; and for each scene the time of day, weather, light, doors and windows, and whether a crowd is there. Each new segment starts from that record. You change it by writing the change in a shot, in the shot where it happens.

Not sure which kind of storyboard you have? The segment header, field by field shows both layouts. Older storyboards work differently: why the next segment opens like the last one ended.

## What is in the record

| About | What is remembered |
| Each person | Which scene they are in · sitting, standing, lying, kneeling or crouching · riding something · what they hold · which look they wear (a state image, if you made one) · marks on the body (blood, bruises, scratches, burns, mud, dust, sweat stains) and where on the body · wet or dry · dead or alive · unconscious or not |
| Each scene | Time of day · weather · the light · whether doors and windows are open or shut · which look the room is in (for example after damage) · whether a crowd of passers-by is there |
| Each object | Who is holding it, or which scene it is in · which look it has |
| Who is next to whom | When two people stand together and one walks away, the other stays where they were |

SceneMixer's record keeps what is still true when the shot ends, not what happens during it. A smile, a flinch or a half-second reach for a glass belongs to that shot’s text and does not carry over.

## How a change gets into the record

Write it as a plain event in the shot where it happens, as you would in a script. The record changes the moment the text says so, and stays changed until the text changes it back.

| You write, in segment 3 | Segment 4 starts with |
| @Mara takes off her coat and lays it over the back of the chair. | Mara without the coat; the coat on the chair |
| @Leo is hit across the face; blood runs from his left eyebrow. | A cut on Leo’s left eyebrow, still there |
| @Leo washes his face at the sink. | The blood washed away |
| @Mara slumps to the floor, unconscious. | Mara lying on the floor, unconscious |

Reverse the change the same way: she puts the coat back on, the cut has healed, he wakes. If the text contradicts the record — a character who died walks into a later shot — the text wins and the record follows it, so a mistake in your wording does not break the segment.

## What you can see and change in the editor

- Time · Weather on the segment header, with Change, applies to the whole run of segments in that scene. Segment header.
- SceneMixer's In frame line under the text shows who each shot will show. Notes under a segment.
- State images for looks that last: a torn coat for the next five segments is a state image, not a sentence you repeat. Costume changes and scene states.
- Video edited appears when you trim a clip or change its visuals; later segments then continue from the edited clip. Edit this clip.

## What it does not do

- It does not show the video the previous clip. There is no last frame attached to the next segment, which is why you can generate segments in any order and why regenerating one never changes the next.
- It does not invent what you did not write. If someone is hurt between segments and no text says so, the record has no wound. Put it in the first shot of the segment where it should show.
- It does not decide framing. Whether a segment opens on a close-up or a wide shot is in the first shot’s text.
- It keeps the record per character. Say who the change happens to: use the character’s name or the @ chip in the shot.

## When something does not carry over

- Find the shot where the change should have happened. Is it written as an event (she is cut on the cheek) or only implied (her cheek looks bad)? Write the event.
- Read the first shot of the segment that lost it. A line there that says otherwise (she walks in, unmarked) overrides the record.
- If the look should last for many segments, make it a state image and pick it for those segments.
- Save, and generate the segment again. Editing the text is free; SceneMixer re-reads the record from the new text.

Segments you generated before the change keep the video they have; only segments you generate afterwards use the corrected record. When a shot comes out wrong · Shot-to-shot continuity.

## FAQ

### Does the next segment see the clip before it?

No. SceneMixer starts each segment from the images and its own storyboard text, plus the running record of the story. There is no last frame attached.

### What exactly does the system remember?

SceneMixer tracks where people are, whether they sit, stand, lie, kneel or crouch, what they hold, which look they wear, wounds and mud, wet or dry, unconscious or not; for scenes, time, weather, light, doors and windows and crowds.

### My character's bandage disappeared in the next segment.

Write the injury as an event in the shot where it happens, and check the first shot of the next segment does not say otherwise. For a look that lasts, use a state image.

### Does it cost credits to update the record?

No. Editing the text on SceneMixer is free; you pay only for the video you generate.

### Why does a character stay where they were when I did not move them?

SceneMixer's record keeps people where the text left them until the text moves them. Write the move in the shot where it happens.

### Is this the same for ads and remakes?

No. They use the older way on SceneMixer, where each segment starts from its own text and images.

### Do older storyboards work this way?

Not unless you regenerate them. SceneMixer storyboards written before late October 2026 keep the earlier behaviour.

### Can I see the record?

SceneMixer does not show it as a list. You see its effects in the Time and Weather line, the In frame note and the images each segment uses.

Updated November 3, 2026


---

# What You Can Feed In: Text, Images, Video | SceneMixer

URL: https://scenemixer.com/guides/what-you-can-feed-in/
Three ways in: paste a script, upload product photos, or upload a video to remake. What each accepts, and the size limits on all three.

# What You Can Feed In

Quick answer: Three doors, one pipeline behind them. Paste or upload a script and it becomes a drama. Upload up to four photos and it becomes an ad. Upload a video and we read it, then build a new drama from what we read. Text has a length floor and a length ceiling; images are JPEG or PNG; video is MP4 or MOV within the limits shown on the upload box.

The sample projects were all made through the first door — a block of text and nothing else.

## Text: the usual way in

Paste it, or drop a file. We read .txt, .docx, .md and .markdown. A .docx is unpacked to plain text in your browser before anything is sent, so formatting, images and comments inside the document are dropped rather than uploaded.

There is a floor: about a paragraph — roughly 330 characters of English prose, or a hundred Chinese characters. Below that there is nothing to break into episodes. If you paste only a link, you will get a specific message saying so — we do not fetch URLs, and a page of ours reading someone else's page is not something you want in the middle of your billing anyway. Paste the text itself.

There is also a ceiling, and it is a hard stop rather than a silent truncation. Long books are handled by reading them in two passes, which raises the ceiling a lot, and we apply whichever ceiling matches how your text will be read.

We refuse rather than truncate because truncation does not announce itself. The failure we measured on over-long input is a script that comes back looking complete and is not — a fifty-episode book parsed into ten episodes, with no error anywhere. A refusal in the first second is a better outcome than a plausible-looking result you only catch three steps later.

Any language. The interface is separate from the language your production documents come out in, and separate again from the language your characters speak — there is a dedicated guide on that.

## Images: reference material and ad photos

Pick whatever your camera or design tool gave you — WebP, HEIC, PNG, JPEG. Your browser decodes it, flattens any transparency onto white and re-encodes it as JPEG before a single byte goes up. One path, no exceptions, including for files that were already fine.

That is why what arrives is always JPEG or PNG: those are the two formats every step downstream accepts, and converting at the door means the server only has to recognise one shape of bytes rather than trust a file extension. Renaming something to .jpg will not get it through.

The only refusals are files your own browser cannot open — HEIC outside Safari, SVG, camera RAW — and you find that out the moment you pick the file, not three steps later.

For an ad on SceneMixer, you can attach up to four photos of the same subject — different angles, a detail, the thing in use. Each up to ten megabytes, sixteen in total. They are read once, for free, before anything is charged, and the read tells you what it thinks the product is so you can correct it before spending.

## Video: what remaking does and does not do

MP4 or MOV, up to three hundred megabytes, and no longer than the length shown on the upload box. The upload is read: we cut it into shots and describe the people, places and props in it.

What happens next is the part to be precise about. The source video is never handed to the video model. It is used to understand, and what comes out is generated from that understanding — new characters from your setup images, new framing, new lines. The result will not look like the original, and it is not meant to.

Which means the honest use is your own footage, or footage you hold the rights to: a rough cut you want rebuilt, a reference you shot yourself, something you have permission to work from. We do not review uploads before they are processed, so the rights question is yours, and our takedown process applies to generated files and share links the same as anything else.

Reading the video is charged, because it is a real model call over every shot we cut out. The rest of the project is charged step by step afterwards, the same as any drama.

## Confirming your rights

The first time you upload a file, a window titled Before you upload asks you to confirm that you have the right to use the material (scripts, images, video, audio) and that every real person whose face or voice appears in it has given consent. Press Confirm and upload to go on; you only need to confirm once, several files in one go ask once, and cancelling stops the upload.

## What the three doors have in common

Behind all three, SceneMixer runs one pipeline: read the material, produce characters, locations and props, break it into episodes, write a storyboard for each, shoot each segment, composite. Ads and remakes simply have something pressing the next button for you.

That is why the editor looks identical whichever door you came in by, and why a guide written about one of them is usually true of the others.

| Input | Accepted | Limits |
| Script text | Pasted, or .txt / .docx / .md / .markdown | At least ~330 characters of English (or ~100 Chinese); a hard ceiling, no truncation |
| Web link | Not accepted | Paste the text instead |
| Reference image | Anything your browser can decode | Converted to JPEG on the way up; checked by bytes |
| Ad photos | Up to 4 of the same subject | 10 MB each, 16 MB total; read free before you pay |
| Source video | MP4, MOV | 300 MB; length as shown on the upload box |

## FAQ

### Can I give you a link to the story instead of the text?

No. SceneMixer does not fetch URLs, and pasting one gets you a message saying so rather than a confusing failure later. Copy the text in.

### My novel is longer than the limit. What now?

SceneMixer reads long text in two passes, which raises the ceiling substantially, and we apply the ceiling that matches how yours will be read. Past that, split it — the limit exists because a read that runs past its budget comes back looking complete while quietly having dropped most of the story.

### Do I have to convert my images first?

No. Your browser converts whatever you pick to JPEG before uploading, so WebP, HEIC and PNG are all fine. The only files SceneMixer refuses are ones the browser itself cannot decode, and you are told at the moment you pick them.

### Does remaking a video copy the original?

No. The source is read, not reshot. SceneMixer never sends the source to the video model. What comes out is generated from the description — your characters, new framing, rewritten lines.

### Can I remake a show I like?

Upload only material you own or are authorised to use. SceneMixer does not screen uploads before processing them, and the responsibility for rights in what you upload sits with you.

### What is the ad price actually covering?

On SceneMixer, the figure on the button is a cap, not a deposit: it is the most the whole run can cost, and you are charged for the steps that actually happen. For a remake, only the read is taken up front.

Updated November 3, 2026


---

# What SceneMixer Gives You for Free | SceneMixer

URL: https://scenemixer.com/guides/what-you-get-for-free/
Welcome credits, setup images that belong to your account not one project, a free storyboard on your first drama, and one free preview. What is covered and what is not.

# What You Get for Free

Quick answer: A new account starts with 50 credits. On top of that: 15 free setup images that belong to your account — not to one project — and can be spent on any drama you have had read, on the default image model (Qwen-Image 3.0); the storyboard for episode 1 of your first drama; one free promo cover (a batch of three posters) on that same drama; and one free preview clip of up to 5 seconds. Reading a script, writing one with AI and generating video are charged normally.

If you would rather see the output before spending anything, the sample write-ups show finished episodes with their settings and their storyboards.

## Setup images — 15, and they follow your account

These are the character sheets, location plates and prop shots that everything downstream references. The allowance is per account, not per project: it is one pool of 15 across characters, locations and props, and whatever you do not use on your first drama is still there for the next one.

Two conditions. It can only be spent on a project whose script has actually been read and paid for — that stops an empty project being used to farm free images. And it does not apply to ad projects or to remakes, because neither of them pays a script-read fee.

## Storyboard — episode 1 of your first drama

This one really is tied to the first drama you ever create, and it is decided the moment that project is created. Regenerating that episode's storyboard, and every later episode, is charged on SceneMixer.

The check looks at whether you have created a project before, including ones you have since deleted, so there is exactly one first drama per account. It also declines when another account registered from the same device or recently from the same address has already claimed it.

## One free promo cover

Your first drama also gets one free batch of promo posters: three candidates for its cover, drawn from your lead characters. On accounts created on or after October 23, 2026, the three are made for you once your lead characters’ sheets are ready and you move on from the characters in step 2, and they wait in Cover art; on older accounts you start the batch yourself. Later batches — a new title, other characters, another mood — are charged like any image. The conditions are the same as for the free setup images: the script has been read and paid for, and it is not an ad project. Promo cover posters.

## One free preview

Each SceneMixer account gets one free preview clip of up to 5 seconds — the opening moment of a segment you wrote, rendered for real. It is there so you can see what the pipeline does with your own material before paying for a full segment.

One per account, not one per project.

## What is not free

Reading a script is charged on every project. So is writing one with AI. Video is charged from the first second: there is no free video allowance.

State images have their own separate small allowance, counted independently of the setup images — using one up does not eat into the other.

## Used is used

The allowance is counted by what you generated, not by what currently exists. Deleting an image and making it again does not give the allowance back. Generations that fail, or that you cancel before they start, do not consume it.

Where anything is unclear — a count that will not resolve, a missing setting — SceneMixer charges rather than giving it away.

## If you run out

Every button shows its price before you press it, and the balance is in the top bar. Credit packs start at $1.49, and membership adds a monthly credit allowance plus a discount on credit purchases, starting at $7.99 a month.

| Item | Scope |
| Welcome credits | 50 on the account |
| Setup images | 15 free — per account, any drama that has been read |
| Storyboard | Episode 1 of your first drama |
| Promo cover | One batch of three posters, on your first drama |
| Preview clip | One per account, up to 5 seconds |
| Video | Charged from the first second |
| Script read / AI scriptwriting | Charged on every project |

## FAQ

### Do the free setup images expire when I start a second project?

No. They belong to your SceneMixer account, so whatever you did not use on the first drama is still available on the next one.

### Can I use them on an ad or a remake?

No. Neither pays a script-read fee, and SceneMixer only lets the allowance be spent on a project whose script has been read and paid for.

### If I delete my first drama, does the storyboard freebie reset?

No. SceneMixer still counts deleted projects, so the next one you create is not a first drama.

### Does a failed generation use up the allowance?

No. SceneMixer releases what failures claimed, and so do cancellations of work that never started.

### Is there any free video?

Only the single preview clip of up to 5 seconds. Full segments are charged from the first second on SceneMixer.

Updated November 3, 2026


---

# What You Get to Keep: Rights and Labels | SceneMixer

URL: https://scenemixer.com/guides/what-you-get-to-keep/
What the terms grant you over what you generate, the two limits attached, how AI video is labelled, and what happens if someone complains.

# What You Get to Keep

Quick answer: Under the terms, what you generate is yours to the extent the law allows, royalty-free, commercial use included — and that covers output made with free credits. Two limits come attached: we cannot promise it is registrable as a copyrighted work, and it is not exclusive to you. Whether finished video carries the visible AI badge we add depends on your plan and your settings.

## What the terms grant

The script or novel you upload stays yours. What comes out the other end is yours to use, including commercially, provided you are within the terms and have paid for it — and output produced with a free allowance counts as paid for. We do not take a cut of anything you make with it and do not charge again at publication.

The exact wording is “to the extent permitted by law”, and that qualifier is doing real work. Whether a machine-generated image or clip attracts copyright at all is unsettled and differs by country. We are not in a position to decide that for you, so we do not claim to.

The same section also has you grant us a licence, and its shape is worth knowing: to run the service (storing, processing, passing material to the model services, showing it back to you), to do content-safety review, and to improve quality on de-identified data. Separately, anything you publish publicly, submit as a sample or enter into a campaign is treated as licensed for showcase and promotion.

## The two limits, stated plainly

SceneMixer cannot promise it is registrable. Several jurisdictions refuse registration for work without sufficient human authorship. If registration matters for what you are making, take advice before you build a business on it.

It is not exclusive. Generation is probabilistic, but it is not infinitely divergent: a similar prompt from someone else can produce something recognisably similar. You get the right to use what you made. You do not get a promise that nobody else will ever hold something like it.

## What we do not promise

We do not warrant that what comes out infringes nobody's rights, and we do not indemnify you if somebody says it does. Responsibility for what you generate — and for what you fed in to generate it — sits with you. SceneMixer's liability is capped in the terms, and the cap is small: it is measured against what you have paid us.

Commercial use is also subject to the terms of the model services underneath. Those are third-party agreements we do not control, and they can change.

Nor do we guarantee any single generation. Results are probabilistic by nature; the terms say so, and the honest version of this product is one where two identical requests can give two different clips.

## How finished video is labelled

We label video at the moment it is delivered: a small badge is burned into the picture in the corner, and a line of AI-generated-content metadata is written into the file. The badge is part of the picture, not an overlay the player draws — it survives being downloaded and re-uploaded.

Those two are ours and they are written in the same pass, so they travel together: without a paid plan you always get both, and on a paid plan we add neither by default — you can switch them back on from settings or from the control next to the model picker in the editor. Where a visible mark is required by the law that applies to you, switch it on.

Marking that the model services themselves apply when they generate is a separate layer and is not something this switch touches. Section 05 of the terms is the rule that governs all of it, not this page.

Independently of any of that, the platform you publish on will have its own disclosure requirement, and meeting it is your job. A badge in the corner is not the same thing as ticking the “AI-generated” box on the platform you upload to.

## If somebody complains

SceneMixer has a notice-and-takedown route, and it explicitly covers generated files and share links, not just uploads. A complaint about something you made here is handled through it.

Two things make complaints much less likely, and both happen before you generate rather than after. Do not feed in material you have no right to use. And do not reproduce a real person's face or voice without documented consent — the terms prohibit that outright, and the legal risk for a likeness sits with you.

## A short pre-publication check

Did you have the right to everything you uploaded? Does the platform you are posting to require an AI disclosure, and have you made it? Is the label on, if your jurisdiction expects it? Do you have your own copy of the file, given that deleting on SceneMixer is one-way?

Four questions only you can answer.

## FAQ

### Do I own the copyright in what I generate?

SceneMixer's terms grant you the output to the extent the law allows, and that includes commercial use. Whether it attracts copyright at all depends on your jurisdiction, and we do not make that claim for you.

### Can I use output made with free credits commercially?

Yes. On SceneMixer, output produced with a free allowance is treated the same as any other output.

### Could somebody else end up with the same clip?

Something recognisably similar, yes. Output is not exclusive to you, and SceneMixer's terms say so.

### Will you defend me if someone claims infringement?

No. Responsibility for what you generate and what you fed in sits with you, and SceneMixer's liability is capped in the terms. There is a takedown route for complaints.

### Will my video carry the AI mark?

The badge and metadata SceneMixer adds: always without a paid plan, and off by default on one — you can switch them back on. Marking the model services apply at generation time is a separate layer that switch does not touch. What is permitted is governed by section 05 of the terms and by the law that applies to you.

### Is the label enough to satisfy a platform's AI disclosure rule?

No. Treat them as separate obligations. The platform you upload to has its own requirement and meeting it is your responsibility.

Updated November 3, 2026


---

# What Is Inside a Project | SceneMixer

URL: https://scenemixer.com/guides/whats-inside-a-project/
A project holds a cast and episodes; an episode holds segments; a segment holds takes. Which level an edit lands on, and what a rebuild loses.

# What Is Inside a Project

Quick answer: A project holds two things: a cast — characters, locations and props, with optional state images — and a list of episodes. An episode holds segments. A segment holds every take you have generated of it. Almost every confusing moment in the editor is an edit landing on a different level from the one you meant.

Open a finished sample episode and the four levels are all on one screen.

## The five levels

Project. The settings, the cast, and the episodes. Everything below inherits from it.

Cast. Characters, locations and props, each with a description and a generated image. A character may also have state images — the same person after a change that a sentence cannot fix, such as a full change of costume or visible injury. Locations and props can have them too: a hall before and after it burns, a blade sheathed and drawn.

Episode. A numbered part of the story with its own storyboard.

Segment. One generation. Fifteen seconds or so of finished video, with the shot text that produced it.

Take. Every time you generate a segment, the result is kept. One of them is the current one; SceneMixer makes switching between them free and instant.

## Which level an edit lands on

This is the thing worth internalising, because the same-looking edit behaves differently depending on where you make it.

Editing a character in the project's cast changes them everywhere, in every episode that has not been generated yet. Editing the same character from inside an episode's asset panel changes them in that episode only — useful when one episode needs a different look and the rest should not follow.

And picking a state image for one segment changes that segment only. Three scopes, three different blast radii.

Renaming is handled for you. When you rename someone on SceneMixer, every reference to them in every storyboard is rewritten in the same transaction, so the text does not quietly stop matching. No two assets in a project can share a name — not two characters, and not a character and a prop — which is what makes that rewrite unambiguous. Projects created before September 2026 may still contain duplicates; in those, a rename updates the cast entry and leaves the storyboard text alone, because there is no safe way to tell which one a reference meant.

## Which image a segment actually uses

Up to five things can claim the right to supply the picture for a given character in a given segment. They resolve in a fixed order, most specific first:

A state you picked for this segment wins. Failing that, an image you overrode for this episode. Failing that, a state the system assigned to this segment on its own when it read the storyboard — the editor marks those “auto”, and it is the one people hunt for, because nobody chose it. Failing that, a state set for the episode as a whole. Failing all four, the base image from the cast.

SceneMixer counts only states that have actually been generated. If a state exists as a description but its image was never made, it is skipped and the next rule down applies — the system falls back to something real rather than sending nothing. If a segment is using a face you did not expect, walking down that list will find where it came from.

## What a rebuild throws away

Rebuilding an episode's storyboard deletes that episode's segments and writes new ones. Anything attached to a segment goes with it: the finished video, its take history, and any per-segment state choices you made by hand.

The episode level survives a rebuild. Your cast survives, your per-episode overrides survive, and the other episodes are untouched. But the per-segment choices are re-derived from scratch, which for most projects looks like nothing happened — and for a project where you hand-picked states shot by shot, looks like the work vanished.

Re-reading the whole script is the level above that: it replaces the episodes that have no storyboard yet. Episodes you have already worked on are left alone.

## Deleting

Deleting anything removes it from the project immediately. Deleting a project, an episode or a cast member also queues its files for erasure. SceneMixer has no restore anywhere you can reach.

Deletion follows the structure: removing an episode removes its segments and their takes; removing a character removes its state images. Before a file is actually erased we check whether anything else still points at it — a character copied into a second project shares the same image — and if something does, the file stays.

| Level | Contains | An edit here affects |
| Project | Settings, cast, episodes | Everything not yet generated |
| Cast member | Description, image, state images | Every episode |
| Episode | Storyboard, segments, per-episode asset choices | That episode |
| Segment | Shot text, takes, per-segment state choices | That segment |
| Take | One generated video | Nothing else — switching is free |

## FAQ

### I changed a character and only one episode updated. Why?

You edited them from inside that episode's asset panel, which is scoped to that episode. Editing the same character in the project's cast on SceneMixer changes them everywhere that has not been generated yet.

### A segment is using the wrong picture of someone.

Five things can supply it, resolving most specific first: a state picked for that segment, an image overridden for that episode, a state the system assigned automatically, a state set for the whole episode, then the base image. Walking down that list finds the source — the automatic one is the usual answer.

### Does regenerating an episode's storyboard affect other episodes?

No. SceneMixer rebuilds only that episode's segments. Your cast and the other episodes are untouched — but everything attached to the discarded segments goes with them, including finished videos.

### Can I reorder segments?

Segments are numbered within their episode and SceneMixer maintains the numbering for you when you insert or delete one. They are not free-floating cards you can shuffle.

### If I rename a character, do the storyboards break?

No. Every reference is rewritten in the same operation, and SceneMixer does not let two assets in a project share a name, which is what keeps it unambiguous. The exception is a project made before September 2026 that still has duplicate names — there the cast entry is renamed and the storyboard text is left as it was.

### Are all my takes kept forever?

On SceneMixer, every take of a segment is kept and you can switch between them for free, for as long as the segment exists. Deleting the segment, the episode or the project removes them.

Updated November 3, 2026


---

# When an AI Video Generation Fails: Which Ones Refund | SceneMixer

URL: https://scenemixer.com/guides/when-a-generation-fails/
A failed segment shows one message, and every message falls into three families: credits returned, on hold for review, or blocked by content rules. What each asks of you.

# When a Generation Fails: What the Message Means and What Comes Back

Quick answer: A failed generation never keeps your credits quietly. The card says Failed and the preview pane shows one sentence that tells you which of three things happened: the credits were returned and you can retry; the credits are on hold while support reviews what went wrong, and you should not resubmit; or the clip was blocked by content rules and you need to change the text. Read the sentence, do what it says, and press Retry or Regenerate when it makes sense.

Open a finished episode to see where a segment's status and message appear.

## Where to look

The segment’s card on the timeline says Failed. Select it and the preview pane shows the message, with Retry where retrying makes sense. Your storyboard text is untouched; so is any earlier version of the clip.

## Family one: the credits came back

Messages like The generation service hit a temporary problem. You weren’t charged, or The task was interrupted by a service update. No credits were used, or support has refunded these credits. SceneMixer puts the held credits back in your balance. Nothing to fix on your side: press Retry, or wait a little if the message says the service is busy.

## Family two: on hold for review

The task was interrupted. These credits are on hold pending a support review — no need to resubmit. Something went wrong after the clip was requested and it is not yet clear whether it was delivered. Support looks at every one of these and either releases the credits or delivers the clip. You do not have to write in, and you should not generate the same segment again in the meantime.

## Family three: blocked by content rules

Blocked by content safety rules — edit the script or description and retry, or The provider’s IP check flagged … as resembling a known franchise or brand. The clip was refused, not lost. Change the shot — the words, the character design, the visual style — and generate again. On SceneMixer, repeated blocks in a short time pause video generation for a while. What to change when a clip is blocked.

## Messages that ask you to retry later

The service timed out, busy (queued / rate-limited), temporarily unavailable, or a platform-side configuration issue that we have been alerted to. These are on our side or the provider’s, not yours. Wait a few minutes and retry; if the message said you were not charged, you were not. Invalid input — check the reference images and duration is the one in this group that is about the segment: an image that could not be used, or a length past the tier’s cap.

## If you think it is wrong

Send feedback from the sidebar with a screenshot of the message and the segment number; a reply arrives in SceneMixer's bell and, if you chose it, by email. How to get help.

| The message says | Your credits | Do |
| You weren’t charged / No credits were used / support has refunded | Back in your balance | Retry |
| On hold pending a support review | Held; released or delivered after review | Wait; do not resubmit |
| Blocked by content safety rules / IP check flagged | Per the message | Change the text or design, then regenerate |
| Timed out / busy / temporarily unavailable | Per the message | Retry in a few minutes |
| Invalid input | Per the message | Check images and the segment’s length, then retry |
| The generated soundtrack / image was rejected by the provider’s content check | Not charged | Retry; if it repeats, change the dialogue or the description |
| Generation failed. Please retry | See your credit history: Refunded on the task’s line means they came back | Retry; send feedback with the segment number if it repeats |

## FAQ

### A segment failed. Did I lose the credits?

No. SceneMixer returned them, or they are on hold while support reviews the failure. The message says which.

### What does 'on hold pending a support review' mean?

The clip may or may not have been delivered, and SceneMixer staff check. Do not resubmit; the credits are released or the clip arrives.

### Can I just press Retry?

Yes, unless the message on SceneMixer says the clip was blocked — then change the text first — or says credits are on hold — then wait.

### Does a failure change my storyboard?

No. SceneMixer leaves the text and any earlier versions of the clip untouched.

### The same segment keeps failing.

If it is blocked, change the shot. If it is a service message, wait and retry; if it persists, send feedback with a screenshot.

Updated November 3, 2026


---

# When a Shot Comes Out Wrong | SceneMixer

URL: https://scenemixer.com/guides/when-a-shot-comes-out-wrong/
A symptom index for bad takes: what to change, what each fix costs, which button destroys paid work, and when a failure gets your credits back automatically.

# When a Shot Comes Out Wrong

Quick answer: Editing the storyboard text is free. Regenerating the video charges again, every time. Switching back to an earlier take is free and instant. And one button — regenerating an episode's storyboard — deletes every finished video in that episode without a refund, so read the first section before you touch it.

The sample episode is worth opening alongside this page: the shot text, the take history and the regenerate controls are all visible on it.

## First: the one button that destroys paid work

Regenerating an episode's storyboard rebuilds that episode from scratch, and every video you have already paid to generate in it is discarded with the old segments. The credits are not returned — they were spent on clips that were delivered. The editor makes you confirm, and refuses outright if a video in that episode is still generating.

Regenerating one segment's video is a different thing and is safe: the old take stays in that segment's history and switching back is free. The mechanics of both are covered in fixing one shot without redoing everything. This page is about what to change when a take comes out wrong, and what each change costs.

## What is free and what costs

Free: editing any of the storyboard text, changing which setup image a segment uses, switching between takes you have already generated, previewing, and re-reading anything you already have.

Charged: generating video. On SceneMixer, each attempt is a separate model call and is billed as one. There is no “try again free” — that is the honest shape of a service that pays per call upstream.

Which is why the order matters. Fix the text first, look at it, and only then spend. A cheap tier is a reasonable place to find out whether a shot reads before you spend a good one on it.

## A symptom index

Most bad takes trace back to something in the shot text rather than to the model having a bad day. The table below is the short version; each fix is an edit to the segment, which costs nothing until you press generate again.

## An action or prop you wrote did not appear

The video draws what a shot shows, and small things slip. Make the action something the camera can see: give it its own shot, frame it close — a close-up on her hand turning the key — and give it a second or two. Name the prop with @ in that shot and keep it on the Props line. Something worn on the body, such as a ring, a badge or an earpiece, reads far better in a close shot than in a wide one. A mechanism, such as a lock opening or a case being loaded, reads best as one clear motion rather than a list of steps. Then regenerate that segment. Props and locations in a shot.

When the clip is right apart from one thing in the picture — a colour, the weather, an expression — SceneMixer's Edit this clip under the player may fix it without regenerating; prices are on the pricing page. Edit this clip.

## What the Stop button does

Stop is not cancel-the-order. If the job has already been picked up and sent upstream, stopping it releases the interface but the credits are settled as spent — the call was made and we were billed for it.

If the job has not been picked up yet — still queued, or parked because your slots were full — stopping it does refund, in full, because nothing was ever sent.

A segment starts saying “generating” the moment you submit it, which is before anything has been sent — so a segment showing that is often still fully refundable, and there is no way to tell from the card. If you have changed your mind, press stop and find out. Once it really has gone upstream, stopping costs the same as letting it finish, and letting it finish at least gives you a take.

## When a failure refunds itself

Automatic: the model never received a usable request; the model answered that it was busy or rate-limited; a clip was refused by content moderation.

Moderation refunds have brakes on them, because an unlimited auto-refund on refused content is an invitation to generate refused content all day. SceneMixer has a per-account limit within a 24-hour window, a site-wide daily ceiling, and a master switch. The part worth knowing: crossing the per-account limit does not just stop the refund — it suspends your video generation for an hour, and the counter does not reset when the suspension lifts, so each further rejection suspends you again.

Not automatic: a timeout or a dropped connection mid-call. From our side those look identical whether the model ran or not, so the credits stay frozen — neither charged nor returned — and a person settles it. The same applies if the model produced a clip but we failed to deliver it to you: the amount stays frozen rather than being charged, because you did not receive anything.

In all of those cases the task’s line in your credit history says which of the three happened: Refunded, no tag (charged), or Under review.

## When to stop editing and just reshoot

Generation is probabilistic. The same text can give a good take and a bad one, and after two or three attempts with no change in the text you are sampling, not fixing.

The signal that editing will help is that you can point at a sentence and say what it is telling the model to do wrong. If you cannot, the text is probably fine and the shot is a dice roll — in which case the cheapest move is to generate again, keep the better take, and move on. Both takes stay in the history either way.

| Symptom | Usually caused by | Fix | Cost |
| A person appears who should not be in the shot | The shot's in-frame list includes them — that list, not the prose, decides who gets a face | Edit the shot so they are described as off-screen, or remove the mention | Free until you regenerate |
| The wrong character's face | Two characters described too similarly, or a missing setup image | Give each a distinct look in the asset library, generate the setup image | Setup image is charged |
| Wrong outfit or wrong state (before/after an event) | The segment is using the base image rather than a state image | Generate the state image and pick it for that segment | State image is charged |
| Framing repeats the previous segment | The previous segment's last frame is attached as a reference for continuity | Edit the previous segment's text — that unhooks its frame. Regenerating alone re-uses it | Editing is free; the regeneration is charged |
| Background crowd appears and disappears between segments | Only some segments declare background people | Add or remove the background-people line consistently across those segments | Free until you regenerate |
| Segment is longer or shorter than expected | The shot durations in the text sum to the segment length | Edit the per-shot seconds | Free until you regenerate |
| Lines are spoken by the wrong voice | The voice tag on the line | Edit the tag, or reassign the voice | Free |

## FAQ

### Does regenerating a segment delete the old one?

No. SceneMixer keeps each take in that segment's history, and you can switch back to any of them for free. The destructive action is regenerating the whole episode's storyboard, which discards every segment in it.

### Is editing the storyboard free?

Yes. Text edits, image choices and version switches cost nothing. Only generating video on SceneMixer is charged.

### I pressed stop. Do I get my credits back?

Only if the job had not started yet. Once it has been sent upstream, stopping on SceneMixer releases the interface but the credits are settled as spent, because the call was made.

### A generation failed. Where did my credits go?

Three possibilities: refunded automatically (the model never ran, was busy, or refused on moderation), charged (it ran and delivered), or still frozen pending a manual check (a timeout, or a delivery failure on our side). The tag on the task’s line in your SceneMixer credit history says which.

### Why can I not regenerate the storyboard while a video is running?

Because it would discard the segment you are currently paying to generate. SceneMixer has no override for that one.

### How many times should I regenerate before giving up on a shot?

If you cannot point at the sentence causing the problem, two or three. Past that you are sampling randomness rather than fixing anything, and a cheaper tier or a rewritten shot is a better use of the credits.

Updated November 3, 2026


---

# When a Model Tier Disappears: What Happens Next | SceneMixer

URL: https://scenemixer.com/guides/when-a-tier-disappears-fallback/
What happens on SceneMixer when a tier you selected is withdrawn or unavailable: the fallback to the default, the message you may see, and why the price stays honest.

# When a model tier disappears

Quick answer: If the tier saved in your preferences no longer exists, text and image work falls back to the default tier; in the episode editor, video picks the tier this episode last used if it is still listed, otherwise the default. Either way, the price on the generate button shows the tier that will actually run. A tier that is temporarily unavailable is hidden from the tier list rather than offered and refused. If the editor was open when a video tier went away, you may see The selected video tier no longer exists — refresh and pick another. Refresh, pick, continue.

The default tier and the Recommended badge lists the tier the fallback lands on.

## Two different situations

| Situation | What the site does | What you see |
| A tier is withdrawn for good | Text and image fall back to the default; video to this episode’s last-used tier, else the default | The tier list no longer shows it; the price on the button is the replacement’s |
| A tier is temporarily unavailable | It is hidden from the tier list until it is back | It is missing from the list; pick another tier until it is back |
| The editor was open when a tier changed | The next generate is checked against the current list | “The selected video tier no longer exists — refresh and pick another.” |

## What does not change

Sheets and segments already made keep their pixels and their record of which tier made them. Versions are kept. SceneMixer regenerates nothing on your behalf, and charges nothing for a fallback; the charge happens only when you generate, at the price shown for the tier that runs.

## How the price stays honest

The price on the generate button is computed from the tier that will actually run, after any fallback. If your saved preference points at a withdrawn tier, the button already shows the price of the tier that replaced it. There is no case where you are charged for a tier you did not see named.

## What to do

Refresh the page; open the tier list; choose a tier from it; regenerate what you need. If you relied on a tier’s specific limits (the 30-second takes, the 2K render), check the replacement’s page for its own.

## FAQ

### Will my old segments be regenerated when a tier goes?

No. SceneMixer leaves existing renders and versions untouched.

### Which tier does the fallback use?

For text and images, the default. For video, SceneMixer uses the tier this episode last used if it is still listed, otherwise the default. The page on the default tier and the Recommended badge lists the current defaults.

### Can I be charged the withdrawn tier’s price?

No. SceneMixer's generate button shows and charges the price of the tier that runs.

### Why is a tier missing from the list today?

It is temporarily unavailable; SceneMixer shows hidden tiers again when they are back.

### Do I have to change my preference by hand?

Not for the fallback to work; do it when you want a specific replacement.

Updated November 3, 2026


---

# Wan 3.0 vs MiniMax H3: Which Suits the Scene | SceneMixer

URL: https://scenemixer.com/guides/when-to-pick-minimax-h3-over-wan/
Wan 3.0 against MiniMax H3 on SceneMixer: the longer take with more references versus the denser, higher-resolution segment, and how to choose for the scene in front of you.

# Wan 3.0 vs MiniMax H3: which suits the scene

Quick answer: : MiniMax H3 runs takes up to 15 seconds at 2K with nine images and three voice clips; the Wan 3.0 steps run up to 30 seconds with ten images and five voice clips, from 10 to 38 credits per second. Pick by the length of the take and the size of the cast.

The still above is from Neon Debt, an episode this site took from script to composite on Wan 3.0 · 720P. The same storyboard renders on any tier; the tier is a setting, not a rewrite.

MiniMax H3 and Wan 3.0 · 480P each have a full page.

## The differences that decide it

| | MiniMax H3 / H3 Fast | Wan 3.0 tiers |
| Segment length | Up to 15 seconds | Up to 30 seconds |
| Resolution | 2K / 768P | 480P / 720P / 1080P |
| Cuts inside one generation | Yes, several shots per segment | Yes, several shots per segment |
| Reference images | Up to 9 | Up to 10 |
| Voice clips | Up to 3, about 14 s in total | Up to 5, about 14 s in total |
| Sound | Always generated | Generated with the picture |
| Price per second | 25 / 16 credits | 10 (480P) to 38 (1080P) credits |

## Scenes that suit H3

Scenes headed for a large screen, where 2K earns its price: a confrontation between two faces, a reveal with an insert shot, anything that fits in 15 seconds and will be watched big. A lean cast helps on SceneMixer, since H3 takes fewer references.

## Scenes that suit Wan

A walk-and-talk that must not cut; a slow push-in that needs the whole take; a crowded frame that needs many character sheets attached; drafting an entire episode cheaply before choosing keepers.

## Switching between them

Each segment remembers its tier. Moving a segment from Wan to H3 may leave some references out, since H3 takes fewer, and a segment longer than 15 seconds cannot use H3 at all — the tier is greyed out for it; moving the other way needs neither. The composite stitches whatever versions are active.

## FAQ

### Can H3 do a 30-second take?

No. SceneMixer segments on H3 are up to 15 seconds.

### Is 2K better than 1080P for phones?

The difference is small on a phone; it shows on large screens.

### Why was a reference dropped when I moved to H3?

H3 takes up to 9 images and 3 voice clips; SceneMixer keeps the most important references and drops the rest.

### Can I use H3 for one segment and Wan for the rest?

Yes; on SceneMixer, mixing tiers inside an episode is normal.

### Which is cheaper?

Wan 3.0 · 480P is the cheaper place to draft on SceneMixer. At the top end, compare H3 at 25 credits per second with Wan 1080P at 38.

Updated November 3, 2026


---

# Pressed Back and the Videos Are Gone? Reading the Episode Cards

URL: https://scenemixer.com/guides/where-are-my-videos-reading-the-episode-cards/
Your clips live inside each episode, not on the episode list. What 'n/m videos done' and 'Composed' mean, the Start Editing button, and returning to your episode.

# Where Are My Videos? Reading the Episode Cards After You Go Back

Quick answer: Nothing is lost when you press ← Back to episodes: the clips stay inside the episode’s segments. The episode list shows them as a count on each card — 3/10 videos done in green — and a Composed badge once the episode has been joined into one video. Coming back from the editor, the list selects the episode you were in and scrolls to it; press Start Editing on that card to go back in.

Open a project, go to Step 3 and look at the card of the episode you worked on: the green count tells you how many of its segments have a clip.

## Where the clips are

Every clip belongs to a segment, and every segment to an episode. The episode list (Step 3) is one level up, so it shows episodes, not clips. Open the episode and the clips are on the timeline, each segment with its versions. The joined, full-length videos are under Composite History in the editor. What is inside a project.

## Reading a card

- n/m videos done (green) — n of the episode’s m segments have a finished clip. It counts the same segments that SceneMixer's Composite would join.
- Composed — the episode has been joined into a full video at least once.
- The rest of the card is as before: thumbnail, title, and the counts of characters, locations and segments.

A card with no count has no clips yet. Generate the storyboard and then the segments.

## Going back in

An episode whose storyboard is ready has Start Editing → on its card (in ad projects it says Start filming). Press it to open the episode editor. You can also select the card and use the button in the bottom bar. On a phone, the button sits on the same card.

## Back lands on your episode

The editor’s ← Back to episodes returns to Step 3 with that episode selected and scrolled into view, so on a long series you are not dropped back at episode 1. The browser also remembers the last episode you edited in each project: open the project on the same computer later and that episode is selected. On another device, SceneMixer's list starts from the first episode with a storyboard.

## Sample projects open on episode 1

Opening a sample project now takes you straight into the first episode’s editor, read-only, so you see finished clips right away. Its ← Back to episodes shows the sample’s episode list. Learn from a sample project.

| You see | It means |
| 3/10 videos done | 3 of 10 segments have a clip |
| Composed | The episode was joined into one video |
| No count on the card | No clips yet |
| Start Editing → | The storyboard is ready; open the editor |
| The card is highlighted after Back | That is the episode you just edited |

## FAQ

### I pressed Back and my videos disappeared.

They are inside the episode. Press Start Editing on that card in SceneMixer; they are on the timeline.

### Where is the full episode video?

In the SceneMixer editor, under Composite History, once you have composited.

### Why does the count differ from the number of segments?

SceneMixer counts only segments with a finished clip; generating or failed ones are not.

### Does the list remember my place on my phone too?

SceneMixer remembers per browser: each device keeps its own last-edited episode.

Updated November 3, 2026


---

# Where Your Credits Go in One Episode | SceneMixer

URL: https://scenemixer.com/guides/where-credits-go-in-one-episode/
Five things are charged and everything else is free. What each one costs, how the freeze-then-charge cycle protects you, and what happens when something fails.

# Where Your Credits Go in One Episode

Quick answer: Five steps cost credits: reading the script, generating setup images, generating each episode's storyboard, shooting each segment of video, and compositing the episode. Editing, previewing, matching voices, switching versions and re-reading what you already have are all free. Video is by far the largest line, and its price is per second of finished footage.

Every figure on this page is pulled live from the pricing API, so what you read here is what the buttons will say. The pricing page has the full table.

## The five charges

Reading the script — 4 credits per 3,333 characters, charged once. It produces the characters, locations, props and episode breakdown.

Setup images — 7 credits each on the default tier. Your account gets 15 free ones to spend across any dramas you have had read.

Storyboard — 1 credit per 333 characters of that episode's share of the script, charged per episode. Regenerating costs the same as the first time.

Video — from 10 credits per second of finished footage, depending on the model tier you pick. Regenerating a segment charges again.

Compositing — 1 credit per second of the finished episode.

## A worked example

On SceneMixer, a ninety-second episode shot on the cheapest tier and composited comes to about 990 credits. The storyboard for that episode is a small fraction of it, and the script read is paid once for the whole project rather than per episode.

The shape of the bill is always the same: video dominates, everything else is rounding. That is why the free allowance on a first project covers the cheap steps — they are where you find out whether the expensive one is worth running.

## Freeze, then charge

When you start something, its cost is frozen — held aside, not spent. If it succeeds, SceneMixer charges the frozen amount. If the model call fails, it is released back to you.

Every task is one line in your credit history (the latest in the credits panel in the top bar, View all for the full list), tagged In progress, Refunded or Under review where it applies, so a balance that looks wrong can always be traced to a specific line.

## What happens when something goes wrong

What refunds automatically: a request the model never received (a bad parameter, a reference image we could not fetch), the model being rate-limited or busy, and a clip stopped by content moderation. A generation you cancel before it starts also refunds on SceneMixer.

What does not refund automatically: a timeout, a 5xx or a network error mid-call. We cannot tell from our side whether the model ran and we were already billed, so the credits stay frozen rather than being either charged or given back, and a person settles it. A generation already running when you cancel is charged — the call has been made.

If a video is produced but we fail to deliver it to you, the amount stays frozen rather than being charged, and it is settled by hand. You are not billed for something you did not receive.

## Discounts apply when you buy, not when you spend

The discount is on the buying side, not the spending side. On SceneMixer, a membership discount and a promo code both come off the price of a credit pack or a subscription, and the combined discount is capped. What a generation costs in credits is the same number for everyone.

The price on a button and the amount actually taken come from the same calculation, so the number you agreed to is the number you pay.

| Step | Charged | Free? |
| Read the script | Once, by length | No |
| Setup image | 7 each | 15 free per account |
| Storyboard | Per episode, by length | Episode 1 on project one |
| Video | Per second, by tier | No |
| Composite | Per second | No |
| Edit, preview, versions | — | Always free |

## FAQ

### Why is video so much more expensive than everything else?

Because it is the only step whose upstream cost scales with the length of what you make. Text and image steps on SceneMixer are one call each; video is one call per segment, per second.

### Do I pay again if I regenerate a segment?

Yes, at the same price on SceneMixer. The earlier version is kept, so you are paying for an alternative, not a replacement.

### What if a generation fails halfway?

SceneMixer refunds an upstream failure. If a video was produced but could not be delivered to you, the amount stays frozen and is settled by hand rather than charged.

### Does membership make generating cheaper?

No — membership makes buying credits cheaper on SceneMixer, and it gives you a monthly credit allowance. A generation costs the same number of credits for everyone. The price on the button is always what is actually taken.

### Can I see where my credits went?

Yes — click your balance in the SceneMixer top bar: the latest entries are there, and View all opens your full credit history, one line per task, purchase or grant.

Updated November 3, 2026


---

# Where Is My History? Versions, Text, Composites and Credits

URL: https://scenemixer.com/guides/where-is-my-history-versions-text-composites-credits/
History on SceneMixer means seven different things: earlier clips of a segment, earlier storyboard text, past composites, credit and order records, feedback, notices.

# Where Is My History? The Seven Kinds, and Where Each One Lives

Quick answer: “History” is in a different place depending on what you are looking for. Earlier clips of a segment are in Version History, under the player in the editor. Earlier storyboard text is behind the History button on the segment’s storyboard. Earlier joined episodes are in Composite History in the editor’s top bar. Credits are in Credit History: the latest under your balance at the top right, the full list on the billing page; payments are in Order History on the billing page. Feedback you sent is in the feedback window’s History tab, and replies and notices are under the bell.

Open an episode, click a segment that you generated twice, and look under the player for Version History.

## Version History: earlier takes of a segment

Every time you regenerate a segment, the earlier clip is kept. When a segment has more than one, Version History appears in the right pane under the player, with each take and its label: generated, trimmed, AI-edited, redubbed, enhanced. Use one to make it the active take; it then plays, downloads and goes into the composite. Segment versions.

## Text history: earlier storyboard text

The History button on a segment’s storyboard opens Text history: SceneMixer always keeps the original text, along with the text from before each of your last five changes. Restore one to bring it back. Text history.

## Composite History: earlier joined episodes

Each time you composite an episode, the result is listed in Composite History in the editor’s top bar, with its status. Open one to play, download or share it. Composite and share.

## Credit History and Order History

Click your balance at the top right: the latest entries are listed under Credit History, and View all opens the full list on the billing page — one line per task, purchase or grant, with the project and segment a task was for. SceneMixer lists payments in Order History on the billing page, with unpaid ones under Pending Orders. Reading your credit history.

## Feedback and notices

Feedback you sent, with any reply, is in the feedback window under History. Replies from support and site notices also arrive under the bell at the top. How to get help.

## What has no history

Deleted projects, episodes and segments are not kept in a bin you can open; deletion is immediate. A re-parsed episode’s old storyboard and videos are not kept either. Deleted projects and files.

| You are looking for | Name on screen | Where |
| An earlier clip of a segment | Version History | Editor, right pane, under the player |
| Earlier storyboard text | History, then Text history | On the segment's storyboard |
| An earlier joined episode | Composite History | Editor top bar |
| Credits in and out | Credit History | Your balance, top right; the full list on the billing page |
| Payments | Order History, Pending Orders | Billing page |
| Feedback you sent | History tab | Feedback window |
| Replies and notices | Notifications | The bell |

## FAQ

### I cannot see Version History.

It appears once a segment has more than one take. SceneMixer's sample projects do not show it.

### How far back does text history go?

SceneMixer keeps the original text, and the text before each of your last five changes.

### Can I get back a deleted project from history?

No. Deleted projects are not kept on SceneMixer in a place you can restore from.

### Is there a list of everything I generated?

Per project: the episode cards show which episodes have videos, and each SceneMixer segment keeps its versions. Reading the episode cards.

Updated November 3, 2026


---

# Where the Music in an Ad Video Comes From

URL: https://scenemixer.com/guides/where-the-music-in-an-ad-comes-from/
Every SceneMixer ad gets its own music, written for that ad and mixed under the dialogue all the way through the end card. Where it comes from, what you can change, and how to use your own track.

# Where the Music in an Ad Video Comes From

Quick answer: Every ad gets its own music, written for that ad. When you make an ad, the music is planned from the ad’s storyboard and composed as a single piece; the clips themselves are generated without music of their own — or, when the piece is not ready in time, with the video model’s own music, the way ads used to be scored — and the final cut mixes the piece in under the dialogue — a little more forward than in a drama — and carries it across the end card to a proper ending. You add nothing, and there is nothing to pick from: it is this ad’s own score. Preview any clip with the Music switch on or off (older clips without the ad’s score do not show the switch); the finished ad always includes music.

Generate one segment of an ad with the Epic vibe and one with Warm: the end card colours follow the vibe, and the ad’s score is written for the ad itself.

## How an ad gets its music

The music is written for the ad, not picked from a library. Once the ad’s storyboard exists, the music is planned from it — the story, the length, where the lines fall — and composed as a single piece. The clips are normally generated without music of their own, so when the final cut is assembled, that one piece is mixed in under the dialogue and runs from the first beat to the end of the film, in the order the cut actually plays. If the piece is not ready by the time a clip is generated, that clip is made with the video model’s own music instead, as ads were scored before. Insert or drop a shot before you assemble it and the music still lines up. There is nothing for you to add, before or after. How sound works in a generation: Dialogue and Sound in One Generation.

## What the vibe controls

Each ad vibe on SceneMixer (Epic, Funny, Warm, Minimal, Twist, Mouth-watering close-ups, Show how it works, A day in the life) sets the ad’s look — its visual, scene and costume style — and the end card’s colours. Ad Options.

## One score, not one per clip

Because the music is composed for the ad as a whole, it does not restart or change character at every cut — the way an ad is scored on television. The clips are generated without music of their own so there is never a second score underneath; if the piece is not ready in time, that clip carries the video model’s own music instead, as before. Background music in an AI short drama covers the same system for dramas.

## Where you hear it

In the SceneMixer editor, a clip that has the ad’s score as its own track previews with the Music switch on by default (the switch beside the player controls it, and a clip downloaded with it off carries no music); flip it to hear the clip without the score, which is handy while reviewing the edit. Clips without the ad’s score — older clips carry their music in the take itself — do not show the switch. The finished ad always carries music. Compositing and sharing.

## Dialogue and the mix

Spoken lines come from the ad script and are performed by the people in the picture; ambience follows the scene. The music sits under both, and under a line it is eased down automatically — the ad’s music rides a little more forward than a drama’s, so the spot feels scored without burying the words. SceneMixer evens out loudness across the final cut; the mixing is done for you, and there is no level to set.

## The end card

The music runs across the end card and comes to an end with the film, so the spot closes on a resolved note rather than cutting to silence. The card is a fixed few seconds at the end. The End Card.

## Using your own music

Download the final cut (the original-quality file), open it in an editor, mute or lower the ad’s track and lay yours under it. If you only want a clip without music to work from, download that clip from SceneMixer with the Music switch off.

| Sound | Comes from | Changed by |
| Music | Written for the ad, mixed in at the final cut | Replace in an editor after download |
| Dialogue | The ad script, spoken on screen | Editing the script or the shot |
| Ambience | The scene | Editing the shot |
| End card | Carries the music to the end | — |
| Loudness | Evened out in the final cut; eased down under lines | — |

## FAQ

### Can I pick a song?

Not inside SceneMixer: the music is written for the ad. To use a different track, download the final cut and replace the audio in an editor.

### Is the ad scored as one piece or several?

SceneMixer writes one piece for the ad, and it runs through the whole final cut, so the music does not change style from clip to clip. If the piece is not ready in time, the clips are generated with the video model’s own music instead, as in an older ad.

### Is there a no-music option?

Not for the finished ad on SceneMixer — it always carries music. A clip that has the ad’s score as its own track can be played with the Music switch off, and a clip downloaded that way has no music; clips without the ad’s score do not show the switch.

### Does the end card have music?

Yes. SceneMixer's music runs across the end card and comes to an end with the film.

### Does the vibe change the visual style?

Yes. Each vibe comes with its own visual, scene and costume style, and sets the end card’s colours; in advanced mode you can change the three styles in SceneMixer's step 1.

### My older ad has different music in every clip. Why?

Ads made before October 5, 2026 were generated with a music description per clip, so the music can shift between clips. Ads made since are scored on SceneMixer as one piece written for the ad; when that piece was not ready in time, the clips carry the video model’s own music instead.

Updated November 3, 2026


---

# Where Your Files Live and What Deleting Does | SceneMixer

URL: https://scenemixer.com/guides/where-your-files-live/
Uploads go straight to object storage, never through our servers. What a link really is, which files use your quota, and what deleting does.

# Where Your Files Live

Quick answer: Your files sit in object storage and are reached by long, unguessable addresses. Those addresses are not secrets — anyone holding one can open it. Only what you upload counts against your storage quota; everything we generate is on us. Deleting removes it from your account immediately and the file itself is cleaned up afterwards, with a check that nothing else is still using it.

## Uploads do not pass through us

When you upload, your browser asks for a short-lived permission slip and then sends the bytes directly to object storage. The file never travels through our application servers. That is why a large upload is as fast as your connection allows rather than as fast as our queue allows, and why an interrupted upload leaves nothing half-written in your project.

The permission slip is good for fifteen minutes and only for the one address it was issued for. When the upload finishes, we look at the bytes that actually arrived — the real size and the real file type — rather than believing what the browser said it was sending.

## What a link really is

Every image and video you see on SceneMixer is shown through a signed address with an expiry on it. Think of it as a long, unguessable URL rather than as a locked door. We do not publish these addresses and they are not listed anywhere, but if one is copied out of the page, whoever holds it can open the file for as long as it is valid.

That is worth knowing before you paste a preview link into a group chat. If you want to show someone a finished clip, use the share button instead — that produces a proper page with a stable address you can reason about, and it is covered in its own guide.

The addresses are deliberately long-lived and stable rather than rotated every few minutes. A URL is also a browser's cache key: an address that changes constantly means your own gallery re-downloads every thumbnail every time you open it.

## Material we hand to a model service

To generate anything from a reference — a character sheet, a location plate — that reference has to be fetchable by the service doing the generating. So SceneMixer publishes a copy of it at an address that service can read, and hands over the address.

Two consequences worth being plain about. First, a published reference is readable by its address, not just by us. Second, it may be copied into storage in whichever region the model service runs in, which is not necessarily the region your account's own data sits in. If you are working with material that must not leave a particular jurisdiction, that is the constraint to weigh before uploading it.

## What counts against your storage

Only what you upload. Every image and video we generate for you — setup images, every take of every segment, every composited episode — is stored at our expense and does not consume your quota. The quota exists to bound how much raw material one account parks with us, not to meter the thing you came here to make.

The quota is reserved at the moment the upload is authorised and given back when you delete. An upload you start and abandon leaves its reservation behind for a while; a background job reconciles those away, so the number settles back to what is actually stored.

## What deleting does

On SceneMixer, deleting an asset, an episode or a project takes it out of your account straight away. The underlying file goes into a cleanup queue and is removed later rather than on the spot.

The delay is not a recycle bin. There is no restore — not on the site, and not by asking. The window exists for our own operational safety, and nothing about it is a service we offer you. Treat deleting as final.

Before a file is actually removed we check whether anything still points at it. Copying a character into another project shares the same underlying image, and an episode override can be the same file as something in your library. If a reference is still live, the removal is dropped and the file stays.

## Closing your account

Closing an account erases the personal details on it — the address it was registered with, the phone number, the profile — and replaces them with a tombstone. The account row itself is kept, because orders, payments and the credit ledger have to remain auditable and they reference it.

Two things to know before you do it. SceneMixer clears any remaining credits and does not refund them. And the objects belonging to the account are queued for removal, which is the same one-way street as any other delete.

| File | Counts against quota | On delete |
| Something you uploaded | Yes | Queued for removal; reserved space returned |
| A generated setup image | No | Queued for removal if nothing else uses it |
| A segment take | No | Queued for removal |
| A composited episode | No | Queued for removal |

## FAQ

### Is a preview link private?

A SceneMixer preview link is unguessable, not private. We do not publish it, but anyone who gets hold of it can open the file until it expires. For showing work to other people, use the share page instead.

### Do my generated videos use up my storage?

No. Only files you upload count toward your SceneMixer quota. Everything generated for you is stored at our expense.

### Can I undo a delete?

No. There is no restore on SceneMixer. Files are cleaned up after a delay, but that window is internal — treat deleting as permanent.

### Does deleting a character break another project that used it?

No. Before a file is removed, SceneMixer checks whether anything still points at it, and if something does, the removal is dropped and the file stays.

### What happens to my files if I close my account?

The personal details on the account are erased and replaced with a placeholder, the account row is kept for billing audit, remaining credits are cleared on SceneMixer without refund, and the objects are queued for removal.

### Where is my material processed?

References SceneMixer hands to a model service are copied to storage that service can reach, which may be in a different region from where your account's data normally sits.

Updated November 3, 2026


---

# Which Image Tier for Character, Location and Prop Sheets

URL: https://scenemixer.com/guides/which-image-tier-for-sheets/
A decision table for SceneMixer's ten image tiers: a whole cast, faces in close-up, 2K locations, long descriptions, state images and cover posters.

# Which image tier for your sheets

Quick answer: Start on the default, Qwen-Image 3.0, for everything — it is the only tier the free sheets are drawn on. Move the faces that carry the series to Qwen-Image 3.0 Pro, or to Nano Banana Pro on English-language projects with detailed costumes. Use SeedDream 5.0 Lite for 2K locations on a budget and SeedDream 5.0 Pro for photographic skin. For state images and the promo cover, which are drawn from an existing sheet, GPT Image 2.5 Sunburst is built to hold the reference. The choice is one setting and applies to new images only, so you can switch per group.

The default tier has the full explanation of the two reading habits every choice here turns on: Qwen-Image 3.0.

## The table

Ten image tiers can be picked in Settings → Model Preferences → Image Model. Prices are per image.

| You need | Pick | Per image | Why |
| A whole cast, cheaply; the free sheets | Qwen-Image 3.0 (default) | 7 | Follows plain descriptions; the only tier the free sheets are drawn on |
| Leads that must hold up in close-up | Qwen-Image 3.0 Pro | 18 | More identity in the face; marked Recommended in Settings; Chinese descriptions |
| 2K locations and props on a budget | SeedDream 5.0 Lite | 8 | 2K, reads the description to the end, low price |
| 2K and photographic skin | SeedDream 5.0 Pro | 25 | Reads the description end to end; strongest skin |
| English supporting cast and places, a step up from the default | Nano Banana 2.1 | 19 | About 2K; Google’s faster model; cheapest of the new tiers |
| Detailed costumes and faces from long English descriptions | Nano Banana Pro | 46 | About 2K; keeps the most named detail; the most expensive tier |
| Volume at 2K from another maker | GPT Image 2.5 Flare | 25 | About 2K; OpenAI’s fast general-purpose model |
| State images and covers drawn from an existing sheet | GPT Image 2.5 Sunburst | 25 | About 2K; OpenAI’s precise model, holds a reference subject |
| A cover with the title painted in | Grok Imagine Image 2.0 | 35 | About 2K; xAI’s model, strong on lettering; very long descriptions are shortened |
| A different look for faces | MAI-Image-2.6 | 31 | About 1 megapixel; Microsoft’s model; strict content filter |

Free sheets: every SceneMixer account’s 15 free sheets are drawn on the default Qwen-Image 3.0 only. Pick another tier and those images are charged at that tier’s price without touching the allowance; switch back and whatever is left is still there. See free sheets are per account.

## Two habits that decide most cases

How much of the description is followed. The Qwen tiers follow the start more reliably than the end; the SeedDream tiers read to the end. Character descriptions on this site are long, with identity first, which is why the Qwen tiers still do well on a cast: what matters most comes first. The same order — identity first, decoration last — also protects you on Grok Imagine, where a very long description is shortened at the end.

How literally phrasing is followed. All tiers act on what you describe; none reliably follows what you tell it not to draw. The SeedDream tiers follow result-style phrases most literally, so describe causes. See say what is there, not what is not.

## A working pattern

- Generate the whole cast on the default tier, using the free sheets. This is where you get the wording right.
- Look at the leads. Switch to Qwen-Image 3.0 Pro, Nano Banana Pro or SeedDream 5.0 Pro and regenerate only those.
- Switch to SeedDream 5.0 Lite or GPT Image 2.5 Flare for the locations if you want 2K plates.
- For state images and the cover, switch to GPT Image 2.5 Sunburst (or Grok Imagine Image 2.0 for a cover with a lot of lettering).
- Switch back. Each switch on SceneMixer affects new images only.

## What not to do

Do not iterate on a description at a premium rate: get the wording right on the default tier, then regenerate the keepers. Do not regenerate a whole cast on a premium tier when only the leads are seen in close-up. Do not expect a tier switch to redraw what you have — regenerating is a new image and a new charge.

## Each tier on its own page

- Qwen-Image 3.0 and Qwen-Image 3.0 Pro
- Seedream 5.0: Lite and Pro
- Nano Banana: 2.1 and Pro
- GPT Image 2.5: Sunburst vs Flare
- Grok Imagine Image 2.0
- MAI-Image-2.6

## FAQ

### Which tier is the default?

Qwen-Image 3.0.

### Which tier has the Recommended badge?

Qwen-Image 3.0 Pro.

### Which tiers render 2K?

SeedDream 5.0 Pro and Lite, Nano Banana 2.1 and Pro, GPT Image 2.5 Sunburst and Flare, and Grok Imagine Image 2.0. SceneMixer's MAI-Image-2.6 draws at about one megapixel.

### Do the free sheets depend on the tier?

Yes. The 15 free sheets per account are drawn on the default Qwen-Image 3.0 only; on any other SceneMixer tier images are charged from the first one.

### Which tiers read English descriptions best?

The Nano Banana tiers, Pro for detail and 2.1 for a lower price. Chinese descriptions read better on the Qwen and SeedDream tiers.

### Which tier for the promo cover?

On SceneMixer, the cover is drawn on your selected tier. GPT Image 2.5 Sunburst holds your leads' sheets well; Grok Imagine Image 2.0 is the one to try for lettering.

### Can I use different tiers for characters and locations?

Yes, by switching SceneMixer's preference between the two groups.

### Where are the Gemini 3.1 Flash and Gemini 3 Pro tiers?

The Nano Banana tiers you can pick on SceneMixer now are Nano Banana 2.1 and Nano Banana Pro. Gemini 3 Pro Image is the model behind Nano Banana Pro.

### Does switching tier change my existing sheets?

No. On SceneMixer, only new and regenerated images use the new tier.

Updated November 3, 2026


---

# Which Model Is Best? Pick a Video, Image or Text Tier by Purpose

URL: https://scenemixer.com/guides/which-model-is-best-pick-by-purpose/
Every video, image and text tier on the site: what each family is good at, where its price shows, and a decision table by purpose.

# “Which Model Is Best?” There Isn’t One — Pick the Tier by What the Shot Needs

Quick answer: There is no single best. Video tiers differ in resolution, maximum segment length and price per second; image tiers in how they read a description and how sharp a face they draw; text tiers in how carefully they structure a storyboard and what they cost. For most projects the defaults are the right call: Wan 3.0 · 480P at 10 credits a second for drafts, Wan 3.0 · 720P for finals, the default image and text tiers everywhere else. Switch tiers only for a reason listed below.

Open the model selector in the editor’s top bar: every video tier shows its resolution and its longest segment; the price per second is next to each tier in Settings.

## Video: four families

Four families, each in two or three resolutions. Wan 3.0 (480P / 720P / 1080P) is the default family: segments up to 30 seconds, dialogue and sound generated with the picture; its 480P tier is the usual way to draft. Seedance 2.5 (480P / Lite / Pro) also takes 30-second segments, at a higher price per second in its upper tiers. Seedance 2.0 (480P / Fast / standard) takes 15-second segments. MiniMax H3 (H3 / H3 Fast) takes 15-second segments, and has no 480P. Wan 3.0 and H3 both cut between shots inside one segment when the storyboard writes the cuts. Every price is per second; a 12-second segment on the default tier costs 12 × 10 credits. Full list with maximum lengths: which tiers accept 30 seconds.

## Video: pick by purpose

Drafting an episode: the default 480P tier. Generate every segment once, fix the storyboard where it is wrong, then regenerate only the segments you keep at a higher resolution — the draft-then-final workflow. A final for phones: SceneMixer recommends the 720P tier; 1080P for landscape or a screen larger than a phone. A long single take (a speech, a slow reveal): a 30-second family. A line the default tier keeps mispronouncing: regenerate that segment on MiniMax H3. Anime and stylised looks: any family; the visual style preset does the work, and 480P hides less than you would think.

## Image: two questions

Sheets are references, so the questions are: does the tier draw the face you described, and does it keep the background plain? The default Qwen-Image 3.0 reads Chinese and English descriptions well; Qwen-Image 3.0 Pro is the recommended step up for lead characters. SeedDream 5.0 Pro reads a long description in full and draws skin and cloth with a more photographic feel; SeedDream Lite is its lighter sibling. SceneMixer shows the price per sheet on the button before you generate. Writing the description so any tier draws the right face: Write a Description That Works.

## Text: the tiers that parse and write storyboards

Text tiers read your script and write the storyboard. The default Doubao Seed 2.1 Turbo handles most scripts; Doubao Seed Evolving is the recommended step up when a storyboard keeps coming back thin or mis-structured; Doubao Seed 2.1 Pro and Qwen3.8 Max are for long, dense material; Qwen3.8 Flash is a fast option for simple scripts. The parse and storyboard fees scale with the tier’s multiplier; each tier’s rate is shown next to it in Settings. Change the text tier in Settings; to use a different tier for one episode, switch before generating that episode’s storyboard and switch back after.

## How to change a tier, and what it does not change

Video tier: the button in the editor’s top bar, which also changes the video tier in Settings; each segment keeps the tier it was made with. Image and text tiers: Settings. Changing a tier on SceneMixer never regenerates anything by itself; existing sheets and clips stay. A tier that is switched off is replaced automatically, by the default or, for video, by the tier the episode last used. Settings, Item by Item and How Model Tiers Work.

| You want | Pick | Why |
| A full draft | Wan 3.0 · 480P | 10/s, 30-second segments |
| Final for phones | Wan 3.0 · 720P | Recommended; 19/s |
| Final for big screens | Wan 3.0 · 1080P or Seedance 2.5 Pro | 1080P; 38/s and 75/s |
| Lead character sheets | Qwen-Image 3.0 Pro | Recommended image tier |
| Storyboards that keep coming back thin | Doubao Seed Evolving | Recommended text tier |

## FAQ

### Which model is the best?

None outright. Draft on the default video tier, finish on 720P or 1080P, and switch only for a specific need: a 2K picture on MiniMax H3, or a line the default keeps mispronouncing (try H3 for that segment).

### Does a more expensive tier keep faces more consistent?

Consistency comes from the sheets and the storyboard, not the tier. A higher tier gives sharper frames, not a steadier face.

### Can I mix tiers in one episode?

Yes. On SceneMixer, the video tier is chosen per segment; the composite joins them regardless.

### Can I use a different text tier for one episode?

Yes. SceneMixer reads the text tier each time a task starts; switch it in Settings before generating that episode’s storyboard and switch back after.

### Where do I see the price before switching?

Under SceneMixer's Settings, next to each tier: per second for video, per sheet for images, as a rate for text. The generate button on a segment also shows that segment’s price.

Updated November 3, 2026


---

# Which Photos Work for an AI Ad Video | SceneMixer

URL: https://scenemixer.com/guides/which-photos-work-for-an-ad/
Up to four images: a product, a shopfront, a dish, a car, an app screenshot. What makes a photo usable, what happens to people in it, and which screenshots are refused.

# Which Photos Work for an Ad

Quick answer: On the home page, open the Ad Video tab and press Add images: up to four JPG or PNG files under 10 MB each, of something you have the rights to. The first image is the thing being advertised and fills the frame; the others show angles, details or the place it lives. A shop or a car with people in the photo is fine — people are removed. A screenshot with a real face is refused, and a full-page screenshot must be cropped to a phone or computer screen shape. After you press Make my ad, the image is read and you can correct the product name before anything is made.

How an ad is made from one photo walks the whole flow; this page is about the photos themselves.

## Where

Home page, Ad Video tab. Add images opens the file picker; Add another image adds more, up to four; Remove image takes one out. The note under the box says the rules: up to four images — product, shopfront, dish, car, home or app screenshot — JPG or PNG — make sure you own the rights.

## What makes a photo usable

The thing being advertised fills most of the frame and is in focus. A shopfront sign, the dish itself, the car, the room, the product on a table: anything the reading can recognise. If it cannot, the message says so — We couldn’t work out what is being advertised here — try clearer images — and offers Choose another. SceneMixer charges nothing for a refused image.

## The first image is the subject

The first image is the featured product: the one the ad is built around and the one the end card shows. The others are treated as angles, details or the place it is used, and become the sets the shots are made in. Put the hero shot first.

## People in the photo

A restaurant photo has diners; a car photo has a driver. That is fine: the reading notes it, and the composer says People will be removed. The place or the product is kept, the people are taken out, and the ad’s own cast is generated instead — so nobody who happened to be in your photo ends up in your ad.

Screenshots are the exception. A screenshot with a real face is refused (This screenshot has a real face in it. Please use one that shows only the interface), because editing the face out would smear the interface with it. Use a clean screen.

## Screenshots

On SceneMixer, an app or website screenshot is used as it is; it is never redrawn, so the interface stays readable. It must be roughly a phone or computer screen shape: a full-page capture is far taller than any video frame and is refused with This screenshot is too tall or too wide — crop it. Crop to one screen and upload again. Ads for apps and websites.

## What the reading returns

Press Make my ad and the composer says Reading your image…. A moment later a dialog says AI has read your product and shows what it understood: the category (Product, Food & drink, Apparel, Vehicle, Store / property, App, Website, Service) and a Product or place name field you can correct. If the reading has misunderstood the thing entirely, Pick different images. The one line you typed under the images — what it is and who it is for — is read too, and wins over the picture when they disagree. Nothing is charged until you choose how to make the ad. The two ways to make it.

## Signs and brand names

SceneMixer reads a name printed on a shopfront, a package or a car as the brand and prints it on the end card. Video models do not write legible text, so the sign in the picture is not something to rely on; the end card is where the name is delivered. The end card.

| Photo | Works? | Note |
| Product on a plain background | Yes | Best as the first image |
| Shopfront or restaurant with customers | Yes | People are removed; the place is kept |
| Dish, drink, garment, car, room | Yes | Fill the frame, in focus |
| App or website screenshot | Yes, if it is one screen | Never redrawn; no faces |
| Full-page screenshot | No | Crop to a screen shape first |
| Screenshot with a real face | No | Use a clean screen |
| Blurry, tiny, or ambiguous | No | Try clearer images |

## FAQ

### How many photos can I add?

SceneMixer accepts up to four, JPG or PNG, each under 10 MB.

### My shop photo has customers in it.

Fine. People are removed and the place is kept.

### Why was my screenshot refused?

Either it has a real face in it, or it is a full-page capture. Use a clean, one-screen crop.

### Can I change the product name after the reading?

Yes. The dialog after the reading on SceneMixer has a name field; correct it before choosing how to make the ad.

### Is the reading charged?

No. SceneMixer charges nothing until you choose one-click or expert mode.

Updated November 3, 2026


---

# Which Script Files Work: TXT, DOCX, Markdown, and What Does Not

URL: https://scenemixer.com/guides/which-script-files-work-txt-docx-markdown-and-what-does-not/
The script box takes pasted text and three file types. What happens to formatting and images in a Word file, why links and PDFs are refused, how to fix odd encoding.

# Which Script Files Work (TXT, DOCX, Markdown), What Is Ignored, and What Is Refused

Quick answer: Paste text, or drop a .txt, .docx or .md file. Word files are read in your browser: the text comes through, headings become plain lines, images, tables and comments are dropped. Markdown is read as text with its markers left in, which the parse ignores. Not accepted: PDFs (export to Word or copy the text), links to a page (paste the text itself), and very short text that cannot be a story. Encoding should be UTF-8; a file that arrives as gibberish was saved in an old encoding — open it in any editor and save as UTF-8. The counter under the box shows the characters read and the limit for your language; a script over the limit is split into volumes and each is its own project.

Drop a .docx: the box fills with its text and the counter updates; if it shows far fewer characters than the document has, the file was mostly images.

## The three ways text gets in

Paste from anywhere; drop or choose a file; or write with AI Create. All three end in the same box and go through the same parse. Upload a Script and Parse It.

## What a Word file loses

Text, line breaks and paragraph order survive. Bold, colours, fonts, headings-as-styles, footnotes, comments, tracked changes, images and tables do not; a script laid out as a table (name column, line column) comes out as running text, so check that speaker and line still sit together. SceneMixer keeps scene headings typed as text; headings applied only as a style become ordinary lines, which is fine.

## What is refused

Links: the box says it cannot open links; paste the text. PDF, images and archives: refused as soon as you pick the file, with Only TXT, DOCX and Markdown files are supported — please pick another file. Nothing is charged; copy the text or export to Word. Too short: a sentence or a title is refused before any charge. Too long: over the single-parse limit; split into volumes. Length Caps and How Characters Count.

## If the reading finds no characters or episodes

A file that looks like text but holds no readable story can still get through the box. When the reading cannot find characters and episodes in it, step 1 says The script couldn’t be read into characters and episodes (was the file plain text?). These credits are on hold pending a support review — no need to resubmit. Leave it; support releases the credits after checking. To try again in the meantime, paste the text itself into a new project.

## Encoding and odd characters

Text should be UTF-8, which is what every modern editor saves. Old files in a regional encoding show as garbage in the box; re-save as UTF-8. Invisible control characters are removed automatically. Curly quotes, dashes and full-width punctuation are fine and are kept.

## Writing the script for the parse

Dialogue as name-colon lines, delivery in brackets, captions marked; scene breaks as plain lines. The parse reads prose as well, but SceneMixer keeps lines written this way word for word. Write Dialogue So It Is Picked Up.

| Input | Result |
| .txt | Read as is |
| .md | Read as text; markers ignored by the parse |
| .docx | Text and breaks kept; formatting, images, tables dropped |
| PDF | Refused; export or copy the text |
| A link | Refused; paste the text |
| A title or one sentence | Refused as too short |

## FAQ

### My Word file has the script in a two-column table.

It will come out as running text; check that names and lines still pair up, or convert the table to text first.

### Can I upload several files at once?

SceneMixer takes one file per project; join them first, or parse each as a volume.

### Does the parse keep my scene numbering?

Scene headings survive as text; SceneMixer's parse decides episodes itself.

### Why does the counter differ from Word's count?

SceneMixer counts what was read, per the site's rules for your language.

### Is my file stored?

SceneMixer stores the text with the project; the file itself is not kept.

Updated November 3, 2026


---

# Which Text Tier for Reading Scripts and Storyboards | SceneMixer

URL: https://scenemixer.com/guides/which-text-tier-for-reading-and-storyboards/
A decision table for SceneMixer’s text tiers: when the default is enough, when to step up for a failing storyboard, and where the free first storyboard applies.

# Which text tier for reading and storyboards

Quick answer: Read every script and storyboard every episode on the default tier first. Step up only per episode, when a storyboard has failed its checks twice or a long script came back with episodes missing: Evolving or Pro at the base rate, then Qwen3.8 Max for dense Chinese scripts. Qwen3.8 Flash is for cheap, fast drafts. The preference applies per task, so nothing forces a whole series onto one tier.

The default tier explains the three jobs a text tier does and what the site checks.

## The table

| Situation | Pick | Rate |
| Anything routine | Doubao Seed 2.1 Turbo (default) | 50% |
| A storyboard failed twice | Doubao Seed Evolving (Recommended) or 2.1 Pro | 100% |
| Dense Chinese script, large cast | Qwen3.8 Max | 133% |
| Quick cheap draft to see an episode’s shape | Qwen3.8 Flash | 10% |

Rates are percentages of the base text rate; the pricing page turns them into credits.

## What the site does before you need to switch

Every storyboard draft is checked for format and language; if it fails, the site asks for one more draft on the same tier before reporting a failure. Long scripts are read in two passes automatically. Upstream busy errors are retried once. So a single failure is not a reason to switch; two in a row on the same episode is.

## The free first storyboard

SceneMixer's storyboard for episode 1 of your first project is free the first time it is generated — on the default text tier only. Generate it on any other tier and it is charged at that tier’s rate. While it is still unused, Settings shows a First one free badge on the default tier.

## Chinese or English

The default and Pro tiers handle both. For dense Chinese, Qwen3.8 Max is the natural step up; for a hard English script, Evolving or Pro. Reading a script on a cheap tier and storyboarding the important episodes on a pricier one is the usual way to spend.

## FAQ

### Does the text tier affect the video?

Only through the storyboard it writes; SceneMixer's video tiers render what you approve.

### Which tier is Recommended?

Doubao Seed Evolving.

### Do I need to re-read the script when I switch?

No. SceneMixer reads a script once per project; switching affects new tasks, such as regenerating a storyboard.

### Is the pricier tier always better?

No. On routine episodes the tiers produce comparable storyboards; the step up pays off on long, crowded episodes.

### Is the free first storyboard available on every tier?

No. SceneMixer offers it only on the default text tier.

Updated November 3, 2026


---

# While the Script Is Being Read: Stages, Leaving, Failure | SceneMixer

URL: https://scenemixer.com/guides/while-the-script-is-being-read/
What happens after you press Next in step 1: the stages, how long it takes, whether you can leave, what it costs, and what a failure does to your credits.

# While the Script Is Being Read

Quick answer: Press Next in step 1 and the reading starts: a panel titled Reading the script walks through finding characters, scenes and props, splitting episodes and shaping the visual style. It takes minutes, longer for a long script. You can browse elsewhere — progress stays in a pill in the corner and the project moves to step 2 on its own when it is done. The reading is priced by script length and charged when it finishes; if it fails, the credits are refunded and you try again.

A finished sample is what a reading produces once every later step is done; step 2 is where you will land first.

## What starts it

In step 1, once the project name, visual style, scene style, costume style, aspect ratio and type are set, Next saves the settings and starts the reading. Before that the button waits for the three style choices. If Next is greyed.

## The stages

The panel is titled Reading the script and shows the stage it is on: Finding characters, Finding scenes, Finding props, Splitting episodes, Shaping visual style, and for a long script Preparing episodes. Under it, short hints describe what is being worked out — the story core, the characters, the pacing. The names of characters and places appear as they are found.

## How long

Minutes for a typical episode-length script; longer for a long one, roughly in proportion. SceneMixer keeps preparing the later episodes of a long script in the background after you have moved on, so you can start on episode 1 while episode 30 is still being readied; an episode that is not ready yet says so when you try to storyboard it.

## Leaving the page

You can. The panel says it: you will move on automatically when it is done; feel free to browse elsewhere — progress stays in the corner. Close the panel and a small pill remains; open another project, read a guide, or close the tab. The reading finishes without you, and the project opens at step 2 next time.

## What it costs

SceneMixer prices the reading by the length of the script: 4 credits per 3,333 characters of English text, shown on the button before you press it. It is charged when the reading completes. Character, scene and prop images are separate and come in step 2, with 15 free per account. Where credits go.

## If it fails

The panel says Parsing failed. Credits have been refunded. Please try again. Nothing is kept from the failed attempt and nothing is owed. Press Next again. If the script itself is the problem — too long for a single reading, too short to be a story, or a link instead of text — you are told before the reading starts. Uploading a script.

## What you get

The panel changes to Script Parsed with a summary: title, protagonist, era, adaptation strategy, the counts of characters, scenes, props and episodes, and an outline per episode. Read it, then Next to SceneMixer's step 2 where the cast gets its images. Checking the summary.

| Stage | What is being worked out |
| Finding characters | Who is in the story, what each looks like |
| Finding scenes | Where it happens, what each place looks like |
| Finding props | The objects the story depends on |
| Splitting episodes | Where the episode breaks fall, by type |
| Shaping visual style | How the whole piece will look |
| Preparing episodes | Long scripts only: readying later episodes in the background |

## FAQ

### Can I leave while the script is being read?

Yes. SceneMixer keeps progress in a pill in the corner and the project opens at step 2 when it is done, even if you closed the tab.

### How long does the reading take?

Minutes for a typical script on SceneMixer, longer in proportion for a long one.

### What happens if it fails?

SceneMixer refunds the credits, and you press Next again. Nothing from the failed attempt is kept.

### Why does an episode say it is still being prepared?

SceneMixer keeps readying the later episodes of a long script in the background. Wait a moment and try again.

### Is the reading charged before it finishes?

SceneMixer charges the reading when it completes. A failed reading is refunded.

Updated November 3, 2026


---

# Who Is in Frame: Naming Characters and Off-Screen Voices | SceneMixer

URL: https://scenemixer.com/guides/who-is-in-frame/
Name every visible character with an @ chip and leave out the ones who are not. How to write off-screen voices, single-person shots and background people.

# Who Is in Frame

Quick answer: A shot draws the people you name in it. Put an @ chip on every character who is visible — @Mara turns from the window — and the shot uses their images. Do not put a chip on someone who is only heard, or who is in the room but not in this shot: an unseen name invites a face that should not be there. For a one-person shot, add In frame only: @Mara. For a voice from outside the picture, write the line with (voice-over) in its delivery note and leave the speaker out of the picture description.

Newer storyboards. In storyboards written or regenerated since late October 2026 there is no In frame only line to write: each shot’s cast is worked out by the system and shown under the segment as In frame. You still put an @ chip on every character who is visible. Notes under a segment · The segment header, field by field.

Open a finished episode and read a two-person scene. Notice which shots name both people and which name only one.

## The rule

Whoever is visible gets a chip. Whoever is not, does not. That is the whole rule, and it does more for face consistency than anything else you can write. Why a named character keeps her face.

A name typed by hand, without the chip, is not the same: the editor warns you after saving that it is not a character in this project, and no image is used for it. Type @ and pick the character.

## One person in the shot

Write the chip in the action and add the line In frame only: @Mara. That tells SceneMixer the frame holds one person, even if someone else is speaking or is nearby. It is the single most useful line for close-ups and reaction shots.

If she is reacting to someone we cannot see, say where she looks — she looks to frame left — rather than naming him. Naming him puts him in the picture.

## Two people, and who faces whom

Name both. Say the arrangement once: @Mara across the table from @Jules, or an over-the-shoulder shot from behind @Jules on @Mara. The scene’s automatic AI 3D staging keeps them on the same sides for every later shot, so you do not need to repeat it. How positions hold across cuts.

The one who is not speaking gets lips closed, no lip movement. It stops a listener mouthing the other person’s line.

## A voice from off screen

Someone speaks but is not in the picture: the narrator, a phone call, a shout from the corridor. Write the line normally, put (voice-over) or (off-screen) at the start of its delivery note, and tag it with the speaker’s voice chip. Do not put the speaker’s Cast chip in the picture text. The line is heard on SceneMixer; the face is not drawn. Writing dialogue lines.

## Background people

A crowd, guests, passers-by are not cast. They are declared once in the segment header, at the end of Set dressing: Background people: wedding guests seated in the pews, murmuring. Every shot in that segment then has them where they belong. Without that line, a shot that mentions “the congregation” has nowhere to put it. The segment header.

## A quick check before you generate

Read each shot and ask: who can I actually see? Those names have chips. Anyone else mentioned is a direction (she looks toward the door), a sound, or a voice-over. Then save and generate.

| Situation | Write |
| Only Mara is visible | In frame only: @Mara |
| Mara and Jules are both visible | Chips for both; say once who is where |
| Jules speaks from another room | His line with (off-screen) and his voice chip; no Cast chip for him |
| Mara reacts to someone we cannot see | she looks to frame left — no name |
| A crowd fills the room | Background people: in the Set dressing line of the header |

## FAQ

### What happens if I name a character who is not in the shot?

The shot tends to put a face where there should be none, or to swap faces. Name only the visible.

### How do I write someone speaking from off screen?

Write the line with (voice-over) or (off-screen) in its delivery note and the speaker's voice chip. Leave the speaker out of the picture text.

### Do background people need chips?

No. In SceneMixer, declare them once in the Set dressing line as Background people.

### What does 'In frame only' do?

It tells the shot that one named person fills the frame. Use it on close-ups and reaction shots.

### Does the listener need anything?

Add lips closed, no lip movement to the person who is not speaking, so they do not mouth the line.

Updated November 3, 2026


---

# Why a Generation Takes Time | SceneMixer

URL: https://scenemixer.com/guides/why-a-generation-takes-time/
Three limits decide when your clip starts: your slot count, a site-wide cap at the model service, and the model's own run time.

# Why a Generation Takes Time

Quick answer: Three things stand between the button and the clip. Your account runs a fixed number of jobs at once — the rest wait their turn and nothing fails. The model service has a cap shared by everyone on the site, so other people's work is in front of yours. And the model itself takes minutes, not seconds. The estimate on screen is an average of what recently finished, not a promise.

Everything described here is visible on a real episode: open a finished sample episode and look at the timeline — every card on it was one trip through the queue.

## Your own limit is a number of slots, not a speed

Your account runs a set number of generations at the same time. Start more than that and the extra ones park: they sit in a waiting state, and as soon as one of your running jobs finishes, the next one is promoted into the free slot automatically. You do not have to come back and press anything.

Parking is not a failure. Nothing errors, nothing is lost, and the credits for a parked job are already held aside — frozen, not spent. If you cancel it before it starts, they come straight back.

How many slots you get depends on your plan. A paid plan raises the number of jobs that can run at the same time, and the top plans also put your jobs in a priority queue ahead of everyone else's. What no plan changes is how long a single generation takes — that part is the model's, and it is the same model either way.

Three short interactive steps are deliberately never parked: writing a script with AI, suggesting a style, and matching a character voice. They run even when your slots are full — they are seconds-long text calls, and it would be absurd for them to queue behind your own ten-minute video. They do still occupy a slot while they run, so starting one when you are already at your limit can briefly park the next thing you submit.

## The shared limit

Behind your slots there is a second cap: how many jobs the whole site may have in flight at a model service at once. That counter is shared across every machine SceneMixer runs, so it is genuinely global — when the site is busy, your job waits behind other people's.

This is why the same segment can start instantly at one hour and wait at another, with nothing about your project having changed.

## Where the number on screen comes from

The estimate is arithmetic, not a forecast. We take how many jobs are queued on the same model service at your priority or above, divide by how many the workers there can run in parallel, and multiply by how long the last fifty successful jobs on that service took. Then we pad the result by a fixed safety factor, because an estimate that runs out before the job does is the worst kind of wrong.

So it is backward-looking by construction. It gets less accurate exactly when things are unusual — which is also when you are most likely to be watching it.

## What actually takes the time

Almost all of it is the model. Our own part — assembling the prompt, resolving which reference images to attach, writing the result back — is a couple of seconds at most.

For a rough sense of scale, a SceneMixer internal measurement taken in August 2026 put the average at around a minute and a half for a setup image, around nine and a half minutes for one segment of video, and around thirteen minutes to composite a full episode. Treat those as orientation, not as a service level: they were one measurement, they move with the tier you pick and with load, and nothing keeps them current.

The one number under your control is the tier. A cheaper, lower-resolution tier is usually also a faster one, which is why the fastest way to find out whether a shot works is to try it cheap first and only then spend the good tier on it.

## If it seems stuck

On SceneMixer, a job that has genuinely gone wrong does not sit silently — it ends. If the model never received the request, or answered that it was busy or rate-limited, the credits are released back to you. If the call timed out or the connection broke mid-flight, we cannot tell whether the model ran and billed us, so the credits stay frozen rather than being either charged or refunded, and a person settles it by hand.

What that means for you: a balance that looks short after something failed is usually held, not gone. Click your balance in the top bar to see your credit history: each task is one line, tagged Under review while held and Refunded once returned.

| Limit | Set by | What happens when you hit it |
| How many of your jobs run at once | Your plan | Extras park and are promoted automatically |
| How many jobs the site has at a model service | Site-wide, shared | You wait behind other people — unless your plan includes the priority queue |
| How long one job runs | The model and the tier you picked | Nothing to do but wait |
| AI script writing, style suggestions, voice matching | Exempt | Runs even when your slots are full |

## FAQ

### Does buying a plan make generation faster?

It raises how many of your jobs can run at the same time, gives you a monthly credit allowance, and on the top plans puts your jobs in a priority queue. A plan on SceneMixer does not change how long one generation takes — that is the model's run time, and it is the same model on every plan.

### My job says it is waiting. Have I been charged?

SceneMixer freezes the credits, which means held aside rather than spent. If the job runs and succeeds they are charged; if you cancel it before it starts, they are released back in full.

### Why is the estimate wrong?

It is the queue on that model service divided by how many jobs the workers can run there in parallel, times how long the last fifty jobs on it took, padded by a safety factor. SceneMixer's inputs are all measurements of the recent past, so the estimate is least reliable exactly when conditions are changing.

### Can I jump the queue?

Not per clip. The top plans do include a priority queue, which puts your jobs ahead of other people's in the shared line; below those, what you get is more slots of your own. A cheaper tier on SceneMixer also usually finishes sooner.

### Should I start everything at once or one at a time?

Start everything. Anything over your slot count parks on SceneMixer and is promoted the moment a slot frees, which is strictly faster than you noticing and pressing the button yourself.

Updated November 3, 2026


---

# Why a Segment Was Split in Two, and What 'Continues' Means

URL: https://scenemixer.com/guides/why-a-segment-was-split-in-two-and-what-continues-means/
A segment whose shots add up to more than one generation allows is cut in two at a shot boundary. Where the cut lands, what 'continues' means, why they cannot merge.

# Why a Segment Was Split Into Two, and What the ‘Continues From the Previous Segment’ Note Means

Quick answer: A segment is one video generation, and a generation has a maximum length (15 or 30 seconds by tier). When the storyboard writes a segment whose shots add up to more, the site first shortens non-dialogue shots; if it still does not fit, the segment is cut into two at a shot boundary, as evenly as possible, never between a shot and the one that continues it. The second piece gets its own header: its opening state describes its first shot, and carries a note that it continues from the previous segment, so the clip starts where the last one stopped. Props are listed per piece. Each piece is priced and generated on its own. They cannot be merged back into one segment; if you want a single clip, shorten the shots until they fit.

Open a split pair: the second segment’s header opens with the continues note and its first shot; the first segment ends at the shot before it.

## Why a segment is split

Segment length is decided in Step 3 (15 or 30 seconds); the storyboard aims at it but a dense passage can overshoot. Rather than fail the whole episode, the site cuts the segment: shots are kept whole, and the cut goes where the two halves are most even. Only a single shot longer than the cap by itself is a failure. 15 or 30 Second Segments.

## What the second piece looks like

Its header is rebuilt for its own first shot: the shot size, who is in frame, what is on screen; and a note that it continues from the previous segment. SceneMixer numbers its shots from 1 again. The two pieces are generated separately, like any two segments; the second follows on from the first through that header. Why the Next Segment Opens Like the Last One Ended.

## Price and generation

Two segments, two generations, each priced by its own length. The total is close to what one long segment would have cost, since video is priced per second; the only difference comes from each piece being rounded up to whole seconds and to the 4-second minimum. Changing Segment Duration.

## Merging back

Not offered: a merged segment would again exceed the cap. To get one clip, edit the shots so their seconds fit the segment length, then delete the second piece. Or accept the pair; a cut at a shot boundary is a normal edit point.

## Ads

SceneMixer fixes ad segments at 15 seconds and scales their shots to fit instead of splitting; a split never happens in an ad.

| Situation | Result |
| Shots add up to slightly over the cap | Non-dialogue shots shortened; no split |
| Still over after shortening | Split at the most even shot boundary |
| A shot continues the previous one | Kept together across the cut |
| One shot alone exceeds the cap | Task fails; edit the shot |
| Ad segment | Scaled, never split |

## FAQ

### Why was I not asked before the split?

The alternative was failing the storyboard; SceneMixer's split keeps everything you wrote.

### Can I move the cut?

Yes: edit the two segments' shots in SceneMixer, or insert and delete segments by hand.

### Does the second piece know the scene?

Yes. Its header on SceneMixer carries the location, time and props for its own shots.

### Does the split cost extra?

SceneMixer prices each piece by length; the sum is about the same as one long generation.

### Why does the second clip open on the same framing as the first one ended?

Its header on SceneMixer describes where the first piece left off. Edit its first shot to open on a different framing, then regenerate it.

Updated November 3, 2026


---

# Why a Task Sometimes Waits 15 Seconds More | SceneMixer

URL: https://scenemixer.com/guides/why-a-task-sometimes-waits-15-seconds-more/
The automatic retry on SceneMixer: when a service says busy, the site tries once more for you, sometimes after a 15-second pause. What triggers it and what it costs.

# Why a task sometimes waits 15 seconds more

Quick answer: When the service behind a tier answers busy or too many requests, the site does not fail your task at once. It tries once more: on another account for the same model if one is free, otherwise on the same one after a 15-second pause. That pause is the extra wait you occasionally notice. If the retry also fails because the service is busy, the task fails with a message saying the service hit a temporary problem and you were not charged; for other temporary errors the credits are held until support checks the task. Content rejections and invalid input are never retried; those need a change from you.

Why Pro tiers queue covers the other kind of waiting, before a request is even sent.

## What triggers the retry

| The service said | Retried? | Why |
| Busy / saturated | Yes, once | It usually clears in seconds |
| Too many requests | Yes, once | On text tiers and the Qwen image tiers the site first keeps resubmitting for up to about a quarter of an hour, then tries another account once; other tiers try once straight away |
| A transient error | Yes, once | It often clears in seconds |
| Content check did not pass | No | Retrying the same material would fail again; edit it |
| Invalid input (references, duration) | No | The request itself must change |

## How the retry runs

If another account for the same model is available, the retry goes there immediately. If not, the site waits 15 seconds and resubmits on the same account. For video, SceneMixer only retries within the first 30 minutes of a task; later than that, the task is left to complete or fail on its own.

## What it costs

Nothing extra. A task is charged once, on delivery, whichever attempt delivered it. If both attempts fail because the service was busy, the credits come back to your balance; for a few other kinds of failure they are held until support checks the task.

## What you can do

Wait. The retry is automatic, and nothing you do in the meantime moves it forward. If a tier is failing with busy repeatedly at a busy hour, generate on another tier for a while, or come back in a few minutes.

## FAQ

### Why exactly 15 seconds?

It is the pause SceneMixer takes before resubmitting on the same account when no other account is free; long enough for a busy service to clear.

### Does the retry double the price?

No. SceneMixer bills one delivery, one charge.

### Why was my rejected task not retried?

Content rejections and invalid input need a change from you; retrying the same request would fail the same way.

### Can I turn the retry off?

No; SceneMixer's retry is site-wide and free.

### Will a video task that fails after 40 minutes be retried?

No. On SceneMixer, the retry only happens within a video task’s first 30 minutes.

Updated November 3, 2026


---

# Consistent AI Series: Same Cast, Props and Rooms | SceneMixer

URL: https://scenemixer.com/guides/why-a-whole-series-stays-consistent/
One cast, one prop list and one layout per room, kept the same from the first shot to the last episode. What that looks like on screen, and what you never have to do.

# Why a Whole Series Stays Consistent on SceneMixer

Quick answer: A SceneMixer project is not a pile of separate clips. It is one cast, one prop list and one set of locations, fixed the moment your script is read, and every shot of every episode is generated from that same set. Faces, outfits, objects and the room you are standing in stay the same across the whole series. The system also builds an automatic AI 3D model of every scene, so people stay where they sat and the cuts make sense. You do nothing extra — you approve the cast once.

Watch it rather than read about it: open a finished episode and step through the segments on the bottom timeline. Same faces, same room, same props, shot after shot.

## One project, one cast

When you upload a script, SceneMixer reads it and produces three lists: the characters, the locations and the props. Those lists are the project. Each entry gets its own image in step 2, and that is the image the rest of the series is built from.

Episode 7 does not invent a new lead. It draws on the same entry you approved in step 2 of episode 1. Rename a character, change her description or replace her image, and every shot you generate from then on follows the new entry.

## Four things that stay put

Faces. The lead has one face. It is the same face in a close-up, in a wide shot, in episode 12. More on character consistency.

Clothes. A character wears what her entry says she wears. When the story needs a change of outfit, you make a state for it and choose which segments use it — the change is deliberate, not drift. How states work.

Props. A sword, a letter, a red flip phone: each is an entry with an image, and a shot that mentions it gets that object, not a look-alike. More on props.

Rooms. A location is generated once and used for every scene set there. The kitchen in episode 3 is the kitchen from episode 1. More on locations.

## The one people do not expect: where everyone is standing

Consistency is not only faces. In a scene with four people at a table, who sits where has to hold from one cut to the next, and the camera has to cut to angles that exist in that room.

SceneMixer builds an automatic AI 3D model of every scene: the room, the furniture, and where each person is placed. Every shot in that scene is generated against the same model. The person at the head of the table is still at the head of the table after the cut; the window stays behind the same chair; a reverse shot shows what is actually on the other side. Nothing to set up — it happens when you generate. More on spatial consistency.

## What it looks like when you watch

Ten shots in a row without a re-introduction. A cutaway to the second person and back, and the first person is in the same seat holding the same cup. The pen that was put down in shot 4 is on the table in shot 5. On SceneMixer, the next segment opens where the last one ended. More on continuity between shots.

That is what lets an episode be a hundred seconds long and still read as one scene rather than a montage.

## What you can change on purpose

Consistency is a default, not a cage. Edit a character’s description and regenerate her image. Make a state for a costume change or a damaged location. Use Edit (this episode) in the editor to swap a character’s image for one episode only. Regenerate one shot without touching the rest.

Each of those is a decision you make once, in one place. The series follows it from there.

| What stays the same | Where it is set | How you change it on purpose |
| A character’s face and build | Step 2, the character’s entry | Edit the description and regenerate, or upload your own image |
| What a character wears | The character’s entry | Add a state for the other outfit and pick which segments use it |
| A prop’s look | Step 2, the prop’s entry | Edit and regenerate the prop image |
| A location | Step 2, the location’s entry | Regenerate it, or add a state for a changed version |
| Who sits where, which wall is behind whom | Automatic, per scene | Change the storyboard text; the layout follows |

## FAQ

### Do I have to do anything to keep characters consistent?

No. Approve the cast once in step 2. SceneMixer generates every shot in every episode from those same entries.

### What if I want a character to look different in one episode?

Use Edit (this episode) in the editor to give that episode its own image, or add a state and assign it to the segments that need it. SceneMixer leaves the rest of the series unaffected.

### Does the layout of a room really stay the same across shots?

Yes. SceneMixer gives each scene an automatic AI 3D model, and every shot in the scene is generated against it. Seats, doors and who stands where hold across cuts.

### Will a prop look the same in every shot?

A prop is an entry with its own image. Shots that mention it get that object. Mention it with @ in the SceneMixer storyboard so the shot knows which one you mean.

### Can consistency break?

A shot can still come out wrong, the same way any generated shot can. When it does, regenerate that one shot; the rest of the series on SceneMixer does not move.

Updated November 3, 2026


---

# Why Characters Swap Sides Between Cuts (and the 180° Rule)

URL: https://scenemixer.com/guides/why-characters-swap-sides-between-cuts/
In AI video, people often flip left and right at the cut. What the 180-degree rule is, why shot-by-shot generation breaks it, and how a world for each scene keeps sides.

# Why Characters Swap Sides Between Cuts

Quick answer: Two people talking sit on an imaginary line drawn between them. Keep every camera on one side of that line and she stays on the left looking right, he stays on the right looking left, and the cuts feel invisible. That is the 180-degree rule. AI video breaks it because each shot is usually drawn on its own: nothing tells the close-up which side of the frame he belongs on, so he lands on either side. On projects with SceneMixer’s world model, everyone has a place in one measured room and every camera is placed in that same room, so the shot, the reverse shot and the over-the-shoulder are planned together and people keep their sides.

Open a finished episode, find two people talking, and watch three cuts in a row. Check one thing only: is each person on the same side of the frame every time they appear?

## The 180-degree rule in two minutes

Draw a line through the two people in a conversation. Film crews call it the axis, or the line of action. Put the camera anywhere on one side of it: a wide shot, over his shoulder, over hers, a close-up of either. In every one of those shots she is on the left and looks right, he is on the right and looks left. Their eyes meet across the cut.

Now put one camera on the other side. In that shot she is suddenly on the right, looking right, and so is he. The viewer reads it as the two of them swapping seats, or as both of them looking at someone else. They cannot say what is wrong, but they feel the jump.

## Why AI video breaks it

- Each shot is drawn on its own. “Close-up of Leo, angry” says nothing about which side of the frame Leo is on or which way he looks. The model picks, and it can pick differently every shot.
- The reverse shot invents a room. Behind him there should be the other side of the room she is facing. Drawn alone, it is whatever wall seems plausible: the window moves, a door appears.
- Eyelines drift. If she looks slightly left and he also looks slightly left, they are no longer looking at each other.
- Walking direction flips. Someone leaves frame on the right and enters the next shot from the right, walking back the way they came.
Writing “she is on the left” into every shot helps a little, but text is a weak way to hold geometry. The fix is to give the shots a shared space.

## How a world keeps sides

On SceneMixer projects with the world model, each scene gets a world before it gets a camera: a 360° view of the set, the room measured, and a place for every character. Then the cameras for that scene are placed inside the same room. A reverse shot is a camera turned to look at him, so it shows what is really behind him; who is on the left follows from where the camera stands, not from a coin toss. Eyelines and distances follow from the positions too.

Camera 2

Camera 3

Camera 4Four cameras in the same demo set. Turn a camera around and it finds another part of the same room, the doors or the windows, never a newly invented wall.

More on what the world is: What is a world model in AI video? Projects without it use the earlier automatic staging, which also places people from the storyboard: Everyone stays where they sat.

## Write the conversation so it can be staged

You do not have to write camera positions. You do have to say where people are and who faces whom, once, early in the segment. Then each shot can just name its framing:

- Opening: Wide shot. @Mia sits at the round table by the window; @Leo sits across from her, his back to the bar.
- Cut to over @Leo’s shoulder, medium close-up on @Mia. She slides the envelope across.
- Cut to the reverse: over @Mia’s shoulder on @Leo. He does not touch it.
- Cut to close-up, in frame only: @Mia. She waits.
“Across from her” and “his back to the bar” are what put the two of them on opposite sides of the line with the room behind each. The full shot format: How to write one shot. A phrasebook of positions: Write positions the world can read.

## When crossing the line is fine

- The camera moves across, on screen. A shot where the camera circles behind her takes the viewer with it. Write the move: “the camera circles round behind @Mia”.
- A new wide shot. A fresh wide shot of the room resets the viewer’s map; after it, a new line can apply.
- Someone moves. If he stands and walks to the window, the line now runs between her and the window. Write the walk so the new position is clear.
Breaking the rule on purpose is a choice; breaking it at random is a glitch. The difference is whether the viewer saw how the camera or the person got there.

## If sides still swap

- Are both people named? Everyone visible should be an @ chip in the shot. Who is in frame.
- Does the text say who faces whom? “Across from”, “beside”, “behind” place people; “they talk” does not.
- Are both segments in the same location? Two segments set in two different scenes are two different rooms.
- Does the project have the world model? New SceneMixer projects on paid plans do; older and free projects use the earlier staging.
- Is it one shot? Regenerate just that shot: Edit one shot and regenerate.

| Shot | Where the camera is | What the viewer should see |
| Wide | To one side, far enough to hold both | Her on the left, him on the right, the room around them |
| Over his shoulder | Behind him, same side of the line | Her face, his shoulder on the right edge |
| Reverse, over her shoulder | Behind her, same side of the line | His face, her shoulder on the left edge, the room behind him |
| Close-up | Near the over-the-shoulder spot | One face, looking off the same way as in the wider shot |
| Insert | Anywhere near the object | The envelope on the table; sides do not matter |

## FAQ

### What is the 180-degree rule?

Draw a line between two people in a scene and keep every camera on one side of it. Then each person stays on the same side of the frame and their eyes meet across the cut.

### Why do characters switch sides in AI videos?

Because each shot is usually generated on its own, and the text of a shot rarely says which side of the frame each person is on. The model picks, and can pick differently every time.

### Does SceneMixer follow the 180-degree rule?

On projects with the world model, the cameras for a scene are placed in one measured room where everyone has a place, so a conversation is planned with people keeping their sides. You can still ask for a camera move that crosses.

### Do I have to write where the camera goes?

No. Say where people are and who faces whom. If you do name an angle, such as over her shoulder or from the doorway, SceneMixer follows it where the room allows.

### Can I break the rule on purpose?

Yes. Write the move that takes the viewer across: the camera circling behind someone, a new wide shot, or a character walking to a new spot.

### Why did the window move between shots?

Without a shared room, each shot invents its own background. With the world model on SceneMixer the room is the same for every shot in the scene; if one shot still drifts, regenerate that shot.

### What about people walking?

Keep the direction consistent: someone who leaves frame on the right should come in from the left in the next shot. Name the destination, like walks to the door, so it stays in the same place.

Updated November 3, 2026


---

# Why Is This Button Greyed Out? Every Case | SceneMixer

URL: https://scenemixer.com/guides/why-is-this-button-greyed-out/
Next, Generate Video, Composite Episode, Delete segment, Remove Badge, Upload: every button that can be unavailable, why each time, and what unlocks it.

# Why Is This Button Greyed Out?

Quick answer: A greyed button is waiting for one thing, and hovering or pressing it usually says what. Next in step 1 waits for the three style choices. Generate Video waits for a script, for a running storyboard or video to finish, or for the segment to fit the tier’s cap. Composite Episode waits for every segment to be done. Delete segment refuses while a video generates and on the last segment. Remove Badge, uploads, custom styles, AI fine-tune and adding assets in the editor are membership features, and the button says so.

Open a finished episode: every segment is done, so Composite Episode is available — and Delete segment would refuse on the last one.

## Step 1: Next

Next stays off until the visual style, scene style and costume style are all chosen. They are pre-filled with a suggestion when the project was created; confirm or change each, and the button lights. Choosing the style settings.

## Step 2: Generate All and Generate

Generate All does nothing when every entry of that tab already has an image or is generating. A single entry’s Generate is unavailable while it generates. A state cannot be generated before its main image exists: This asset needs its main image first.

## Step 3: Parse Storyboard

Unavailable while that episode’s storyboard is being written, and while the episode is still being prepared after a long-script read (This episode is still being prepared — try again in a moment). Re-parse is refused while any of the episode’s segments is rendering video.

## Editor: Generate Video

Off, or refused with a message, when: the segment has no script yet; the episode’s storyboard is still generating; this segment’s video is already generating; an edit is still syncing; the segment exceeds the selected tier’s cap (over cap on the card); the segment’s dialogue is not in the project’s dialogue language (confirm to proceed); or you have as many jobs running as your plan allows. The messages.

## Editor: Composite Episode

Available only when every segment is done. Until then the tooltip counts: N/M segments done, composite available when all complete. Also unavailable while a composite is already running.

## Editor: Delete segment, Save

Delete segment refuses while that segment’s video generates (wait for it to finish) and on the last segment (At least one segment must remain). Save is off when nothing has changed.

## Membership features

On the current plan settings these are for members, and the button or panel says so: removing the AI badge (Upgrade to remove the Badge), uploading your own images and voices, custom style text in step 1, AI fine-tune on an image, and adding assets to an episode from the editor. SceneMixer leaves states (costume changes and the like) open to everyone. What a paid plan changes.

## Ads and remakes

An ad that is being made by AI keeps its editing buttons off until it is done (AI is still making this ad); cancel is off once shooting has started. A remake’s parse button does not exist: the project was read from your video, so there is nothing to parse.

| Button | Off because | Unlocks when |
| Next (step 1) | A style choice is missing | All three styles chosen |
| Generate Video | No script / storyboard generating / video generating / edit syncing / over cap / at the concurrency limit | The named thing finishes or is fixed |
| Composite Episode | Not every segment is done, or a composite is running | All segments done |
| Delete segment | Video generating, or last segment | Generation ends; or leave one segment |
| Remove Badge, Upload, custom style, AI Fine-tune, Add assets | Membership feature | A paid plan |
| Cancel (ad) | Shooting has started | — decide when it is done |

## FAQ

### Next in step 1 will not light up.

Choose all three style settings: visual, scene and costume.

### Generate Video is greyed on one segment only.

That segment has no script, is over the tier's cap, or is already generating. The card says which.

### Composite Episode is greyed.

Not every segment is done yet. The tooltip counts how many are left.

### Why can't I delete this segment?

Its video is generating, or it is the only segment left.

### Why is Upload greyed?

Uploading your own images and voices on SceneMixer is a membership feature on the current plan settings; the button says so.

Updated November 3, 2026


---

# Why Pro Tiers Sometimes Queue: Per-Minute Limits | SceneMixer

URL: https://scenemixer.com/guides/why-pro-tiers-queue-rate-limits/
Why a sheet or segment can wait before it starts on SceneMixer: your plan’s limit on tasks at once, the per-minute limit on the Pro image tier, and what to do.

# Why Pro tiers sometimes queue

Quick answer: Two limits can make a task wait. Your account runs a set number of tasks at the same time, depending on your plan; anything beyond that waits and starts as soon as one finishes — on every tier. On top of that, the Qwen-Image 3.0 Pro tier accepts fewer requests per minute than the default Qwen tier, so at busy moments its sheets wait a little and the site resubmits them for you. A sheet that takes a minute longer than usual on Pro is waiting its turn, not stuck; give it a minute before doing anything.

Why a generation takes time covers the other reasons a task can wait.

## The two limits

Tasks at once, per account. Your plan sets how many of your tasks run at the same time. Generate All on twenty sheets starts the first few and the rest wait; each starts when one finishes. This applies to every tier. Requests per minute on the Pro image tier. Qwen-Image 3.0 Pro accepts several times fewer requests per minute than Qwen-Image 3.0, which is why its sheets can wait where the default’s do not. None of this changes the price; a task that waited costs the same as one that started at once.

## What the site does

| What happens | What you see |
| Your plan’s number of tasks at once is reached | The extra tasks wait and start one by one |
| A Pro sheet lands beyond the minute’s limit | The card stays at generating a little longer |
| The site waits and resubmits, backing off a little each time | The task starts, usually within a minute |
| The limit is hit repeatedly at a busy hour | The task starts later than usual; still no extra charge |
| The service itself is saturated | “The generation service hit a temporary problem. You weren’t charged — feel free to try again.” — the site already tried again once |

## What you can do

Expect a big batch to trickle. Generate All starts a few at a time because of your plan’s limit; a higher plan runs more at once. Draft on the default tier. It accepts more requests per minute than Pro; move only the keepers to Pro, a few at a time. Give it a minute. A waiting task starts on its own; nothing you do in the meantime moves it forward.

## How to tell a queue from a failure

A waiting task has no error message and its card keeps spinning. A failure names its cause — content check, upstream busy, invalid input — and its credits come back on SceneMixer or are held until support checks it. If a card has been spinning for many minutes, refresh; if it is still spinning after that, write to support naming the project, episode and segment.

## FAQ

### Is a task that waited charged more?

No. SceneMixer charges it once, on delivery, like any other. If it fails, the credits come back or are held until support checks it.

### Does the default tier ever wait?

Yes, when your SceneMixer plan’s number of tasks at once is reached. It rarely hits the per-minute limit the way Pro does.

### Does Generate All count as one task?

No; SceneMixer counts each sheet as its own task, so a large batch fills your plan’s tasks-at-once and, on Pro, the per-minute limit.

### Should I generate it again while it waits?

No. A waiting task starts on its own; asking again does not move it forward.

### Do video segments wait too?

SceneMixer counts them towards the same tasks-at-once limit, and at busy hours a segment can also wait for capacity.

Updated November 3, 2026


---

# Why Shared AI Videos Carry the SceneMixer Badge, Even for Members

URL: https://scenemixer.com/guides/why-shared-videos-carry-the-scenemixer-badge/
Members can make clips without the AI badge, but anything shared through a SceneMixer link carries it. What changes, what stays yours, and the one-minute wait.

# Why Every Shared Video Carries the SceneMixer Badge

Quick answer: Every video shared from SceneMixer — a share link, a post from the share menu, the file your phone’s share sheet sends — carries the small SceneMixer badge in the bottom-right corner. If your clip has no badge (members can switch it off), a separate badged copy is made for sharing; your clip and your downloads are not touched. Making that copy takes about a minute the first time you share a clip; after that it is instant.

If you are a member and want an unbadged file for your own channel, use Download, not Share.

## What changes and what does not

- Shared: the video on the share page, its preview image when the link is posted, and the file your phone’s share sheet sends all carry the badge.
- Not changed: your clip in the editor, your versions, and Download — they stay exactly as they were, badge or no badge according to your setting.
- If your clip already has the badge, it is shared as it is.

The AI badge switch and What Download saves.

## Why

A shared video travels on without you: reposted, clipped, embedded. The badge tells anyone who meets it that it was made with AI, which is what AI-content labelling rules in many places expect, and it is the same everywhere, whoever made the clip. Keeping it off your own downloads is a membership choice; keeping it on what leaves through our share links is not. AI short drama compliance.

## The one-minute wait

The first time you share a clip without a badge, the share buttons show Adding the SceneMixer badge to your video, about a minute… and become active when the copy is ready. Share the same clip again later and it is immediate. If you change the clip — a new version, a trim, an edit — the new clip needs its own copy.

## Links you shared before

Share links made before this rule, for clips without a badge, now show the badged copy too. The first person to open such a link may see the video is being prepared for about a minute while the copy is made; after that the page plays as normal. It happens once per clip. Search engines see the page’s poster meanwhile, so nothing drops out of search. Share links: who can see them.

## Where the badge sits

Bottom-right, small, the same mark generated clips carry when their badge is on. It is burned into the picture of the shared copy, so it stays when the video is reposted.

| | Carries the badge? |
| Your clip in the editor | Only if your badge setting is on |
| Download | Only if your badge setting is on |
| Share link page | Always |
| File sent by the phone's share sheet | Always |
| Preview picture of a posted link | Always |

## FAQ

### I'm a member and removed the badge. Why is it on my shared video?

SceneMixer’s shared copies always carry it. Your own clip and downloads stay without it.

### Can I share without the badge?

Not through SceneMixer's share options. Download the file and post it yourself.

### Does the shared copy cost credits?

No.

### Why was my share button spinning?

The badged copy was being made; it takes about a minute on SceneMixer, once per clip.

Updated November 3, 2026


---

# Why Sheets Are Plain: White Background, Even Light, No Mood

URL: https://scenemixer.com/guides/why-sheets-are-plain-white-background-flat-light-no-mood/
Sheets are not stills from your film; they are references every shot is matched to. Why a moody sheet makes worse clips, and where the cinematic look actually comes from.

# Why Character, Scene and Prop Sheets Look Plain: White Backgrounds, Even Light and No Mood, on Purpose

Quick answer: A sheet is a reference: when a shot is generated, the sheets of everything in it are handed to the model as ‘this is what these look like’. The model copies what it sees — and it copies the light and the mood as faithfully as the face. A character sheet lit like a film noir poster brings that lighting into a sunny picnic; a scene plate with fog and god-rays fights the storyboard’s ‘hard noon light’ in every shot; a prop sheet on a wooden table brings the table. So sheets are deliberately neutral: characters on white with a plain side light that shows the bones, scenes as empty rooms with their layout and light source readable, props alone on white lit to show material. The cinematic look is written per shot in the storyboard, where it belongs.

Generate two clips of the same shot, once with the default neutral sheet and once with a sheet you regenerated with “dramatic lighting” in the description: the second carries that light into the shot regardless of the storyboard.

## What a sheet is used for

Every clip generation receives the sheets of the characters, states, scene and props named in the shot. They are the model’s only knowledge of what these look like, and it reproduces them closely: face, clothes, materials, and whatever light and background are in the image. Consistency across a series comes from this. Same Face in Every Shot.

## Why mood on a sheet is a problem

Mood is light. A sheet with a rim light, a colour cast, a smoky background or a silhouette carries those into every shot that uses it, on top of what the storyboard asks for that shot. The two fight, and the result is muddy or inconsistent from shot to shot. A neutral sheet can be lit any way the shot needs; a moody sheet can only be lit its own way.

## What each kind of sheet is optimised for

Characters: SceneMixer draws them full length on white, one plain side light, so the model sees the face and the build clearly; casting photo, not poster. Scenes: the empty location with three layers of depth and a visible light source, no people, no effects; a plate to shoot in. Props: the object alone on white, light only to reveal material. States: the same, changed only in the one thing the state is about. Scene Image and Description, Props.

## Where the cinematic look comes from

From the storyboard: each shot names its light (source, direction, quality), its time of day, its weather, its lens. The clip is generated from that text plus the neutral sheets. The visual style chosen in SceneMixer's Step 1 sets the medium and finish for everything. If you want a moodier film, write it into the shots, not into the sheets. Describe the Cause, Not the Result.

## Uploading your own images

Your own photo works as a sheet under the same rule: a plain, well-lit picture on a simple background serves every shot; a stylised portrait serves only shots that look like it. Crop out other people and text. Your Own Reference Images.

| Sheet | Kept plain | Because |
| Character | White, one side light, no expression drama | Face and build must read in any light |
| Scene | Empty, layout and light source visible, no fog or effects | Every shot in that room inherits it |
| Prop | Alone on white, material lit | Any hand, any table, any light later |
| State | As the base, plus the one change | Only the change should differ |

## FAQ

### Can I make the sheets look cinematic anyway?

You can regenerate with mood words; expect that mood in every shot and fights with the storyboard's light.

### Why is the character sheet a full-length shot?

Build and posture matter for consistency; the face is the priority within it.

### Does the visual style apply to sheets?

Yes, SceneMixer's visual style sets the medium and finish; the neutrality is about light and background.

### Why no people in scene plates?

They would appear in every shot of that room.

### Can a prop sheet show the prop in use?

No; the hand and the setting would come along. Usage is written in the shot.

Updated November 3, 2026


---

# Why Storyboard Labels Turn Into Pills When You Edit (Non-Chinese UI)

URL: https://scenemixer.com/guides/why-storyboard-labels-turn-into-pills-when-you-edit/
On a non-Chinese interface, the fixed parts of a storyboard show in your language and become pills in edit mode. Why, what a pill allows, what happens to pasted text.

# Why the Storyboard’s Labels Become Pills When You Edit, and What You Can Do With Them

Quick answer: A storyboard has two kinds of text: the words that describe the picture and the lines, and a fixed structure around them — shot numbers, header fields, voice tags, the opening-state line, the cut that starts each shot. The structure is stored in one fixed form so that every later step (checks, clip generation, translation) recognises it; on a non-Chinese interface it is shown in your language. When you edit, those labels become pills: you cannot type inside one, but you can delete it or move it, and the site guarantees that what you save is the exact fixed form. Text pasted from outside is handled too: common English spellings of the labels in the blocks you changed are recognised on save; copy-paste inside the editor keeps everything exact.

Open a segment in edit mode: the shot markers and header labels are pills; click between two pills and type; the pills stay put.

## What the pills are

Shot markers ([Shot 3]), header fields (Location:, Time:, Set dressing:, Props:, Opening:), the in-frame line, the voice tag after a spoken line, off-screen markers, and the cut that starts a shot (Cut to close-up). Each is one pill. What the fields mean: The Segment Header Fields.

## Why they are not plain text

SceneMixer reads the labels in the steps after you: the checks that count seconds, the clip generation that attaches sheets by name, the translation that keeps lines intact. If a label were typed slightly differently, those steps would not recognise it and the segment would quietly lose that structure. A pill cannot be misspelt. On a Chinese interface the labels are already in their stored form and are edited as text.

## What you can do with a pill

Delete it (select and backspace), move it (cut and paste, or drag), put text before or after it. The @ menu inserts character, scene, prop and voice references, not header labels. What you cannot do is type inside it or partially delete it. Two pills next to each other, or a pill at the very start of a line, leave a spot you can click into. Editing the Storyboard.

## Pasting text from elsewhere

If you paste a segment written elsewhere with English labels — from a document, from another tool — the common spellings ([Shot 2], Location:, In frame only) are recognised by SceneMixer when you save, in the blocks you changed. Unusual spellings are kept as plain text and lose their meaning; check the result. Copying from the editor and pasting back into it restores everything exactly.

## Voice tags and off-screen markers

The voice tag after a line ((voice: Mei)) binds the line to that character’s voice; it is a pill and should stay at the end of the line. Off-screen markers (voice-over, off-screen, narrator) in the delivery note are shown in your language and stored in fixed form as well. Dialogue Lines.

| Element | In reading mode | In edit mode |
| Shot marker | [Shot 3] | Pill; delete, move, insert |
| Header field | Props: … | Pill label, free text after it |
| Voice tag | (voice: Mei) | Pill at the end of the line |
| Off-screen marker | (voice-over) | Pill inside the delivery note |
| Everything else | Your words | Type freely |

## FAQ

### Can I remove a pill by mistake?

Yes; the segment then lacks that structure. Undo, or type the label in English at the start of its line (Location:, [Shot 3]); common spellings are restored on save.

### Why does the Chinese interface show plain text?

The stored form is Chinese; no display change is needed there.

### I pasted from my phone and a label became plain text.

An unusual spelling; retype it in a common form (Location:, [Shot 3]) at the start of the line, or undo the paste.

### Do pills affect the clip?

They are how SceneMixer's clip generation knows the structure; keep them.

### Can I rename a label?

No. On SceneMixer, labels are fixed; what follows them is yours.

Updated November 3, 2026


---

# Why the Next Segment Opens Like the Last One Ended, and How to Cut

URL: https://scenemixer.com/guides/why-the-next-segment-opens-like-the-last-one-ended/
Each segment is generated on its own, so openings can repeat. What a segment starts from, how to get a clean cut, and how to carry a wound or a coat across the join.

# Why the next segment can open like the last one ended — and how to change it

Quick answer: Each segment is generated on its own: the video model gets the character, scene and prop images and this segment’s storyboard text, not the clip before it. So two segments open on a similar picture when the text asks for it, or when the same people are in the same room. If an opening repeats the last framing, rewrite the first shot’s framing and regenerate that segment. What must carry over — a bandage, a coat taken off, someone lying down — is carried differently depending on the storyboard: in newer storyboards the site keeps a running record from your earlier text, so write the event where it happens; in older ones, write it into this segment’s Opening frame or first shot.

Open segment 2 of an episode, read its first shot, and compare it with the last shot of segment 1. Which kind of storyboard you have: the segment header.

## What each segment starts from

Two things: the images of everyone and everything named in the segment (with any state you picked for it), and the segment’s storyboard text. The finished clip before it is not part of that. That is why you can generate segments in any order, and why regenerating segment 3 never changes segment 4.

In newer storyboards there is a third thing: the running record of the story so far — who is where, what they hold, wounds, wet clothes, the time and the weather — built from the earlier segments’ text. It is what makes a bandage appear in segment 4 when segment 3 said it was put on. What the system keeps track of.

## Why two openings can still look alike

- The first shot names the same framing as the previous segment’s last shot, for example two medium shots of the same person. Older storyboards sometimes did this.
- In an older storyboard, the Opening frame of this segment describes the same position and distance the last segment ended on.
- Both segments are in the same room with the same people, so the picture is naturally close; only the framing tells them apart.

## How to get a clean cut

Edit the segment and give the first shot a clearly different framing or angle from the last shot of the segment before: a close-up on her hands after a wide shot of the room, or the reverse angle on him after a shot over his shoulder. In an older storyboard, check that the Opening frame line agrees with that first shot. Save, and regenerate this segment only. Editing text on SceneMixer is free; the new video is charged like any generation. Or let AI edit rewrite the first shot for you.

## Keeping state across the join: newer storyboards

Because the clip before is not a reference, a state only survives the cut if the record has it. The record learns from the text:

- Write the change as an event in the shot where it happens (@Mara takes off her coat and lays it over the chair, a cut opens above @Leo’s left eyebrow).
- If a state should last for many segments — a torn coat, a wrecked room — make it a state image and pick it for those segments rather than repeating words. Use a character state in one scene.
- If the first shot of this segment says otherwise (she walks in, unmarked), the text wins over the record. Fix the shot.
- If the change happened between the two segments, off screen, put it in this segment’s first shot: @Leo, a fresh bandage on his forearm, sits down.

## Keeping state across the join: older storyboards

On SceneMixer, an older storyboard has the Opening frame line and nothing else carrying the join. Write what must carry over there, or in the first shot: she is lying on the bed, a bandage on her left forearm. For a change of costume or a damaged room that lasts several segments, use a state image instead of repeating the words. Costume changes and scene states · Shot-to-shot continuity.

| What you see | Usual cause | What to do |
| Segment opens at the same framing the last one ended on | First shot (or, in an older storyboard, the Opening frame) asks for it | Rewrite the first shot with a new framing, regenerate |
| A bandage or prop disappears at the cut | Newer: the event was never written, or the first shot says otherwise. Older: only the previous clip showed it | Write the event in the shot where it happens, or in this segment's first shot |
| Costume reverts in the next segment | No state picked for this segment | Pick the state image for the segment |
| Regenerating segment 3 did not change segment 4 | Expected | Segments are generated independently |

## FAQ

### Does the next segment see the previous clip?

No. SceneMixer starts each segment from the images and its own storyboard text. In newer storyboards it also starts from the running record of the story, which is built from the text, not from the video.

### Why did my close-up come out as a medium shot?

Check the first shot, and in an older storyboard the Opening frame line; both should say close-up. If the text is right and the shot is still wrong, see when a shot comes out wrong.

### Does regenerating the previous segment change this one?

No. On SceneMixer, only this segment's own text, images and the record built from earlier text affect it.

### Do I have to generate segments in order?

No. Any order works on SceneMixer, and Generate all submits them together.

### Does a cleaner cut cost extra?

Editing the text is free. SceneMixer charges for regenerating the segment like any video generation; see the pricing page.

### My character's bandage vanished at the cut. What do I do?

In a newer storyboard, write the injury as an event in the shot where it happens and check the next segment's first shot does not contradict it. In an older one, write it in the Opening frame or first shot.

Updated November 3, 2026


---

# Why Video Costs More Than Everything Else: Credits Per Second

URL: https://scenemixer.com/guides/why-video-costs-more-than-everything-else/
Parsing is per thousand characters, sheets per image, storyboards per hundred characters, video per second. Why per-second dwarfs the rest and how to spend it well.

# Why Video Costs So Much More Than Sheets and Storyboards

Quick answer: One second of generated video costs about as much as one whole character sheet: the default tier is 10 credits a second, a sheet 7. A 12-second segment is therefore a dozen sheets’ worth, and a ten-segment episode is a hundred-odd. That is the shape of the underlying cost: a video model computes every frame of every second, an image model computes one frame, a text model computes a few pages. The way to spend less is not a cheaper button but fewer regenerated seconds — draft at 480P, fix the storyboard, finish once.

Change one segment from 8 to 16 seconds and watch the price tag double before you press anything.

## The four prices in one episode

Parse: 4 credits per 1,000 characters of the script, once. Sheets: 7 credits each at the default tier, after the free ones. Storyboard: 1 credit per 100 characters of the episode’s share of the script. Video: per second of the tier — 10 at 480P, 19 at 720P, 38 at 1080P — plus the composite at 1 per second. In a typical episode the video lines are more than nine-tenths of the bill. Line by line: Where Your Credits Go in One Episode.

## Why a second of video costs a whole image

An image model draws one frame. A video model draws twenty-four or more frames for every second, keeps them consistent with each other, and on the tiers used here also generates the voices and the sound in the same pass. Text models read and write a few pages. The provider prices follow that work, and SceneMixer's credit prices follow the provider prices, tier by tier. How tiers map to prices: How Model Tiers Work.

## Where the credits really go: retakes

The seconds you keep are a small part of the seconds you generate. A segment regenerated three times at 720P costs three segments. That is why the site pushes the fix upstream: a wrong face is a sheet problem, a wrong action is a storyboard problem, and both are cheap to fix compared with one more take. The draft-then-final routine turns most retakes into 480P retakes: Draft at 480P, Finish at 1080P.

## Seconds you can trim

Every shot in the storyboard has seconds; reaction shots and establishing shots often carry more than they need. Trimming a second from six shots saves six seconds of video across every take of that segment. Spoken shots have a floor (the line must fit); everything else is yours to shorten. Shot Seconds and Timing.

## What does not help

A lower tier does not lower the price of a mistake by much if you then regenerate at a higher one anyway; skipping the storyboard fix and hoping the next take is right is the expensive habit. On SceneMixer, a membership changes how many credits you get each month, not how many seconds a clip needs: What a Paid Plan Changes.

| Item | Priced per | Default price |
| Parse | 1,000 characters of script | 4 credits |
| Character / scene / prop sheet | Sheet | 7 credits |
| Storyboard | 100 characters of the episode | 1 credit |
| Video segment | Second, by tier | 10 credits at 480P |
| Composite | Second of the episode | 1 credit |

## FAQ

### Is there a flat price per episode?

No. SceneMixer prices video per second, so a longer or retaken episode costs more. The card prices add up in front of you.

### Why is 1080P almost four times 480P?

SceneMixer's price follows the provider's price for the higher resolution. Draft at 480P and pay 1080P once.

### Does the free preview count?

No; SceneMixer's preview is free and holds nothing.

### Is dialogue extra?

No. SceneMixer includes voices and sound with the seconds on every tier used here.

### Does a shorter storyboard make video cheaper?

Fewer seconds do. Cut shot seconds you do not need; SceneMixer's storyboard fee itself is small.

Updated November 3, 2026


---

# Working in Your Own Language | SceneMixer

URL: https://scenemixer.com/guides/work-in-your-own-language/
The interface, the storyboard you read, and the language your characters actually speak are three separate things. How each one is decided and which you control.

# Working in Your Own Language

Quick answer: Three languages are in play and they are deliberately separate. The interface is whatever you picked, in sixteen languages. The storyboard you read follows your interface language automatically, translated from the working copy the pipeline uses. The language your characters speak is a setting you choose before the script is read, and it is the only one of the three that changes the finished video.

Switch the site language from the top bar and open any project: the storyboard, the character names and the episode titles follow. This sample episode is the same document you would be reading. The video you already generated does not follow — it says what it was shot saying.

## Why there is a working copy and a reading copy

The storyboard is a technical document. Shot markers, duration fields, asset references and voice tags are a format the rest of the system reads literally — the rules that fix continuity, the code that builds the video prompt, the check that shot seconds add up. If the document were simply translated, those would stop matching and stop working.

So SceneMixer keeps two copies. The working copy keeps its format; the reading copy is what you see. When translating, the format markers are lifted out, carried across untouched and put back, and the result is checked before it is stored.

## Editing in your language

Edit the storyboard while reading it in German and your edit is saved in German, then translated back into the working copy. Only the blocks you actually changed are sent; everything you left alone is kept exactly as it was rather than round-tripped.

While that is happening, generating video for that segment is held for a moment so the shot is built from the updated text rather than the old one.

## When translation has not caught up

Opening an episode whose translation is still running shows a short wait with a way past it — you can always choose to read the original instead, and it lets you through by itself after a few seconds. It is a soft gate on purpose: a translation problem should never be able to lock you out of your own project.

If a translation fails outright, SceneMixer shows the working copy with a retry rather than a blank space.

## The one that changes the video

Dialogue language is chosen before the script is read, and it is applied at that step — the lines are written in your chosen language from the start, and the storyboard copies them through word for word.

Because it is applied at the read, changing it later means re-reading the script. The production document itself stays in its working language; only what the characters say changes.

Not every voice model is equally fluent in every language. Where a default tier is known to occasionally mispronounce a word in your chosen language, that is written on the option itself rather than left for you to discover in a finished shot.

## What follows your language and what does not

Following on SceneMixer: the interface, the storyboard you read, character and location names, episode titles, verification emails, and error messages.

Not following: video you have already generated, and the working copy underneath. Both are records of a decision already made.

| Language | Set where | Changes the video? |
| Interface | Top bar, 16 languages | No |
| Storyboard you read | Follows the interface | No |
| Dialogue | Step 1, before the read | Yes |

## FAQ

### Do I have to pick a reading language?

No. It follows the interface language you already chose, and SceneMixer has no separate control for it.

### Does translating the storyboard cost credits?

No. Reading your project in your own language on SceneMixer is free.

### If I edit in my language, does the system understand it?

Yes. SceneMixer saves your edit as you wrote it and translates it back into the working copy, with only the blocks you touched being sent.

### Can I change the dialogue language after generating?

SceneMixer applies it when the script is read, so changing it means re-reading. Video already shot keeps the lines it was shot with.

### What if a translation fails?

You see the original working copy with a retry. On SceneMixer, nothing is hidden and nothing blocks you.

Updated November 3, 2026


---

# Cinematic Camera Moves for AI Video: Framing, Movement, Cuts

URL: https://scenemixer.com/guides/write-cinematic-camera-moves-for-ai-video/
Write camera direction AI video can follow: framing sizes, moves in plain words (handheld, tracking, push in, aerial), angles, and cuts that change the framing.

# How to write cinematic camera moves for AI video

Quick answer: Write the camera in plain words at the start of each shot: first the framing (wide shot to close-up), then whether the camera is still or moving and how: handheld, following her as she runs; tracking alongside the car; a quick push in on his face; a high wide shot moving forward over the harbour. You do not have to place the camera: say who faces whom and the staging finds the angle. When two shots of the same people follow each other, the second changes the framing or the angle, and storyboards are checked for this.

Try it on one shot of your storyboard: change Medium shot to Low angle, medium shot, handheld following @Mara up the stairs and regenerate that segment.

## Framing first

| Framing | Shows | Use it for |
| Wide shot | The whole place, people small | Where we are; scale; endings |
| Full shot | A whole person, head to feet | Entrances, action, costume |
| Medium shot | Waist up | Conversation |
| Medium close-up | Chest up | The important line |
| Close-up | The face, or one object | Reaction, emotion, the clue |

How to write one shot · the mistakes camera lines most often make.

## Movement in plain words

| Move | Write it as | Feels like |
| Still | fixed camera | Calm, watching, tension |
| Handheld | handheld, following him through the crowd | Urgency, panic, documentary |
| Tracking | tracking alongside her as she walks | Journey, speed |
| Push in | a quick push in on her face | Realisation, shock |
| Pull back | pulling back to reveal the empty hall | Loneliness, a reveal |
| Aerial | high wide shot, moving forward over the rooftops | Opening, scale, “drone shot” |

Say which way a handheld camera moves; “slight handheld” tends to come out as a still shot. A very slow push can also come out still, so make the moves you care about clear and quick enough to see.

## Angles that mean something

Low angle makes someone powerful; high angle makes them small; over the shoulder puts us in a conversation; eye level is neutral. Name who faces whom and SceneMixer's automatic staging puts the camera where that is true. Write positions · Why characters swap sides.

## Cuts that change something

Two shots of the same people at nearly the same framing look like a jump, not a cut. Each cut changes the framing by at least a step, or changes the angle, or changes who we see; storyboards are checked for this when they are written, including across the join between segments. Shot-to-shot continuity.

## A chase, shot by shot

| Shot | Camera |
| [Shot 1] 3s | High wide shot, moving forward over the market roofs; @Mara running below. |
| [Shot 2] 2s | Handheld, following @Mara as she cuts between the stalls. |
| [Shot 3] 2s | Low angle, close-up on her feet skidding on wet stone. |
| [Shot 4] 2s | Tracking alongside @Mara; the crowd blurring past. |
| [Shot 5] 3s | A quick push in on @Mara's face as she sees the dead end. |

Short shots, a different framing every cut, one move each. Shot seconds.

## What not to write

Lens numbers, camera brands and strings of tags (“cinematic, 8K, anamorphic, masterpiece”) do little. Describe what it should look like instead: shallow depth of field, the background soft. No magic words.

## Make it in SceneMixer, start to finish

- Script: paste your own on Upload Script, or write the brief above in AI Create and let it draft the episodes.
- Step 1: Type Film; Scene Style Street Life, Neon Nights; Costume Style Streetwear, Smart Casual; Visual Style Digital HD, Shot on Film; Target Region; dialogue language; 16:9 Landscape. Then read the script.
- Step 2: characters, locations and props as usual; the camera is written in the storyboard, shot by shot, not here. The first 15 setup images on an account are free.
- Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
- Video: render on the default tier at 10 credits a second, fix single shots instead of whole episodes, then put the episode together and post. Before rendering, read each shot's framing and movement and edit the words where the camera should do something else.

| Moment | Framing | Movement |
| Opening | Wide or aerial | Moving forward |
| Conversation | Medium, then medium close-up | Fixed |
| The reveal | Close-up | Quick push in |
| The chase | Full and close-up | Handheld, tracking |
| The ending | Wide | Pulling back |

## FAQ

### How do I write camera moves for AI video?

Start each shot with the framing, then the movement in plain words: fixed, handheld following someone, tracking, a quick push in, pulling back, aerial.

### Do I have to say where the camera is?

No. Say who faces whom; SceneMixer's staging places the camera.

### How do I get a drone shot?

Write it as a high wide shot moving forward over the place.

### Why did my handheld shot come out still?

Say which way it moves, following which action; a vague or very slow move can come out still.

### Why do my cuts look like jumps?

Two shots of the same people at the same framing; change the framing or the angle. SceneMixer checks storyboards for this.

### Do lens and camera names help?

Rarely; describe the look you want instead.

### Can I change the camera after the storyboard is written?

Yes. Edit the shot text in SceneMixer and regenerate that segment.

Updated November 3, 2026


---

# Dialogue in the Storyboard: How to Write It, What Save Tidies Up

URL: https://scenemixer.com/guides/write-dialogue-in-the-storyboard/
How to write a dialogue line in the storyboard, which casual forms save tidies up, the messages it shows, and off-screen, phone and two-speaker lines.

# Dialogue in the Storyboard: How to Write a Line, and What Save Tidies Up

Quick answer: A line lives inside the shot where it is spoken, in four parts: the speaker chip, the delivery note in round brackets, the words after the colon, and the voice chip after the words — @Mara (low, steady): Read it before you say no. (voice: 🔊Mara). You can also write it casually — @Mara says: “Read it before you say no.” — and pressing Save rewrites it into the standard shape and tells you which shots it tidied. When it cannot tell who is speaking, or cannot read a line as a line, it says so instead, and tells you how to write it. Every message is described below.

Open an episode, press Edit on a segment, and find a line inside it: what you type there is exactly what the video says.

## The four parts of a line

@Mara (low, steady): Read it before you say no. (voice: 🔊Mara)

- The speaker chip — type @ (or press the @ button above the text) and pick the character from the Cast tab: the name becomes a chip tied to that character, not plain text.
- The delivery note in round brackets — how it is said, a few words: (low, steady), (through her teeth), (fast, out of breath). The note can be left out; a line with only a colon and the words works too — the voice chip is still needed.
- The words after the colon — what is said, written exactly as it should be spoken. SceneMixer's video reads the storyboard, not your script: if the line is shortened or reworded here, that is what is heard.
- The voice chip after the words — type @ again, open the Voice tab and pick the same character. It reads (voice: 🔊Mara) and says whose voice speaks the line. It is not decoration: a line without it is not treated as a dialogue line at all.

All of it stays inside the shot where the line is spoken; what happens after it — she closes her lips and looks away — follows in the same shot. The Cast chip and the Voice chip are two different things: the Cast chip puts the character in the picture, the Voice chip gives the line their voice. A spoken line usually needs both.

## Writing one, step by step

- Open the episode and press Edit on the segment. The text you see is the machine script — the thing the video is built from.
- Put the cursor in the shot where the line is spoken, after the action sentence: @Mara sets the envelope down. …
- Type @ and pick the character from the Cast tab.
- In round brackets, write how it is said: (low, steady).
- Type a colon, then the words: Read it before you say no. Keep the line as long as the shot can hold — about three words a second in English.
- Type @ again, open the Voice tab and pick the same character; the voice chip lands after the words.
- Press Save. Read the message that appears, if any; the four possible ones are below.

On SceneMixer, lines are spoken in the project’s dialogue language, chosen when the project was set up: write them in that language. Dialogue language · How a character gets a voice.

## What Save tidies for you

You do not have to write the full shape. Saving rewrites the common casual ways of writing a line into the standard one, and answers with one message naming the shots it tidied. Your words are kept — it only rearranges what is around them. The rewrite runs over the whole segment, not just the shot you changed, and it runs on every save that touches this text, including saving an AI edit and restoring an older version from History.

| You wrote | Save rewrites it as |
| @Seren (alarmed, spoken fast) Who are you? (voice: 🔊Seren) | @Seren (alarmed, spoken fast): Who are you? (voice: 🔊Seren) — the missing colon is added |
| @Severin says: “Aeylin. Are you ready?” (voice: 🔊Severin) | @Severin: Aeylin. Are you ready? (voice: 🔊Severin) — the speaking verb goes, the quotes come off |
| @Maia “Okay, okay. I’m quiet.” (voice: 🔊Maia) | @Maia: Okay, okay. I’m quiet. (voice: 🔊Maia) |
| @Erik steps toward her and angrily says, “Don’t call me sweetheart!” (voice: 🔊Erik) | @Erik steps toward her. @Erik (angrily): Don’t call me sweetheart! (voice: 🔊Erik) — the action stays, the manner moves into the note |
| @Jules is heard from the hallway, calling: “Dinner!” (voice: 🔊Jules) | @Jules is heard from the hallway. @Jules (off-screen): Dinner! (voice: 🔊Jules) — a voice from off camera is marked as one |
| @Maia says: “Dad!” (no voice chip) | @Maia: Dad! plus a voice chip for Maia — one is added where the speaker is clear |
| A line already in the standard shape | Left alone, byte for byte — model-written lines included |
| She says firmly: “Stand down.” — no name | Left alone. The editor warns instead: nobody is named, and it will not guess between the people in the shot |
| when @Maia whispers “Mother…” and the pebbles move — reported, not spoken now | Left alone. Reported speech inside a clause is not a line spoken in this shot |

What the rewrite will not do: it does not touch the words themselves; it does not guess a speaker for He / She says; it leaves a sentence whose subject is not clearly the speaker (a name standing as the object, looks at @Maia and says…); it leaves a name that is not one of the project’s characters; and it skips a sentence that was already spoken earlier in the same shot. The tidied text is what is saved from then on, and the version as it was stays in the segment’s History, where you can restore it. What the notes after saving mean.

## The four messages after Save

Tidied. “Tidied the dialogue in shot 2, 5 into "Name (delivery): line" so it is voiced with the character’s voice.” Nothing to do — the shot numbers are the numbering inside this segment, the same numbers in the text. SceneMixer's History holds the previous version.

Not recognized. “A line in shot 3 wasn’t recognized as dialogue, so it will be treated as description and not voiced. Write it as "Name (delivery): line".” There is a line in the shot, but it cannot be read as one: usually a missing colon, a speaking verb with the wrong person next to it, or a voice chip naming a different character than the speaker. The words are not lost — they stay in the shot as description — but nothing binds them to a voice, so the video may say them in some other voice or not say them at all. Rewrite the line in the standard shape before generating.

It doesn’t say who speaks it. “A line in shot 4 doesn’t say who speaks it, so it won’t use a character’s voice. Use the character’s name instead of He / She: "Name (delivery): line".” A pronoun is not a fact: She says could be any woman in the shot, so the editor will not choose for you. Write the character’s name; if a voice chip already names them, the line can be tidied from it.

Missing voice chip. “Some dialogue lines are missing voice tags — their audio may drift.” One or more lines have the shape of dialogue but no voice chip, so no character’s voice is bound to them. Add the chip from the Voice tab: a line without one is not treated as a dialogue line.

One message per kind per save; several shots are listed as “2, 5”. If none of the four appears, nothing was found to tidy or to warn about.

## Voice-over, narration, and phone calls

A line spoken by someone who is not in the picture starts with the mark in its delivery note: @Jules (off-screen): Wait! (voice: 🔊Jules), or @Narrator (voice-over): Ten years earlier. (voice: 🔊Narrator). The line is heard over the shot; the people in the picture keep their lips closed and nobody on screen mouths it. Do not put the speaker’s Cast chip in the picture text either: a character who is not in the shot gets no reference image on SceneMixer and is drawn as off-screen.

For a phone call, the person we cannot see is an ordinary line with the mark, and the note can say where the voice comes from: @Jules (off-screen, on the phone): Get out of the house. The person we see takes their normal line. A narrator who never appears at all is a voice-only character — give them a voice in step 2 like anyone else. Leads, supporting and voice-only characters · Who is in frame.

One thing that is not the mark: “looks off-screen” describes where her eyes go; she is in the shot and speaks on camera. If the voice is what comes from off camera, write it — is heard, off-screen.

## Two people in one shot

Each speaker gets a complete line of their own — speaker chip, note, words, voice chip — written in the order they speak, one after the other in the same shot. Never merge two speakers into one line and never leave one of them without their voice chip. The other people in the shot are handled on their own: SceneMixer keeps whoever is in the picture and not speaking with their lips closed.

## The delivery note

The note is what the acting follows, and it never changes the words. A few physical words work better than a mood: (low, steady), (fast, out of breath), (calling from the stairs), (stress on ‘never’) — a word in quotes has to be one that is in the line. One note per line. The voice-over mark lives in this same note, at its start. Dialogue lines: delivery, voice-over and captions.

## Names read the wrong way, and accents

The line is spoken exactly as written: the video voices the words in the shot, and SceneMixer has no separate pronunciation notation. If a name comes out wrong — a spelling with two readings, an unusual imported name — write it in the line the way it should sound. The character’s name on their sheet is not what gets pronounced; only the line is.

There is no accent or dialect field on a line either. Where an accent lives is the character’s Voice description, and from there it covers every line that character speaks in the episode — put it there first. Give a Character an Accent or Dialect in an AI Drama. When you only want one line to carry a little of it, put it in that line’s delivery note, short and physical ((thick Scottish accent)) — the acting is asked to follow it — and remember it applies to that line and not to the rest. The voice chip decides which voice speaks the line. How character voices work.

## When a line comes out wrong

| What you see | What it comes from | What to do |
| The shot plays silent — the words are never said | The line is not read as dialogue: no voice chip, or a form the reader cannot take apart | Write the standard shape and add the voice chip; check the message after saving |
| It is said in a voice that is not hers | The voice chip is missing (the “audio may drift” warning), or the chip names a different person than the speaker | Both names must be the same character |
| Someone else says it | The line hangs on the speaker chip written before the colon, not on whoever the words sound like | Fix the speaker chip |
| The words differ from my script | The video speaks the storyboard; the reading may have reworded or shortened the line | Fix the words in the shot — the script is not re-read |
| The person on screen talks over a line that should come from elsewhere | The note has no mark | Add (off-screen) or (voice-over) in front of the note |
| My line disappeared from the text | It was never a line the reader recognizes — most often the chip is missing or the name is not chipped | Write it as a full line with @ chips; the words themselves are never rewritten by the editor |

## FAQ

### Do I have to write the line in the standard shape?

No. SceneMixer tidies the common casual forms on save — Name says: “…”, a missing colon, quotes right after the name — and tells you which shots it tidied. Writing the standard shape yourself is still the most predictable.

### The editor rewrote my text. Can I get the old one back?

Yes. Press SceneMixer's History on the segment: the first version and the text before the last few edits are kept and can be restored.

### I wrote She says and nothing was tidied. Why?

Without a name there is nothing to tie the line to: He or She could be anyone in the shot, so SceneMixer's editor does not guess and instead says a line doesn’t say who speaks it. Write the character’s name.

### Which shot does the message mean?

The shot numbering inside this segment — the third block of shot text in it, not the segment’s number in the episode. Saving on SceneMixer rewrites the whole segment, and the message lists the shots it touched.

### Can a line be spoken by someone who is not in the shot?

Yes. Put (off-screen) or (voice-over) first in the delivery note and keep the voice chip; do not put the speaker in the picture text. The line is heard over the shot.

### Can two characters speak in the same shot?

Yes. Write one complete line per speaker — its own speaker chip, note, words and voice chip — in the order they speak.

### Do I pay for the tidy?

No. Editing text on SceneMixer is free; only generating the segment costs credits, at the prices on the pricing page.

### Should I type the character’s name as plain text?

No. Add characters through the @ menu in SceneMixer so the name becomes a chip linked to that character; a plain name in the text is not treated as dialogue, and the line won’t use their voice.

### I changed a line in my script file — will the video pick it up?

SceneMixer reads the script once, at the start. Later edits happen in the editor: the video speaks the storyboard. Changing a line after the clip is made.

Updated November 3, 2026


---

# Write Dialogue in Your Script So It Is Picked Up Word for Word

URL: https://scenemixer.com/guides/write-dialogue-in-your-script-so-it-is-picked-up/
The parse carries dialogue from your script into the outlines; the storyboard copies it word for word. The line shape it recognises, delivery notes, phone, voice-over, ca

# How to Write Dialogue in Your Script So the Parse Keeps Every Line, With Delivery, Off-Screen and Captions

Quick answer: Write each spoken sentence on its own line as Name: words, with a delivery note in brackets before the colon when it matters: Mei (quietly, on the phone): I saw him. The parse keeps lines in that shape, the storyboard copies them unchanged, and the clip speaks them. Dialogue folded into narration (‘she told him she had seen him’) is paraphrased, not spoken. Off-screen speech is a note, not a separate line: (voice-over), (on the phone), (over the radio). Text meant for the screen is its own line: (caption: three years later), though captions are not drawn in the clip at the moment. A group speaking in unison can be a speaker (Crowd: Run!); a group described as murmuring is background sound, not dialogue.

Parse a short scene written both ways — as lines and as prose — and compare the episode outline: the lines survive verbatim, the prose becomes a summary.

## The shape the parse recognises

Speaker: sentence. One sentence per line. Speaker names as they appear in your cast; nicknames are matched when the script introduces them. Brackets between the name and the colon hold the delivery note; brackets inside the line are read as stage direction and not spoken. Screenplay layout (name on its own line, dialogue below) is also read. What the parse does with the rest: Upload a Script and Parse It.

## Delivery notes

Put how the line is said in the brackets: (whispering), (through tears), (flat). The note travels with the line into SceneMixer's storyboard and shapes the performance in the clip. If you leave it out, the storyboard adds a plausible one; yours is better. Keep it to a few words; a sentence of direction is acting instruction for the storyboard, not for the line.

## Off-screen: phone, radio, voice-over

Write the off-screen marker first in the note: Father (on the phone, hoarse): Come home. The clip then plays the line without showing the speaker’s mouth. Voice-over narration is (voice-over); inner thoughts (inner voice). Without the marker, an absent speaker may be put in the picture. Storyboard side of this: Dialogue Lines, Voice-Over and Captions.

## Captions: text the audience reads

Time jumps and place cards go on their own line as (caption: …). On-screen captions are not switched on at the moment, so such a line does not appear as text in the clip, and the video models do not draw readable words on their own. Write time jumps and place cards this way anyway, so they carry over if captions are switched on, and add any readable text in a video editor after download. Text on Screen.

## What gets lost

Speech inside narration becomes summary. Lines with two speakers on one line (A: Yes. B: No.) are usually split correctly but are safer as two lines. Very long speeches are kept but will be cut across shots by the storyboard. SceneMixer translates dialogue in a language other than your project’s dialogue language into it at the parse.

| You write | The storyboard gets |
| Mei (quietly): I saw him. | The line, spoken quietly, on Mei's shot |
| She told him quietly that she had seen him. | A summary; nothing spoken |
| Father (on the phone, hoarse): Come home. | The line, heard, speaker not shown |
| (caption: Three years later) | Kept as a caption line; not drawn in the clip for now |
| Crowd: Run! | A group line, spoken in unison |

## FAQ

### Do I need screenplay formatting?

No. Name-colon lines are enough; SceneMixer also reads screenplay layout.

### Can a line have no speaker?

Then it is narration. Give it a speaker or mark it (voice-over) with a narrator name.

### How do I mark a line as a whisper?

In the brackets before the colon: (whispering).

### Will my exact wording be kept?

Yes, for lines in the recognised shape; SceneMixer's storyboard copies them word for word.

### My script is in Spanish but the project speaks English.

Lines are translated at the parse into the project's dialogue language; set that language in Step 1.

Updated November 3, 2026


---

# Write Fights, Crowds and Group Scenes That AI Video Gets Right

URL: https://scenemixer.com/guides/write-fights-crowds-and-group-scenes-for-ai-video/
AI video is best with few people and one clear action per shot. How to write a fight blow by blow, a crowd as background and a dinner of six without blended faces.

# Writing Fights, Crowds and Group Scenes for AI Video

Quick answer: AI video does best when each shot has a few named people and one clear action. Write a fight as a run of short shots, one blow or one reaction each, rather than one shot of a whole brawl. Put a crowd in the segment’s background-people line and keep it behind the named characters. For a group scene, show two or three people per shot and cut between them instead of framing everyone at once. Show the moment before or after a violent act rather than the act in detail, which also keeps clear of content review.

Find the busiest shot in your episode, count the named people in it, and split it if there are more than three.

## Fights: one blow per shot

Break the fight into beats: the swing, the block, the face of the one who takes the hit, the fall. Give each its own shot of two or three seconds and name who does what with @. Name the weapon or object with @ in the shot where it is used, so it keeps its look. Wide shots of many people fighting at once tend to blur; the cut between close shots carries the energy. How to write one shot · Props.

## Injuries and violence

Show the cause and the result, not the process: the raised knife, then her hand pressed to her side. Detailed wounds and gore are likely to be refused by SceneMixer's content review. Blocked by moderation.

## Crowds

A crowd is not a list of characters. Write it once in the segment’s background-people line — who, where, doing what — and keep the named characters in front. Anyone who speaks or is looked at individually is a character, not part of the crowd. Background people.

## Group scenes: the dinner of six

Four or more named characters in one shot is where faces start to blend. Open on one wide shot if you need the room, then cut to pairs: the host and the guest who answers, the two who exchange a look. Name only the people in frame in each shot; the others are off screen for that moment. My character’s face changed · Who is in frame.

## Big set pieces

Battles, stadiums, car chases with many vehicles: suggest the scale and film the people. A close shot of a soldier with smoke and noise behind him says “battle” better than an army. Sound helps: write it in the shot’s Ambient line.

| Scene | Weaker | Stronger |
| Fight | One 10-second shot of a brawl | Four short shots: swing, block, hit, fall |
| Crowd | A list of twenty extras as characters | One background-people line, named leads in front |
| Dinner of six | Everyone in every shot | A wide opener, then pairs |
| Battle | An army in a wide shot | One soldier, smoke and sound |

## FAQ

### How many people can be in one shot?

Two or three named people is the comfortable range. With four or more, faces start to blend.

### Can I have a long fight in one take?

It is possible, but long takes of fast action tend to drift. Short shots with clear actions hold up better.

### Why did my fight get blocked?

SceneMixer blocks detailed injury or gore. Show before and after instead.

### Do extras need sheets?

No. On SceneMixer, background people come from one line in the segment header.

Updated November 3, 2026


---

# Write Hooks and Cliffhangers Into Your Short Drama Script

URL: https://scenemixer.com/guides/write-hooks-and-cliffhangers-into-your-script/
Short drama lives on the first seconds and the last line of each episode. How to write an opening hook, end every episode on a question, and check both before generating.

# Writing Hooks and Cliffhangers Into a Short Drama Script

Quick answer: A short drama episode has two jobs: stop the scroll in its first seconds, and make the next episode feel necessary in its last. Open on conflict already under way, a line that raises a question, or an image that does not fit, not on scenery. End each episode on an open question: a reveal, a decision not yet made, a threat arriving. Vary the kind of ending so the series does not feel mechanical. Write both into the script before it is read, check the episode list in step 1, and edit an episode’s first shot in the editor if it opens too slowly.

Read the first and last line of each episode in your step 1 summary. If any episode ends on a resolved note, rewrite that ending in your next script.

## The opening hook

- Start late. Begin inside the argument, the chase, the slap, not before it.
- A line that asks. “You were supposed to be dead.” gives the viewer a question in the first breath.
- An image that does not fit. A bride in a hospital corridor; a CEO scrubbing floors.
- Skip the establishing shot. Place and time can come a few seconds later.

## The cliffhanger

End each episode one beat before the answer. Rotate between kinds so viewers cannot predict them:

- A reveal: someone learns who the other person really is.
- An open emotion: a confession left unanswered, a door about to close.
- A turn: the plan was a trap; the ally is the enemy.
- An arrival: a knock, a phone call, a face in the crowd.

## In the middle: one conflict per scene

Short episodes have no room for scenes where nothing is at stake. Give every scene someone who wants something and someone in the way. Inner thoughts read poorly on screen; turn them into a line, a look or an action. Write dialogue so it is picked up.

## Where to do it in SceneMixer

- Before the reading: write openings and endings into your script; the reading keeps the events and the order.
- With AI Create: say in your brief that each episode should end on a cliffhanger. Writing the one-line brief.
- After the reading: read each episode’s outline in step 1 before generating storyboards.
- In the editor: if an episode opens slowly, edit its first shot to start on the conflict. Editing text is free. Edit one shot.

| Episode moment | Weak | Strong |
| First seconds | A wide shot of the city at dawn | She throws the ring at his feet |
| First line | “Good morning.” | “Who told you I was here?” |
| Last beat | They make up and smile | His phone lights up: a message from her sister |

## FAQ

### Does every episode need a cliffhanger?

In a short drama series, yes, except the last. A resolved ending gives viewers a reason to stop.

### Will the reading keep my hook?

SceneMixer keeps the events and their order. Check the outline in step 1 and the first shot in the editor.

### How short is too short for a hook?

A line or an image is enough. The point is a question in the viewer's mind, not length.

### Can I add a hook to an episode already made?

Edit the first shot and regenerate that segment, or move a stronger moment to the start.

Updated November 3, 2026


---

# How to Write Positions in a Storyboard: Sit, Stand, Face, Move

URL: https://scenemixer.com/guides/write-positions-the-world-can-read/
Where people stand, sit and walk comes from your storyboard text. A phrasebook for tables, doors, windows, beds and cars, plus posture, facing and movement, by example.

# Write Positions the World Can Read

Quick answer: Positions come from the storyboard text. Write them as relations to things in the room and to each other: at the head of the table, across from her, in the doorway, with his back to the window. Say the posture when it is not standing, and who faces whom. Write it once at the start of the segment, and again only when someone moves. Prefer room anchors over screen directions: “by the window” holds in every shot, while “on the left” is only true from one camera. Where you say nothing, people are still given a sensible place and kept there.

Open a segment in the storyboard and read its opening shot. If it does not say who is where, add one sentence, such as @Mia sits at the round table by the window; @Leo sits across from her, his back to the bar.

## Where positions come from

When a segment is generated, every character in it is given a place in the scene’s room, read from what the storyboard says. On projects with the world model that room is the measured set built from the scene image; on other projects it is the earlier automatic staging. Either way, the text is the input.

Everyone in the scene always has a place. If the text says nothing about someone, they are put somewhere sensible and kept there, but “sensible” may not be what you pictured. One sentence settles it.

## Anchor people to the room, not to the frame

“On the left” only means something from one camera. Turn the camera around for the reverse shot and the left is now the right. “By the window” means the same thing from every camera, because the window is part of the room.

- Good anchors are things the scene image actually shows: the table, the bar, the door, the window, the bed, the stairs, the car.
- Good relations are between people: across from, beside, behind, facing, with her back to him.
- Weak anchors are screen directions (left, right, foreground) and things the set does not have. If the text puts someone by a fireplace and the scene image has no fireplace, there is nothing in the room to put them by.

## A phrasebook for common places

| Place | Phrases that place people clearly |
| Table | sits at the head of the table · sits across from @Leo · sits beside her, on the window side · stands behind his chair |
| Bar or counter | sits at the end of the bar · leans on the counter facing the room · stands behind the bar |
| Door | stands in the doorway · waits just inside the door · leans against the closed door |
| Window | stands at the window, looking out · sits with her back to the window |
| Sofa, bed | sits on the sofa next to @Mia · sits on the edge of the bed · lies on the bed, facing the wall |
| Desk | sits behind the desk · stands in front of the desk, facing him |
| Stairs | stands at the top of the stairs · sits on the bottom step |
| Car | in the driver’s seat · in the passenger seat · in the back seat behind the driver |
Cars, buses, trains and planes have their own guide: Scenes inside a car, bus, train or plane.

## Posture: say it when it is not standing

Standing is assumed. Anything else should be in the text, in the shot where it starts: sitting, kneeling, crouching, sitting on the floor, lying down, seated in a vehicle. A figure that should be kneeling but is written as nothing will stand.

- @Ava kneels beside the bed.
- @Noah sits on the floor against the wall, knees up.
- @Leo lies on the sofa, one arm over his eyes.
While a segment renders, SceneMixer's 360° preview draws each person in the pose that was read: Look around the set while a shot renders.

## Facing and eyelines

Say who faces whom, and who has their back to what. That decides which way each person looks, and so which cameras can see their face.

- @Mia faces @Leo across the table.
- @Leo turns away from her, toward the window.
- @Ava stands behind @Mia, watching her, unseen.
Why this matters at the cut: Why characters swap sides between cuts.

## Movement: start, destination, one move per shot

A move needs a destination the room has. “He moves” says nothing; “he walks to the door and stops in the doorway” gives a place to end in, and that place holds for the next shots.

- One move per shot. Standing up and crossing the room are two beats; give them two shots or enough seconds.
- Say where they end. The next shot starts from there.
- Keep direction. If she leaves the shot heading for the door, the next shot should find her near the door.

## Outdoors, crowds and checking your work

Outdoor scenes on SceneMixer get a lighter world: the ground, the nearby objects and the distance in four directions. Anchor people to objects that are really there, such as the bench, the fountain or the car, rather than to open ground. Background people are written separately from the named cast: Background people in a scene.

To check: on projects with the world model, select a segment that has not been generated and drag around the set to see what the room holds. While it renders, the figures show where everyone was placed. If someone is wrong, fix the sentence and generate again once the run finishes.

| Instead of | Write |
| They talk. | @Mia sits across from @Leo at the round table. |
| On the left of the screen. | By the window. / With her back to the bar. |
| He moves. | He walks to the door and stops in the doorway. |
| She is down. | She kneels beside the bed. |
| In the car. | In the passenger seat, beside @Leo who drives. |
| Somewhere in the room. | At the end of the bar, facing the room. |

## FAQ

### Do I need to write positions in every shot?

No. Write them once at the start of the segment, then again only when someone moves or changes posture.

### What happens if I write nothing about positions?

SceneMixer still gives everyone a sensible place and keeps them there across the shots. It may just not be the place you pictured.

### Can I use left and right?

You can, but they change with the camera. Anchors in the room, like the window or the door, mean the same thing from every angle.

### What if the scene image does not have the thing I mention?

Then there is nothing in the room to put the person by. Anchor them to something the image shows, or make a new scene image that has it.

### Why is someone standing when I wrote that they sit?

Check that the posture is in the text of the shot where it starts, and that the person is named with an @ chip there. Then regenerate the shot.

### Does writing positions cost anything?

No. On SceneMixer, positions are part of the storyboard text; you pay only for what you generate.

### How many people can I place in one scene?

SceneMixer gives everyone named in the segment a place. For a crowd behind the main characters, describe background people instead of naming each one.

### Can I drag people into place instead of writing it?

No. SceneMixer's 360° preview is read-only. The text is the way to place people.

Updated November 3, 2026


---

# AI Create Writes the Script in Your Language | SceneMixer

URL: https://scenemixer.com/guides/write-the-script-in-your-language/
The script from AI Create comes out in the language you use the site in, one of sixteen. How to switch before you write, how the price follows, where dialogue is set.

# Writing the Script in Your Language

Quick answer: Switch the interface language — the language button at the top, or Settings → Language — before you press Start Writing, and the script is written in that language: Korean interface, Korean script. Sixteen languages are available. The price on the button follows the language, because a Korean or Japanese script counts like Chinese while English and other alphabetic languages count by the reduced rule. The language the cast speaks in the videos is a separate choice, made in step 1 under dialogue language.

Working in your own language covers the rest of the site.

## Switch first, then write

The language you are using when you press Start Writing is the language the script is written in. Change it with the language button in the top bar or Settings → Language; the page switches at once. A brief written in another language is fine — the script follows the interface, not the brief. A Traditional Chinese interface gets a Chinese script.

## The price follows

Chinese, Japanese and Korean count every character; alphabetic languages count 3,333 characters as 1,000, the same rule as the read. So the same Short option shows one price in English and another in Korean. The menu on SceneMixer shows the figure for the language you are in. Length and cost.

## What happens after

The read, the cast descriptions and the storyboards are produced from your script and shown to you in your interface language; the summary and the storyboard text you see in the wizard are in the language you chose. Editing them is the same as for any project. After the script is written.

## The language the videos speak

Separate. In step 1, dialogue language decides what the cast says out loud; by default it follows the site’s production language, so a Korean script does not automatically give Korean-speaking videos — pick Korean there. Some languages carry a note about occasional mispronunciation on the default tier. Dialogue language.

## Writing in a language you do not read

Works, but you cannot check the draft. Better: write in your own language, read it, and set the dialogue language to the target; the videos speak it while you keep a script you understand.

| You want | Set | Where |
| The script in Korean | Interface language: Korean | Top bar or Settings, before Start Writing |
| The videos to speak Korean | Dialogue language: Korean | Step 1 |
| Both | Both of the above | — |
| The script in your language, videos in another | Interface: yours; dialogue: the other | — |

## FAQ

### How do I get the script in German?

Switch SceneMixer's interface to German before pressing Start Writing.

### Why does the same length cost more in Japanese?

Japanese counts every character; alphabetic languages count by the reduced rule. SceneMixer's menu shows the price for your language.

### Will the videos speak the script's language?

Only if you set the dialogue language in SceneMixer's step 1. It is a separate choice.

### Can I write the brief in English and get a Chinese script?

Yes. SceneMixer's script follows the interface language, not the brief.

### How many languages?

SceneMixer supports all sixteen interface languages; Traditional and Simplified Chinese both get a Chinese script.

Updated November 3, 2026


---

# AI Create: Writing the One-Line Brief for a Good Script | SceneMixer

URL: https://scenemixer.com/guides/writing-the-one-line-brief/
What to put in the AI Create box so the script comes out the way you want: genre, lead, conflict, ending, tone, and what to leave out. With before-and-after examples.

# Writing the One-Line Brief

Quick answer: On the AI Create tab, one or two sentences are enough if they carry five things: the genre, the lead (who, and what they want), the conflict (who or what stands in the way), the ending you want, and the tone. A revenge drama: a disowned heiress returns as the family firm’s new investor; her brother tries to bury the truth; she wins the company and forgives no one; cold and fast. That beats a story about a rich family every time. Inspire Me fills the box with an example to edit. Pick the length, press Start Writing.

Four ways to start if you are not sure AI Create is the tab you want.

## The five things

Genre: revenge, romance, cultivation, workplace, horror, comedy. It sets the rhythm and the kind of hook each episode ends on. Lead: one person, named or not, with a want. Conflict: the person or force against them; a story with no opponent drifts. Ending: even one word (she wins, they never meet again) keeps the episodes pointed somewhere. Tone: cold, warm, absurd, slow-burn. Optional: the setting and era, and the number of episodes.

## Before and after

Before: A love story in Shanghai. — After: A slow-burn romance in 1990s Shanghai: a night-shift radio host falls for a caller who never gives her name; his producer wants the segment cut; he finds her on the last night of the show; bittersweet.

Before: Cultivation, main character gets strong. — After: A cultivation revenge story: a sect’s kitchen boy is framed for a theft and cast out; he returns ten years later as a grandmaster to expose the elder who framed him; each episode ends on a reveal; ruthless.

## What to leave out

Camera directions, shot lists, character sheets — those come from SceneMixer's read and storyboard later. Long backstory: the script will invent what it needs. Several stories at once: one brief, one story; write again for another. Real people and brands: keep to invented names.

## Length and language

Pick the length under Length: from a mini at 1,000 characters to a long script at 20,000; the price on the button follows. The script is written in the language you are using the site in. Length and cost · Your language.

## If the first draft is off

Write again with a sharper brief; each writing is priced on its own and every draft stays in Writing History. Or take the draft into a project anyway: the summary, cast and storyboards are editable, and text edits are free. After the script is written.

| Put in | Example | Why |
| Genre | revenge drama | Sets the rhythm and the episode hooks |
| Lead + want | a disowned heiress who wants the firm back | The story has a centre |
| Conflict | her brother buries the truth | Episodes have something to push against |
| Ending | she wins and forgives no one | The episodes point somewhere |
| Tone | cold and fast | Dialogue and pacing match |

## FAQ

### How long should the brief be?

One or two sentences with the genre, lead, conflict, ending and tone. Longer is fine; vaguer is not.

### What does Inspire Me do?

SceneMixer fills the box with an example brief you can edit or replace.

### Should I describe the characters' looks?

No. SceneMixer's read writes detailed descriptions later, and you can edit them in step 2.

### Can I ask for a number of episodes?

Yes, in the brief. The length you pick on SceneMixer sets how much script there is to split.

### The draft is not what I meant.

Write again with a sharper brief. Every draft stays in Writing History.

Updated November 3, 2026


---

# Yearly Plan Credits Arrive Monthly: Why and What It Means

URL: https://scenemixer.com/guides/yearly-plan-credits-arrive-monthly/
Pay for a year and the plan is one payment but twelve monthly grants. What lands on payment day, what arrives each month, what happens on refund, upgrade or expiry.

# Why a Yearly or Quarterly Plan Gives You Its Credits Month by Month

Quick answer: A yearly or quarterly membership is billed once but its credits are granted per month: on payment day you get the first month’s allowance (if no other membership is running at the time) plus the sign-up bonus for the whole term, and each following month the next allowance lands on the same day. Your membership benefits (more tasks at once, the badge switch, member-only features) run for the whole term without gaps, and the expiry date shown is the end of the full term. The reason is simple: a year’s worth of credits handed over on day one is either spent in a week or sits unused, and neither matches how people actually produce. Monthly grants keep the balance steady. Credits already granted do not expire with the term.

Open the membership page after paying for a year: the current period shows this month’s grant and the date of the next one; the expiry shown is a year out.

## What you get on payment day and what comes later

On the day you pay: this month’s allowance (700 for Pro, 2,000 for Studio, 5,000 for Enterprise) and, on a first subscription, the one-time bonus counted on the whole term. Every month after that, on the same calendar day, the next allowance. The membership page shows which month you are in and when the next grant is due. What each tier includes: Membership Tiers Explained.

## Benefits and expiry

The concurrent-task limit, the badge switch and every other benefit apply continuously from payment to the end of the term; SceneMixer never switches anything off between monthly grants. The expiry date is the end of the whole term, not of the current month. The plan is not renewed automatically unless you took it as a subscription (card, PayPal, or Card / Google Pay on the Airwallex tab); a one-period purchase with a local method on the Airwallex tab does not renew; check the membership page for the renewal state.

## If you upgrade mid-term

Upgrading to a higher tier starts the higher tier now and grants the difference for the remainder of the current month, in proportion to what is left of it; the remaining months of the term switch to the higher tier’s allowance. Details and confirmation: Upgrading Mid-Cycle.

## If you are refunded

A full refund cancels the months that have not started and reclaims everything the order has granted, including the sign-up bonus. A partial refund reclaims nothing automatically and puts the membership on hold until support reviews it. Credits bought separately on SceneMixer are untouched. The rules: Refund Rules.

## Unused credits

Credits granted by a plan sit in the same balance as purchased credits and do not expire when the plan ends. Spending order and other balance questions: Do Credits Expire?.

| Moment | What happens |
| Payment day | First month's allowance + sign-up bonus for the whole term; benefits start |
| Same day each month | Next month's allowance |
| Upgrade | Higher tier now; difference for the rest of this month; later months at the new allowance |
| Full refund | Unstarted months cancelled; granted credits and bonus reclaimed |
| Partial refund | Membership on hold until support reviews it |
| End of term | Benefits end; granted credits stay |

## FAQ

### I paid for a year and only got one month of credits.

That is SceneMixer's design: one grant per month, the rest arrive on the same day each month.

### Is the sign-up bonus also monthly?

No. SceneMixer counts it on the whole term and grants it once, on payment day.

### Do the monthly credits expire at the end of the month?

No. On SceneMixer, anything granted stays in your balance.

### Do benefits pause between grants?

No. SceneMixer runs them continuously for the whole term.

### Does a quarterly plan work the same way?

Yes, SceneMixer makes three monthly grants instead of twelve.

Updated November 3, 2026


---

# “Your Edit Could Not Be Synced”: Every Reason and What to Do

URL: https://scenemixer.com/guides/your-edit-could-not-be-synced/
Editing a storyboard in your own language? The status line above a segment says when your edit did not reach the production copy. Each of the eight reasons and its fix.

# “Your Edit Could Not Be Synced”: What It Means and How to Fix It

Quick answer: When you read and edit a storyboard in a language other than the project’s working language, every save is translated back into the production copy — the version video generation uses. A status line above the segment tells you how that went. Your edit could not be synced to the production copy. The previous version is still used for generation. means the translation back failed, and the line after it says why: most often an @ name was changed, the shot count or seconds changed, or a line was separated from its voice tag. Fix that in your text and save again, or press Restore machine translation to throw your edit away and start again from the version generation uses.

Open an English project with the site in another language, click a segment and look above its storyboard text: while an edit is being written back you will see Saved. Syncing your edit to the production copy…

## Who sees these messages

Only people who read a storyboard in a different language from the one the project works in — for example, a Japanese interface on an English project. The storyboard you read and edit is a translation; the one video generation reads stays in the project’s working language. Every time you save, your change is translated back. With an English interface on an English project, and on the Chinese site, there is nothing to sync and none of these lines appear. Working in your own language · Interface, script and dialogue language.

## The status line, state by state

| Line above the segment | What is happening | Button |
| Translating to your language… | A new or changed segment is being translated for you to read. | — |
| Saved. Syncing your edit to the production copy… | Your save is being translated back. | — |
| Your edit is being synced to the production copy — video generation for this segment resumes in a moment. | You tried to generate while it was still syncing; generation waits for it. | — |
| Your edit could not be synced to the production copy. The previous version is still used for generation. + a reason | Your edit did not make it back. The video would still be made from the text before your edit. | Restore machine translation |
| Not translated + a reason | This segment could not be translated for you to read; the original is shown. | Retry |
| We couldn’t translate it into the production language, so your text is used as written for now. | Your text went into the production copy as you wrote it. | — |

## The eight reasons, and what to do about each

| Reason shown | What went wrong | What to do |
| An @ reference (the name after @) was changed or removed — keep those names exactly as they are. | A character, scene or prop chip was edited, retyped or deleted, so it no longer matches the project. | Put the chip back from the @ menu instead of typing the name; save again. |
| The number of shots or their seconds changed — edit the text inside a shot, not the shot headers. | A shot marker or its seconds was added, removed or changed in your language. | Put the shot markers and seconds back as they were and edit only the words inside the shots; or press Restore machine translation and redo the edit that way. |
| A dialogue line or its voice tag changed — keep each line together with its voice tag. | A line and the voice tag after it were split, or a tag was removed. | Keep every line followed by its voice tag; change only the words of the line. |
| The translation came back malformed. Please try again. | A one-off translation error. | Save again. |
| Translation is busy right now; it will retry automatically. | The translation service was busy. | Nothing; it retries on its own. |
| We couldn’t translate it into the production language, so your text is used as written for now. | Usually a segment you inserted and wrote yourself: it could not be translated back, so your own words were kept. | Nothing needed; it is translated again automatically when it can be. Check the segment before generating. |
| Syncing took too long. Save again to retry. | A timeout. | Save again. |
| Something went wrong. Save again to retry. | Any other error. | Save again; if it keeps happening, restore and redo the edit. |

## What Restore machine translation does

SceneMixer's Restore machine translation deletes the translation you edited for this segment and has the segment translated again, fresh, from the production copy — the version that video generation is still using. Your edit in your language is gone after that. Retry on a Not translated line does the same thing: it translates the segment again.

So before you press it, copy the change you wanted (the words, not the chips), restore, then make the change again carefully — inside the shots, keeping chips and voice tags. If the edit was small, fixing the cause and saving again keeps it without restoring.

## Editing in your language without breaking sync

- Insert names with SceneMixer's @ menu. A typed name is not a chip, and changing a chip’s text breaks the reference. The @ menu.
- Leave the labels alone. Shot markers, header fields and voice tags are shown as pills in edit mode; type between them, not inside them. Why storyboard labels turn into pills.
- Leave shot markers and seconds as they are. Adding a shot, removing one or changing seconds in the translated text is the change that most often fails to sync; the reason line itself says to edit the text inside a shot, not the shot headers.
- Copy and paste inside the editor, not from another document, so chips survive. Copy and paste without losing links.
- Wait for the line to clear before you generate. A generation started during syncing waits for it.

## Before you generate: which text will be used

If the segment shows Your edit could not be synced…, generating now makes SceneMixer build the video from the text before your edit. Fix or restore first, wait until the line clears, then generate. If it shows the used as written for now line, the video is made from your own words as typed. The five wait messages · Storyboard text history.

| You did | Likely reason | Fix |
| Retyped a character’s name | An @ reference was changed | Insert it from the @ menu |
| Added or deleted a shot, or changed seconds | The number of shots or their seconds changed | Put the shots back; edit only inside them |
| Moved a line away from its voice tag | A dialogue line or its voice tag changed | Keep each line with its tag |
| Nothing unusual | Busy, timeout or malformed | Save again, or wait |

## FAQ

### What is the production copy?

The version of the storyboard video generation reads, kept in the project's working language. SceneMixer translates the storyboard you read in your own language from it, and translates your saves back into it.

### Is my edit lost when sync fails?

Not at once: your text stays on screen. But SceneMixer's video generation keeps using the earlier version until the edit syncs. Fix the reason and save again, or restore and redo the edit.

### What does Restore machine translation do?

SceneMixer deletes your edited translation of this segment and translates the segment again from the version generation uses. Your edit is gone afterwards, so copy the words you wanted first.

### Why does changing seconds fail?

Shot markers and seconds have to match between the two versions exactly, so a change to them in the translated text does not sync. Edit only the words inside the shots in SceneMixer.

### Why does it say an @ reference was changed when I only fixed a typo?

The typo was probably inside a chip, or the name was retyped as plain text. Put the chip back from SceneMixer's @ menu.

### Can I generate while it is syncing?

You can press generate; SceneMixer holds the generation until the edit has synced, then starts it.

### Do I see these messages on the Chinese site?

No. On SceneMixer, they only appear when you edit a storyboard in a language different from the project's working language.

### Does the dialogue language change when I edit in my language?

No. The language the characters speak is a SceneMixer project setting. Editing in your own language changes only which language you read and write the storyboard in.

Updated November 3, 2026


---

# Your First Five Minutes: What Is on the Home Page | SceneMixer

URL: https://scenemixer.com/guides/your-first-five-minutes/
A tour of the home page after you sign in: the four tabs in the big box, the price on the button, Recent Projects, Featured Projects, the sidebar and the top right.

# Your First Five Minutes

Quick answer: The home page is one big box with four tabs — Upload Script, AI Create, Ad Video, Remake a video — and a button whose price updates as you type. Below it, Recent Projects shows your cards with Step n/3 on each, then Featured Projects you can open and browse. The left sidebar holds Home, Projects, Assets, Membership, Credits and Refer, with Feedback and Settings at the bottom. Top right: your credit balance, Upgrade, the bell, your avatar. Paste a script, press Get Started, and you are in step 1 of a project.

Which of the four tabs to use if you are not sure which fits.

## The big box

Upload Script: paste text or Upload File (TXT / DOCX / Markdown). AI Create: type one idea and a script is written for you. Ad Video: photos in, an ad out. Remake a video: a clip in, a storyboard out. The button at the bottom right shows ✦ and the price of the first step; it changes as the text changes. Upload and parse.

## The strip above

A promo line — First drama: concept sheets and episode-1 storyboard parsing, free — with Learn more. Close it with the cross on SceneMixer; it stays closed. What is free.

## Recent Projects and Featured Projects

Your projects as cards, newest first, each with a thumbnail, its name and Step n/3. Click one to resume where you stopped; hover for Rename and Delete Project. View all opens the full list. Below, Featured Projects are finished dramas you can open, browse and play, but not change. Using a sample.

## The sidebar

Home: this page. Projects: every project. Assets: every image, video and script you have made or uploaded, grouped by type. Membership: plans. Credits: packs and order history. Refer: your referral link. At the bottom, the question mark opens SceneMixer's Ask AI (the Ask window; feedback is sent from inside it) and the ⚙ opens Settings. Settings.

## Top right

Your balance (click for Available, Frozen, Credit History, Buy Credits), Upgrade, the bell for announcements and replies, and your avatar (nickname, ID, Terms, Log out). The Ask button in the corner answers questions about the site from these guides. Getting help.

## The on-page tips

The first time you open a page, a short tour points at its buttons. Close it any time; replay it from Settings → Quick tour → Replay.

| Area | What it is for |
| Four tabs | Start a project from a script, an idea, photos or a clip |
| Button with ✦ | The price of the first step, live |
| Recent Projects | Resume, rename or delete |
| Featured Projects | Open and browse finished dramas |
| Sidebar | Projects, Assets, Membership, Credits, Refer, Feedback, Settings |
| Top right | Balance, Upgrade, bell, avatar, Ask |

## FAQ

### Where do I start?

Paste a script into the box on SceneMixer and press Get Started; the button shows the price first.

### What is Step n/3 on a card?

On SceneMixer, the step shown on the card says how far that project got: 1 outline, 2 cast and sets, 3 episodes. Click to resume there.

### Can I edit a Featured Project?

No. SceneMixer opens them read-only, for browsing and playing.

### Where are my files?

Assets in SceneMixer's sidebar, grouped by type across every project.

### How do I see the tips again?

Settings, Quick tour, Replay.

Updated November 3, 2026


---

# Your Own Reference Images: What Changes When You Replace a Sheet

URL: https://scenemixer.com/guides/your-own-reference-images-what-changes/
Uploading your own image replaces the sheet the video looks at. What the swap changes, what stays (description, existing clips), and how to prepare the image.

# I Already Have Reference Images for My Characters: What Changes When I Upload Them

Quick answer: Yes, you can use your own images: open the character, scene or prop in Step 2 and press Upload & Replace. After you pick a file, a dialog asks how to use it: Optimize into a reference sheet has AI redraw it as a clean sheet (priced like an image), Use this image as is makes the photo itself the sheet. Whichever you choose, every video generated for that character looks at that image from then on instead of the old sheet. The description under the card does not change — it still steers how the shot is written, so bring it in line with the picture. Clips you generated earlier are untouched; state images made from the old sheet are treated as out of date until regenerated. Prepare the image the way a sheet is made: one person, plain background, 9:16 for people and 16:9 for places and props.

Open any character card in Step 2 and look for Upload & Replace next to Regenerate; the current sheet stays until you choose how the upload is used.

## How to replace a sheet

In Step 2, open the card and choose Upload & Replace. JPEG or PNG, a sensible shape (roughly 9:16 for a person, 16:9 for a place or a prop; very tall or very wide images are refused later at generation). After the file is in, a dialog asks how to use it, and the sheet does not change until you choose. Optimize into a reference sheet has AI redraw it as a clean white-background sheet together with the description — the redrawn sheet becomes the reference for the shots, and it is priced like an image. Use this image as is makes the photo itself the sheet, exactly as uploaded; choose it when the image is already sheet-like, since a snapshot brings its background, light and pose into every shot that uses it. Either way the sheet and its thumbnails are replaced everywhere the character appears. Uploading is a paid-plan feature; the details and the alternatives: Use Your Own Photo for a Character or Scene.

## What changes and what does not

Changes: the reference the video model looks at for that character in every segment generated from now on — your exact photo under Use this image as is, or the redrawn sheet under Optimize into a reference sheet; the card, the panel thumbnails, the face crop used in the editor. Does not change: the description (appearance, clothing, voice) — the storyboard and the shot text still read it; clips already generated; other characters. Goes stale: state images (costume changes and the like) that were generated from the old sheet count as out of date on SceneMixer and fall back to text until you regenerate them.

## Bring the description in line

The video is built from two things: the image, and the words. If your image shows a blonde woman in a leather jacket and the description still says a brunette in a coat, the shot text will describe the coat and the model will argue with itself. Edit the appearance and clothing fields to match the picture; keep the identifying features you want kept (a scar, glasses). Write a Description That Works.

## Preparing an image that works as a sheet

A generated sheet is a neutral reference on purpose: full body, plain background, even light, no mood. Your image should be the same kind of thing. Crowded backgrounds, heavy filters, dramatic lighting and other people in frame all leak into every shot that uses it. On SceneMixer, real people need permission — your own face is fine, a stranger’s is not: Content Rules and Moderation.

## Using the same image across projects

An uploaded sheet lives in the project. To reuse it elsewhere, open the matching card in the other project and press Pick existing: the whole entry moves over — the uploaded image, the name, the description and the voice — so the card’s description comes with it and does not have to be written again. Reuse a Character, Scene or Voice Across Projects.

| After you upload | Generated sheet | Your image |
| New segments look at | — | Your image, or the sheet optimized from it |
| Description | Kept | Kept — edit it to match |
| Existing clips | Unchanged | Unchanged |
| State images from the old sheet | — | Out of date until regenerated |
| Thumbnails and face crop | — | Refreshed |

## FAQ

### Do I need a plan to upload?

Yes, on SceneMixer, uploading a replacement is a paid-plan feature; the free path is to edit the description and regenerate.

### Will the face be exactly my image?

The model treats it as a strong reference; likeness on SceneMixer is good but not a copy, and lighting and angle follow the shot.

### Can I upload several angles?

SceneMixer keeps one sheet per character. Use states for a second look (a costume) rather than a second angle.

### My upload was accepted but the video failed with 'invalid input'.

The image's shape or size is outside what the video tier accepts. Crop closer to 9:16 and re-upload.

### Does the upload count against my free sheets?

No. On SceneMixer, free sheets count generations; an upload is not one.

Updated November 3, 2026


---

# Your Scene Image Becomes the Set: How AI Builds a 360° Room

URL: https://scenemixer.com/guides/your-scene-image-becomes-the-set/
A location’s scene image is opened up into a 360° set and measured, then every shot in that scene is filmed inside it. How it works, and what makes a good scene image.

# Your Scene Image Becomes the Set

Quick answer: Every location in your project has a scene image. On projects with the world model, that one picture becomes the set: it is opened up into a 360° view, and the room is measured (walls, ceiling, doors and window openings, where the furniture stands). Every shot in that scene is then placed inside it. Outdoor scenes get a lighter set: the ground, nearby objects and the distance in four directions. A room is measured once and reused across shots and episodes; make a new scene image and the room is measured again from the new picture. There is nothing extra to make and no extra credits.

Open step 2 of a project and look at the Scenes tab: each scene image there is the starting point for that location’s set. Then, in the editor, select a segment set there that has not been generated, and drag around the preview.

## From one picture to a whole set

- The scene image. The picture of the location you already have in step 2.
- Opened up to 360°. The image shows one direction; the set needs all of them. The rest of the room is filled in to match the picture, once, and then kept.
- Measured. Walls, ceiling, door and window openings, and where each piece of furniture stands.
- Used for the scene. Characters get places in it, and every shot’s camera is placed inside it.

## What the measurements are for

- Placing people. “Sits at the round table” needs to know where the table is. “Stands in the doorway” needs the door.
- Placing cameras. SceneMixer places each camera inside the room, at a real distance, and keeps it from looking through a wall.
- Keeping the background. A reverse shot shows the part of the room that is really behind the person, so the window and the door stay put.
How positions are written so they can use the room: Write positions the world can read.

## Indoors and outdoors

Rooms get the full treatment: walls, ceiling, openings and furniture. Outdoor scenes get a lighter set: the ground people stand on, the objects close by, and the distant view in four directions. There is less to measure outdoors, so anchor people to things that are really there, such as a bench, a car or a doorway.

Vehicles are their own case: Scenes inside a car, bus, train or plane.

## Measured once, used everywhere

SceneMixer measures a room once and reuses it for every shot in every segment set there, across episodes. The office in episode 1 is the same measured office in episode 6. That is a large part of why a location looks like the same place each time it comes back; Props and locations in a shot covers the rest.

## If you change the scene image

- A new scene image means a new room. The set is opened up and measured again from the new picture, and segments generated from then on use it.
- Finished segments keep their look. Clips already made were filmed in the old room. Regenerate them if they need to match; each regeneration is charged as usual.
- Fix the description first. If the image is wrong because the description is, change the description, then make the image. Scene image does not match the description.

## What makes a good scene image for a set

- Show the space. A view of the room with floor and walls visible gives the set more to measure than a close-up of one object.
- Include what the scene uses. If they sit at a table and leave through a door, the table and the door should be in the picture.
- Keep people out of it. Characters come from their own images. Anyone painted into the scene image becomes part of the set.
- Match the story’s place. An office for the office scenes, a kitchen for the kitchen; one image per location, not one for everything.
- Make it in the project’s style. The set takes its look from this picture, so a scene image in the wrong style carries that into every shot.

## Which projects build sets

SceneMixer builds a set for each scene automatically in new projects on paid plans, with no extra credits. Scene images themselves are priced as before. What a paid plan changes. Projects created before the world model launched, and projects on the free plan, still use the scene image as the location’s reference with the earlier automatic staging. More on the idea: What is a world model in AI video?

| The scene image shows | What the set gets from it |
| The whole room, floor and walls visible | Walls, ceiling, openings and furniture to measure |
| The table, bed or door the scene uses | Real places to put people and send them to |
| One corner or one object, close up | Little room to measure; most of the set is filled in |
| People in the picture | Those people become part of the set |
| An outdoor place | The ground, nearby objects and the distance in four directions |

## FAQ

### Do I have to make anything extra for the set?

No. SceneMixer builds the set from the scene image you already have. There is no extra step and no switch.

### Does building the set cost credits?

No. SceneMixer runs it on new projects on paid plans with no extra credits. The scene image itself is priced as before.

### Why does the set show parts of the room my image did not?

An image shows one direction and the set needs all of them, so the rest of the room is filled in to match the picture. SceneMixer fills it in once and then keeps it, so it stays the same in every shot.

### I made a new scene image. What happens?

The room is measured again from the new picture, and segments generated from then on use it. Finished segments on SceneMixer keep the old room until you regenerate them.

### Can I edit the set directly?

No. Change the scene image, or its description and then the image. SceneMixer's preview of the set is read-only.

### Does it work outdoors?

Yes, SceneMixer builds a lighter set: the ground, nearby objects and the distant view in four directions. Anchor people to things that are really there.

### Is the same set used in later episodes?

Yes. SceneMixer reuses a measured room across shots and episodes, as long as the scene image stays the same.

### My older project has no 360° set. Why?

Projects created before the world model launched keep the earlier automatic staging. Start a new project on a paid plan on SceneMixer to get sets.

Updated November 3, 2026


---

# Your Uploaded Clip in a Remake: Where It Goes and What It's Used For

URL: https://scenemixer.com/guides/your-uploaded-clip-in-a-remake-where-it-goes-and-what-it-is-used-for/
A remake starts from a clip you upload. What the site does with it, what it never does (feed it to the video model, share it, train on it), and when it is deleted.

# The Clip You Upload for a Remake: Where It Is Stored, What It Is Used For, and What It Is Never Used For

Quick answer: The clip you upload is used for one thing: to be read. The site cuts it into shots, recognises the people, places and props, listens for the spoken lines, and writes an episode outline from that. Everything you then generate — sheets, storyboard, clips — is made from the reading, not from the footage. The footage is never handed to the video model as a reference, so the result does not copy its frames and you are not charged twice for its seconds. It is stored under a private, unguessable address, is not shown to other users, is not used to train anything, and is queued for deletion as soon as the read completes, then permanently removed about a week later. You are responsible for having the right to use the clip; the read does not check ownership.

Compare a remake’s clips with the source: same cast, places and lines, but the framing and the faces are the site’s own; nothing is lifted from the footage.

## What the clip is used for

Reading: cutting into shots, describing each shot, identifying the recurring people and places, listening to each shot for dialogue, noting whether there is music. The outputs are text and a few still frames used during the read; from there the project is an ordinary script project. Source Video Limits, Read Time and Price.

## What it is never used for

Not as a reference video for generation: the video models can take reference clips, and SceneMixer deliberately does not pass yours. That is why remakes do not reproduce the source’s frames and why you pay for the read by the minute and for the output by the second, never both on the same seconds. Not shown to anyone else, not on any public page, not in the sample gallery. Not used to train models.

## Where it is stored

In the site’s storage under an address that cannot be guessed and is not listed anywhere; pieces cut during the read (shots, the audio track, a lower-resolution working copy) are stored the same way. During the read, a working copy may also be placed, under an unguessable address, where the reading service can fetch it; it is removed after the read.

## When it is deleted

The source clip does not wait for the project: SceneMixer queues it for deletion as soon as the read completes and permanently removes it about a week later. The project’s own files go the same way when you delete the project, and deleting your account does that for every project. The Seven-Day Window.

## Rights

Upload only footage you have the right to use. The read does not check ownership, and the result is your project under your account; a remake of someone else’s film is your responsibility, and the content rules apply to what you generate from it. Content Rules.

| Question | Answer |
| Is the footage fed to the video model? | No. Only the reading is used. |
| Are frames copied into my clips? | No. Clips are generated from text and your sheets. |
| Can other users see it? | No. |
| Is it used for training? | No. |
| When is it deleted? | Queued when the read completes; gone about a week later. |
| Who checks I own it? | Nobody; that is your responsibility. |

## FAQ

### Can I download the source back from the site?

SceneMixer does not offer it as a download; keep your original.

### Why does the remake look different from the source?

Because SceneMixer generates it from the reading, with your cast sheets and styles, not from the frames.

### Can I re-read the same clip with different settings?

SceneMixer reads a clip once per project; start a new remake for a new read.

### Are the still frames from the read kept?

A few are used during the read; SceneMixer does not show them anywhere in your project.

### Does the site check copyright?

No. SceneMixer does not check copyright, so upload only what you have rights to.

Updated November 3, 2026


---

# SceneMixer Compared with Other AI Video Tools | SceneMixer

URL: https://scenemixer.com/alternatives/
One page per tool: how SceneMixer compares with other AI video tools for episodic stories, with sourced, dated facts and when the other tool fits better.

Alternatives

# SceneMixer Compared with Other AI Video Tools

Updated November 2, 2026

Quick answer: Each page here compares SceneMixer with one other AI video tool for people making stories in episodes. Facts about the other tool come only from its own public pages, linked and dated; SceneMixer's prices come straight from our live pricing; and every page says when the other tool is the better fit.

## Comparisons

- Adobe Firefly Alternative for AI Drama Series — Adobe Firefly generates video with its own and partner models inside Adobe plans; SceneMixer turns a script into an episodic drama. How they compare.
- AI Video Tools Compared: 47 Tools Next to SceneMixer — 47 AI video tools in one table, in their own words and dated: what each is built around and what it costs to start, next to SceneMixer. Not a ranking.
- Alici AI Alternative for AI Short Drama — Alici plans shots with an agent and generates clips with many models, from $12 a month; SceneMixer turns a script into an episodic drama. When to pick which.
- Boords Alternative for AI Storyboard to Video — Boords is storyboard and animatic software with client review; SceneMixer renders the storyboard into video episodes. How they compare, and when Boords fits.
- Canva AI Video Alternative for Drama Series — Canva's Create a Video Clip makes 8-second clips inside a design suite; SceneMixer turns a script into an episodic drama. How they compare, and when Canva fits.
- CapCut AI Video Alternative for Drama Series — CapCut is a video editor with an AI video maker; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when CapCut fits.
- Dashtoon Studio Alternative for AI Story Video — Dashtoon's creation studio shut down on August 15, 2026; its site still hosts comics. If you made stories there, here is how SceneMixer picks the work up.
- DramaSo Alternative for AI Drama Series — DramaSo breaks down viral clips and previews dialogue as audio first; SceneMixer turns a script into episodes with a fixed cast. How they compare.
- Dramily Alternative for AI Short Drama — Dramily takes a story idea to scripts, locked assets, storyboards and video prompts for short dramas; SceneMixer renders the episodes. How they compare.
- Freepik (Now Magnific) Alternative for AI Drama — Freepik now trades as Magnific: stock content plus many AI video models on yearly credits. SceneMixer turns a script into an episodic drama. How they compare.
- Gemini Omni Alternative for AI Drama Series — Gemini Omni Flash makes and edits short clips in conversation; SceneMixer turns a script into an episodic drama. What Google states, and when Omni fits.
- Google Flow Alternative for AI Drama Series — Google Flow is Google's AI creative studio with Veo and Nano Banana; SceneMixer turns a script into an episodic drama. How they compare, and when Flow fits.
- Grok Imagine Alternative for AI Drama Series — Grok Imagine generates and edits videos up to 15 seconds; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when Grok fits.
- Hailuo AI Alternative for AI Drama Series — Hailuo AI makes video clips with MiniMax models from $14.99 a month; SceneMixer turns a script into an episodic drama. How they compare, and when Hailuo fits.
- Hedra Alternative for AI Drama Series — Hedra offers visual models and the infrastructure to run them; SceneMixer turns a script into an episodic drama. How they compare, and when Hedra fits.
- HeyGen Alternative for AI Story Video — HeyGen makes avatar presenter videos and translations; SceneMixer turns a script into an episodic drama with a fixed cast. When to pick which.
- Higgsfield Alternative for AI Drama Series — Making a series? How SceneMixer compares with Higgsfield on workflow, consistent characters, dialogue and price, and when Higgsfield is the better fit.
- invideo AI Alternative for AI Drama Series — invideo AI pairs agents with an editor and stock footage; SceneMixer generates an episodic drama from a script. How they compare, and when invideo fits better.
- JoggAI Alternative for Story Video With Characters — JoggAI makes avatar-led videos for marketing, training and podcasts, with a free plan; SceneMixer makes dramas with a cast that acts. When to pick which.
- Katalist Alternative for Script to Storyboard to Video — Katalist turns a script into a storyboard and then into ad videos, from $19 a month; SceneMixer does the same for episodic drama. When to pick which.
- Kling AI Alternative for AI Drama Series — Kling AI makes clips with its own models; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when Kling fits better.
- Kling 4.0: Specs, Release Status and an Alternative — Kling 4.0 was announced on September 28, 2026: 3-30 second takes, 4K, up to 10 keyframes. What was announced, what is missing, and where a series pipeline fits.
- Kling 4.0 Flash: 720p, 20 Seconds, Who It Suits — Kling 4.0 Flash is the announced 720p, 20-second model built for volume. Its specs, what it leaves out, and who should wait for the full Kling 4.0 instead.
- Krea Alternative for AI Drama Series — Krea is a creative suite for images, video, 3D and LoRA training; SceneMixer turns a script into an episodic drama. How they compare, and when Krea fits.
- Leonardo AI Alternative for AI Drama Series — Leonardo.Ai is an image-first platform with video and animation on token plans; SceneMixer turns a script into an episodic drama. How they compare.
- LlamaGen Alternative for Comic Stories as Video — LlamaGen's comic-to-video mode animates finished comic panels; SceneMixer turns a comic script into an acted animated episode. When to pick which.
- LTX Studio Alternative for AI Drama Series — LTX Studio is a suite for storyboards, pitch decks and video; SceneMixer turns a script into episodes. How they compare, and when LTX Studio fits better.
- Luma Alternative for AI Drama Series — Luma pairs its Ray models and Luma Agents with third-party models; SceneMixer turns a script into an episodic drama. How they compare, and when Luma fits.
- MagicLight Alternative for AI Drama Series — MagicLight makes long story videos; SceneMixer makes episodic dramas with a fixed cast. How they compare on format, voice and price, and when MagicLight fits.
- Medeo Alternative for AI Drama Series — Medeo creates videos through chat with an editor and stock footage; SceneMixer turns a script into an episodic drama. How they compare, and when Medeo fits.
- Midjourney Video Alternative for AI Drama Series — Midjourney animates its images into 5-second clips you can extend; SceneMixer turns a script into an episodic drama. How they compare, and when Midjourney fits.
- Mootion Alternative for AI Drama Series — Mootion turns ideas and scripts into voiced visual stories; SceneMixer turns scripts into episodic dramas. How they compare, and when Mootion fits better.
- Morphic Alternative for AI Drama Series — Morphic gives you models and workflows to bring stories to life; SceneMixer turns a script into an episodic drama. How they compare, and when Morphic fits.
- MovieFlow Alternative for AI Drama Series — MovieFlow makes short AI videos from a line of text with an agent and several models; SceneMixer turns a script into an episodic drama. How they compare.
- Neolemon Alternative for Animated Story Video — Neolemon makes consistent cartoon characters and print-ready picture books; SceneMixer turns a story into animated episodes that speak. When to pick which.
- OpenArt Alternative for AI Drama Series — OpenArt is an image and video studio with Director and consistent characters; SceneMixer turns a script into episodes. How they compare, and when OpenArt fits.
- Pika Alternative for AI Drama Series — Pika gives you many models for quick clips; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when Pika fits better.
- Pippit Alternative for AI Short Drama — Pippit's drama tool makes videos from a script or prompt with a choice of models; SceneMixer builds an episodic drama from one script. When to pick which.
- PixVerse Alternative for AI Drama Series — PixVerse offers its own models plus many others, with daily free credits; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare.
- Pollo AI Alternative for AI Drama Series — Pollo AI gives you its own and third-party models for quick image and video generations; SceneMixer turns a script into an episodic drama. How they compare.
- Revid AI Alternative for AI Drama Series — Revid AI creates, publishes and remixes short-form videos for growth; SceneMixer turns a script into an episodic drama. How they compare, and when Revid fits.
- Runway Alternative for AI Drama Series — Runway is a generation and editing platform; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when Runway fits better.
- Showrunner Alternative for Making AI Shows — Showrunner lets you watch AI shows and make scenes inside them; SceneMixer builds your own series from a script with a fixed cast. When to pick which.
- Sora Alternative for AI Drama Series — OpenAI discontinued the Sora app in April 2026 and the Sora API in September 2026. If you made stories with Sora, here is how SceneMixer picks up the work.
- StoryShort Alternative for AI Drama Series — StoryShort automates faceless short videos and auto-posts them; SceneMixer turns a script into an episodic drama. How they compare, and when to stay.
- Topview Alternative for AI Drama Series — Topview Drama Studio and SceneMixer both turn scripts into episodes. How they differ on languages, credits and price, and when Topview is the better fit.
- VEED Alternative for AI Drama Series — VEED is a browser video editor with an AI studio of many models; SceneMixer turns a script into an episodic drama. How they compare, and when VEED fits.
- Veo Alternative for AI Drama Series — Veo is Google's video model with native audio; SceneMixer turns a script into an episodic drama. How they compare, and when Veo is the better fit.
- Vidnoz Alternative for AI Story Video — Vidnoz makes avatar presenter videos from templates, with a free daily allowance; SceneMixer turns a script into an episodic drama. When to pick which.
- Vidu Alternative for AI Drama Series — Vidu makes 3–16 second clips with multi-reference consistency; SceneMixer turns a script into an episodic drama. How they compare, and when Vidu fits.

## How these pages are written

- One tool per page. We compare SceneMixer with one product at a time; we do not rank tools or score them.

- Their facts, their words. Every statement about another product links to that product's own page and shows the date we checked it.

- Our prices are live. SceneMixer's numbers are filled in from the same source as the pricing page each time these pages are rebuilt.

- When to stay. Each page says when the other tool fits better.

- Rechecked regularly. Facts are rechecked at least every 45 days and after any price or plan change we notice. Corrections: support@scenemixer.com.

## What SceneMixer is for

SceneMixer turns a script or novel into a series: it reads the story into a cast, locations and an episode outline, draws reference art once, writes a storyboard for each episode, renders it segment by segment against the same character sheets, and joins each episode into one file. See the short drama generator, sample dramas and the guides.

Sample dramas made end to end in SceneMixer · see them all


---

# Adobe Firefly Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/adobe-firefly/
Adobe Firefly generates video with its own and partner models inside Adobe plans; SceneMixer turns a script into an episodic drama. How they compare.

Alternatives

# An Adobe Firefly Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Adobe Firefly is Adobe's generative AI workspace for images, video and audio: its video generator turns a prompt or an image into clips with the Adobe Firefly Video Model and partner models — the page lists Google's Veo 3.1, Luma AI Ray3, Kling 3.0 and Runway — and an editor to arrange, trim, extend and re-voice them. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Firefly for clips and images inside the Adobe ecosystem; choose SceneMixer to produce a series with a fixed cast.

## A workspace inside Adobe, or a series pipeline

Adobe sizes Firefly plans in generative credits a month, with a count of five-second videos on each paid tier and a free tier of daily generations1. SceneMixer starts from the story: the outline decides the episodes, and every segment is rendered against the same character sheets.

## At a glance

Adobe Firefly facts as stated on its own pages, checked November 2, 2026

| | Adobe Firefly | SceneMixer |

| Built around | A generative AI workspace for images, video and audio, with partner models beside the Adobe Firefly Video Model12 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Video | A text prompt or an image (first frame optional) to a clip; refine, trim, reorder, swap models mid-edit, regenerate or extend any segment in its editor2 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Partner models | Google Veo 3.1 (and Veo 3.1 Fast), Luma AI Ray3, Kling 3.0 and Runway listed beside Adobe's own2 | Tiers by resolution and price per second; the model behind each tier is managed for you |

| Plans | Standard 2,000 generative credits a month, Pro 4,000, Pro Plus 10,000, Premium 50,000; shown in local currency (A$16.49, A$32.98, A$79.99 and A$329.99 a month incl. GST when we checked)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Videos per plan | Up to 20, 40 and 100 five-second videos a month on the first three plans; unlimited access to the Firefly Video Model on Premium1 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| Free | Free daily generations across image, video and audio, with models such as Nano Banana and Luma Ray3.14; limits refresh every day1 | 50 free credits and 15 free reference images for new accounts |

| Commercial use | “Videos generated with the Adobe Firefly model are safe for commercial use”, as Adobe's video page puts it2 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Eagle's Vow”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Adobe Firefly if…

- You already work in Adobe apps — its plans list Photoshop on web and mobile, Adobe Express Premium and Adobe Fonts1.

- You want Adobe's own video model with its statement on commercial use, next to partner models in one workspace2.

- You want to compare models on one prompt — Adobe says you can swap models inside a workflow or try one prompt across models2.

- You also need audio tools — speech, sound effects, music and audio and video translation are listed on its plans1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Adobe Firefly lands in SceneMixer

| In Adobe Firefly | Closest step in SceneMixer |

| Generate video from a prompt or image | A storyboard segment, rendered on a video tier. How to write one shot |

| Assembler and the video editor | The episode editor, with one card per segment on the timeline. The episode editor, a tour |

| Generate sound effects and music | A built-in score written per episode. Background music |

| Extend any segment | The next segment opens where the last one ended. Shot-to-shot continuity |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; prompts you wrote for single clips can become shots in the storyboard.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer an Adobe Firefly alternative?

For turning a written story into a multi-episode drama with a fixed cast, yes. For images, clips and edits inside Adobe's apps, Firefly is built around that.

### How much does Adobe Firefly cost?

On November 2, 2026, its plans page listed Firefly Standard with 2,000 generative credits a month, Pro with 4,000, Pro Plus with 10,000 and Premium with 50,000, priced in local currency (A$16.49, A$32.98, A$79.99 and A$329.99 a month for us). Check adobe.com for your region.

### Is Adobe Firefly free?

Its plans page, checked November 2, 2026, said you do not need a plan to get free daily generations across image, video and audio, which refresh every day.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15. Native dialogue vs dubbing.

## Sources and corrections

- Compare plans that include generative AI – Adobe Firefly (adobe.com/products/firefly/plans.html) — plans, monthly generative credits, videos per plan, free daily generations
- AI Video Generator – Adobe Firefly (adobe.com/products/firefly/features/ai-video-generator.html) — inputs, models, editing, commercial-use statement

Facts about Adobe Firefly come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Adobe Firefly's site before you buy. Adobe Firefly is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# AI Video Tools Compared: 47 Tools Next to SceneMixer | SceneMixer

URL: https://scenemixer.com/alternatives/ai-video-tools-compared/
47 AI video tools in one table, in their own words and dated: what each is built around and what it costs to start, next to SceneMixer. Not a ranking.

Alternatives

# 47 AI video tools next to SceneMixer, in one table

Updated November 2, 2026

Quick answer: 47 AI video tools next to SceneMixer in one table, each described in its own words as of the date shown: what it is built around and what it costs to start. The table does not rank or score anything; the one-to-one page behind each row says when that tool is the better fit.

## 47 tools, one table

The first two columns quote each tool's own page; the date on each cell is when we checked it. SceneMixer's row is filled in live from its pricing data.

| Tool | Built around | Price to start | Compared with SceneMixer |

| SceneMixer | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. | Pricing |

| Adobe Firefly | A generative AI workspace for images, video and audio, with partner models beside the Adobe Firefly Video Model12 | Standard 2,000 generative credits a month, Pro 4,000, Pro Plus 10,000, Premium 50,000; shown in local currency (A$16.49, A$32.98, A$79.99 and A$329.99 a month incl. GST when we checked)1 | Adobe Firefly vs SceneMixer |

| Alici AI | A Video Super Agent that plans shots and generates clips; image and video studios; a short-drama use case1 | Basic $12/month (150 credits, up to 720p); Pro $25 (500 credits); Creator $59 (1,440); Ultimate $149 (3,800); first-month discounts shown; yearly “save 30%”2 | Alici AI vs SceneMixer |

| Boords | Pre-production: scripts, storyboards and animatics, shared with clients for feedback and sign-off2 | Solo $39/month ($26 billed yearly), one user; Pro $75 ($50), up to 5 users; Team $125 ($85), up to 10; Agency $250 ($165), up to 301 | Boords vs SceneMixer |

| Canva | A design suite for presentations, social posts and video; Create a Video Clip generates a clip from text12 | Free; Pro US$144 a year; Business US$250 a year per person; Enterprise on request. AI tools share a monthly allowance that varies by plan; an AI Pass add-on adds more1 | Canva vs SceneMixer |

| CapCut | A video editor (desktop, online, mobile) with an AI video maker, templates, captions and effects12 | Free; Pro 850 monthly AI credits and 1 TB of storage; Ultra 12,179 monthly AI credits; Teams per seat. Shown in local currency (AUD 37.99 and 111.99 a month for Pro and Ultra when we checked)1 | CapCut vs SceneMixer |

| Dashtoon | Dashtoon Studio shut down on 15 August 2026 and is no longer available; creation continues on Frameo, its new platform for AI video and short drama1 | Not stated on the pages we read: dashtoon.com/pricing returned “page could not be found”2 | Dashtoon vs SceneMixer |

| DramaSo | Break down a trending clip into a beat sheet, or paste your own script1 | Free: 80 welcome credits once; Pro $39/month with 3,000 credits (“≈ 55s of finished film” on the default model)1 | DramaSo vs SceneMixer |

| Dramily | One SaaS for short-drama production: scripts, reusable assets, storyboards, video prompts, previews, covers and export-ready episodes1 | Free: 50 evaluation credits. Creator $49/month with 1,250 credits; Studio $99 with 2,600; Producer $199 with 5,300; yearly saves 10%1 | Dramily vs SceneMixer |

| Freepik | A creative platform: 250M+ stock photos, videos, vectors and PSDs plus AI image, video and audio tools across many models1 | Premium 240K credits a year; Premium+ 600K, with unlimited generations on some models; Pro Starter 1.5M; credits valid for 1 year; shown in local currency (A$18.50, A$45 and A$109.50 a month billed annually when we checked)1 | Freepik vs SceneMixer |

| Gemini Omni | A video model for high-speed generation and conversational editing, used through the Gemini API1 | Not stated in the release notes (checked November 2, 2026)1 | Gemini Omni vs SceneMixer |

| Google Flow | Google's models: Gemini Omni, Nano Banana and Veo 3.1 (native audio); an agent and user-built tools such as a Storyboard Studio1 | Google AI Plus $4.99/month, Google AI Pro $19.99, Google AI Ultra $99.99 and $199.99; prices vary by market1 | Google Flow vs SceneMixer |

| Grok Imagine | Video generation in xAI's Grok, with an API for developers1 | Per second, by duration and resolution; rates on xAI's pricing page, not on the documentation page (checked November 2, 2026)1 | Grok Imagine vs SceneMixer |

| Hailuo AI | A video and image generator from MiniMax: “native multimodal generation, precise multimodal editing, production-ready content creation”2 | Free: one-time trial credits, Hailuo models only. Standard $14.99 with 1,000 credits a month; Pro $54.99 with 4,500; Master $119.99 with 10,500; Max $184 billed yearly with 27,000; discounts shown on the page1 | Hailuo AI vs SceneMixer |

| Hedra | “Real-world AI, from research to runtime”: inference, sandboxes and model training for developers1 | Pay-as-you-go API billing, “no subscription, no seats, no minimum”: video generation from 1.22¢ a second, image generation from 0.59¢ a generation2 | Hedra vs SceneMixer |

| HeyGen | “Create, localize, scale, and collaborate on video, no camera or crew needed”: avatars, voices, translation and team tools1 | Free: 3 watermarked videos a month. Creator $29/month ($24 billed yearly) with 600 credits and 1080p. Pro $49 ($41 billed yearly) with 1,000 credits and 4K. Business $149 plus $20 a seat1 | HeyGen vs SceneMixer |

| Higgsfield | An AI-native creative suite with 35+ models and tools1 | Free plan with limited use; Starter $19/month billed yearly (270 credits a month); Plus $59/month or $47 billed yearly (1,200); Ultra $129 or $99 billed yearly (3,000)2 | Higgsfield vs SceneMixer |

| invideo AI | AI agents and generative models plus an editing timeline, stock footage and your own footage1 | Starter $20 per seat a month billed monthly ($10 billed annually), 400 credits; Plus $60 ($30), 2,000 credits1 | invideo AI vs SceneMixer |

| JoggAI | AI avatars presenting a script: marketing, training, podcast and product videos; talking photos; translation1 | Free $0 (200 credits, 31 days); Starter from $24/month billed yearly (144,000 credits a year); Creator $40.66/month; Team $215/month for 3 seats2 | JoggAI vs SceneMixer |

| Katalist | Script to storyboard to video for marketers and agencies; product swaps into winning creatives; UGC and promo videos1 | Essential $19/month; Pro $39/month; Unlimited $99/month (2 seats); Enterprise on request; an annual block lists different figures2 | Katalist vs SceneMixer |

| Kling AI | Kling 4.0 (“dynamic motions, Omni reference, native 30s-generation and multiple keyframes”) and Kling 3.0 / 3.0 Omni (“synchronized audio-video generation, intelligent storyboarding and element reference”)1 | Free Basic plan; Standard $10/month list, $6.99 for the first month then $8.80 a month, 660 credits a month1 | Kling AI vs SceneMixer |

| Krea | Image, video, 3D and lip-sync models with real-time generation, Nodes workflows and the Krea Agent1 | Free: 100 units a day; Pro $35/month or $21 billed yearly with 20,000 units a month1 | Krea vs SceneMixer |

| Leonardo.Ai | “Create across image, video, design, and motion”: one subscription, 80+ models1 | Free: 150 fast tokens a day, Auto model only, public creations. Essential $12 a month (8,500 tokens), Premium $30 (25,000), Ultimate $60 (60,000); yearly billing up to 20% off; prices ex. tax1 | Leonardo.Ai vs SceneMixer |

| LlamaGen | AI comic, manga and anime art generation; a Comic to Video mode for finished panels12 | Standard $18/month (15,000 credits); Pro $30 (45,000); Unlimited $96 (135,000); Production Unlimited $200 (270,000); annual saves up to 40%1 | LlamaGen vs SceneMixer |

| LTX Studio | “Professional creator, marketer or filmmaker” plans with storyboards, pitch decks and node-based Flows1 | Free: 800 one-time credits; Lite $15/month ($12 billed yearly) with 8,000 credits; Standard $35/month ($28 billed yearly) with 28,0001 | LTX Studio vs SceneMixer |

| Luma | Its Ray models plus third-party image and video models, and Luma Agents1 | Plus $30/month ($25 billed yearly) with 10,000 credits and commercial use; Pro $90 ($75) with 40,000; no free plan listed1 | Luma vs SceneMixer |

| MagicLight | Long-form story videos, “up to 50 minutes with steady characters, scenes, and tone”1 | Standard: 8,000 credits a month, “up to 130 min video generation” and “up to 290 s image-to-video generation”1 | MagicLight vs SceneMixer |

| Medeo | Creating videos using chat, with a video editor and collaborative editing tools1 | Free plan; Pro $39/month ($28 billed annually) with 700 credits (≈ 3 min of full AI video)1 | Medeo vs SceneMixer |

| Midjourney Video | Animating Midjourney images, or an uploaded image used as a start frame, on the web1 | A video job costs about 8 times an image job, per the announcement; usage comes from the Midjourney plan1 | Midjourney Video vs SceneMixer |

| Mootion | Turning text, ideas, images and scripts into cinematic visual stories1 | Standard $15/month billed monthly ($12 billed yearly), 1,500 credits; EDU plan $6/month ($5 billed yearly)2 | Mootion vs SceneMixer |

| Morphic | All models plus workflows; shared credits and extra seats on Pro and Max1 | Basic $19/month with 2,400 credits; Standard $24 with 3,6251 | Morphic vs SceneMixer |

| MovieFlow | A video agent and several models: “Your idea. A movie. In minutes.” (Beta)1 | Standard $15.83/month billed yearly ($19 monthly) with 1,800 credits; Pro $32.50 ($39) with 3,880; Ultra $124.92 ($142) with 14,300; Starter $5.75 billed yearly with 1,0002 | MovieFlow vs SceneMixer |

| Neolemon | Illustrated children's books: consistent cartoon characters, a multi-character scene creator, speech bubbles, covers and print-ready export2 | Creator plan $29 a month, billed monthly: 600 credits (up to 150 generations), commercial use included, credits do not roll over1 | Neolemon vs SceneMixer |

| OpenArt | An AI creator studio for images and video, with 100+ image, video and audio models1 | Starter $14 per seat a month billed monthly ($13 billed annually), 4,000 credits1 | OpenArt vs SceneMixer |

| Pika | All models on every plan; parallel video, image and audio generations1 | Free: credit packs only; Starter $10/month with 900 credits; Creator $35/month with 3,1501 | Pika vs SceneMixer |

| Pippit | A video suite: AI Drama Generator, Story Studio, avatars, talking photos, product videos1 | Free plan with daily credits and no card; paid plans list 1,575, 3,885 and 35,500 credits a month and “Remove watermark”; prices shown in local currency2 | Pippit vs SceneMixer |

| PixVerse | PixVerse models plus many other image and video models; batch creation and off-peak modes1 | Standard $8/month on the yearly plan ($96), 1,200 credits every 30 days, up to 720P1 | PixVerse vs SceneMixer |

| Pollo AI | Its own Pollo models plus third-party video and image models1 | Free, Lite (400 credits a month), Pro (800) and Ultra (5,000), with 1–6 parallel tasks1 | Pollo AI vs SceneMixer |

| Revid AI | Create, publish, learn, repeat: rendering, publishing, performance stats and Auto-Mode1 | Growth $39/month ($32 billed yearly) with 2,000 AI credits; Ultra $199/month with 12,0001 | Revid AI vs SceneMixer |

| Runway | Its Gen-4.5 model plus other video and image models, an agentic creative collaborator and editing1 | Free: 125 one-time credits; Standard $15/month ($12 billed annually), 625 credits = 52s of Gen-4.51 | Runway vs SceneMixer |

| Showrunner | Watching AI shows and making scenes within them; remixing; starring in a show by uploading yourself1 | Not stated on the pages we checked | Showrunner vs SceneMixer |

| Sora | App discontinued April 26, 2026; API discontinued September 24, 20261 | Not stated on the pages we checked | Sora vs SceneMixer |

| StoryShort | Automated short videos with voiceovers, background music and auto-posting1 | Starter $39/month ($23/month billed yearly) with 400 credits1 | StoryShort vs SceneMixer |

| Topview | “A one-line story idea, a short plot outline, a full drama script, a novel chapter, or a serialized story concept”1 | Plus $36/month billed annually (100 credits a month); Business $44/month billed annually against a $75 list price (3,000 credits a year, delivered up front); Ultra $50/month billed annually against $150 ($599.9 a year; 500 credits a month); Team $56 a seat2 | Topview vs SceneMixer |

| VEED | A browser video editor with a generative studio: AI video, avatars, dubbing, subtitles and translation12 | Creator 6,000 credits, Pro 30,000, Studio 180,000, a user a month billed yearly; shown in local currency (AU$17, AU$35 and AU$56 when we checked)1 | VEED vs SceneMixer |

| Veo | A video generation model, Veo 3.1: sound effects, ambient noise and dialogue generated natively with the picture1 | Not stated on Google's Veo pages. Flow, which lists Veo 3.1, is free with a Google account or part of Google AI Plus ($4.99/month), Pro ($19.99) and Ultra ($99.99 and $199.99); prices vary by market4 | Veo vs SceneMixer |

| Vidnoz | AI avatars, video templates and voiceover: “Scale Video Production 10X Faster”12 | Starter $26.99/month ($19.99 billed yearly), 450 credits a month, 1080p, no watermark; Business $74.99 ($56.99 billed yearly), 900 credits a month, voice clone and video translation1 | Vidnoz vs SceneMixer |

| Vidu | Video generation with multi-reference consistency, 2D animation, off-peak mode and batch generation1 | Free plan; Standard listed at $10/month (shown at $8.25 with a first-purchase discount), 800 credits a month1 | Vidu vs SceneMixer |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## How to read this table

- Built around is the first row of each tool's comparison page: the job the product is organised around, in its own words.

- Price to start is the entry plan or free tier as listed on the tool's pricing page on the date shown. Plans change often; follow the link for the exact source and the date.

- Not a ranking. A clip-first video model, a faceless-video pipeline and a series pipeline are different products. The one-to-one pages say when the other tool is the better fit.

- SceneMixer's row is not a quote: it is injected from the same source as the pricing page each time the page is rebuilt.

## What SceneMixer is for

SceneMixer turns a script or novel into a series: it reads the story into a cast, locations and an episode outline, draws reference art once, writes a storyboard for each episode, renders it segment by segment against the same character sheets with dialogue performed in any of 15 languages, and joins each episode into one file. New accounts start with 50 free credits and 15 free reference images.

## Frequently asked questions

### Which of these tools is best?

This page does not rank them. Each one is built around a different job; the one-to-one pages say when the other tool is the better fit, and the right answer depends on whether you are making one clip, one film or a series.

### Where do the facts about other tools come from?

Only from each tool's own public pages. Every cell carries the date we checked it, and the linked page lists the exact sources.

### Why is SceneMixer in the table?

Because this is SceneMixer's site, and readers compare it with these tools. Its numbers are filled in from the same source as the pricing page each time the page is rebuilt; the other tools' numbers are quoted as of the date shown.

### How often is this table rechecked?

At least every 45 days and after any price or plan change we notice. Corrections to support@scenemixer.com.

### What does SceneMixer do that most of these do not?

It produces a whole series from one script or novel: an episode outline, reference sheets for every character and place, a storyboard per episode, video segment by segment with dialogue performed in any of 15 languages, and one composited file per episode, at $0.06 a second and up.

### Can I move a project from one of these tools to SceneMixer?

Bring the script or outline as text and, on a paid plan, upload character images you have the rights to in place of generated sheets; the outline, storyboards and video are then generated here.


---

# Alici AI Alternative for AI Short Drama | SceneMixer

URL: https://scenemixer.com/alternatives/alici/
Alici plans shots with an agent and generates clips with many models, from $12 a month; SceneMixer turns a script into an episodic drama. When to pick which.

Alternatives

# An Alici AI Alternative for AI Short Drama

Updated November 1, 2026

Quick answer: Alici is an AI video creation platform: its Video Super Agent turns an idea, a photo, a product or a short script into a shot plan and generates clips with third-party video models, and it has an AI short drama use case that turns “a story idea or script into a multi-scene short drama”. SceneMixer is built for the series: a cast drawn once, an episode outline, a storyboard per episode, video segment by segment with the lines performed, in 15 languages. Choose Alici for short clips and promos planned by an agent across many models; choose SceneMixer when the same faces and rooms must carry twenty episodes.

## An agent that plans clips, or a pipeline that builds episodes

Alici's plans are sized in credits that videos and images share: Basic lists 150 credits a month as “12 videos at 480p, or 100 images”, and the top plan 3,8002. SceneMixer prices video by the finished second and lists it before each stage runs.

## At a glance

Alici AI facts as stated on its own pages, checked November 1, 2026

| | Alici AI | SceneMixer |

| Built around | A Video Super Agent that plans shots and generates clips; image and video studios; a short-drama use case1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Inputs | A photo, product, character, short script or idea; viral video links for recreation1 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Characters across shots | Not stated as a guarantee; the brief and shot plan are kept together across generation1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Video models | MiniMax H3, Seedance 2.5, Google, Kling, Runway, Grok, Vidu, Wan, HappyHorse; availability and durations vary by workflow1 | Tiers by resolution and price per second; the model behind each tier is managed for you |

| Plans | Basic $12/month (150 credits, up to 720p); Pro $25 (500 credits); Creator $59 (1,440); Ultimate $149 (3,800); first-month discounts shown; yearly “save 30%”2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Resolution | Basic up to 720p; other plans up to 4K; “maximum video resolution varies by model”2 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Glass Floor”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Alici AI if…

- You want an agent to plan a clip from a photo or a product and pick among many models for you1.

- You make promos and short pieces rather than episodes — Alici lists marketers, business owners and social teams first1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Alici AI lands in SceneMixer

| In Alici AI | Closest step in SceneMixer |

| Video Super Agent shot plan | The storyboard: numbered shots with seconds, who is in frame and the lines, editable as text |

| Short drama use case | A Short Drama project: episodes with hooks, cast drawn once. Upload a script and read it |

| Credits shared by images and videos | One credit balance; images, reads and video priced per item and shown before you press |

## Moving a story over

- Paste the script you would have given the agent — the read returns the cast, locations and an episode outline.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer an Alici alternative?

For an episodic drama with a fixed cast and performed dialogue, yes. For agent-planned clips across many video models, Alici is built around that.

### How much does Alici cost?

On November 1, 2026 its pricing page listed Basic at $12 a month with 150 credits, Pro at $25 (500), Creator at $59 (1,440) and Ultimate at $149 (3,800), with first-month discounts. Check alici.ai/pricing for current prices.

### Does SceneMixer have a free plan?

New accounts get 50 free credits and 15 free reference images; there is no monthly free allowance beyond that.

## Sources and corrections

- Alici AI (alici.ai) — what the platform is, tools, models, who it is for
- Pricing – Alici AI (alici.ai/pricing) — plans, credits, resolution

Facts about Alici AI come only from the pages listed above, as checked on November 1, 2026. Prices and plans change often, so check Alici AI's site before you buy. Alici AI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Boords Alternative for AI Storyboard to Video | SceneMixer

URL: https://scenemixer.com/alternatives/boords/
Boords is storyboard and animatic software with client review; SceneMixer renders the storyboard into video episodes. How they compare, and when Boords fits.

Alternatives

# A Boords Alternative for AI Storyboard to Video

Updated November 2, 2026

Quick answer: Boords is pre-production software for video teams: scripts, storyboards and timed animatics, shared with clients through one link so they can comment on the exact frame and sign off, with exports to Premiere, Resolve and After Effects. SceneMixer starts from the same storyboard but goes on to make the video: a script becomes an episode outline, a cast drawn once, a storyboard per episode, and each segment rendered with characters performing their lines on screen in 15 languages. Choose Boords to plan and get a shoot or an edit approved; choose SceneMixer when the storyboard itself should become the finished episode.

## A board to approve, or a board to render

Boords describes its job as getting “every video signed off before production starts”, with unlimited clients reviewing and commenting without an account on every plan1. In SceneMixer there is no handoff to a production crew: the storyboard is the input to the video model.

## At a glance

Boords facts as stated on its own pages, checked November 2, 2026

| | Boords | SceneMixer |

| Built around | Pre-production: scripts, storyboards and animatics, shared with clients for feedback and sign-off2 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Output | Frames, a timed animatic with your voiceover, PDF; export to Premiere, Resolve or After Effects at the approved timing2 | A rendered video file per episode: every segment generated, with voices and score, then composited |

| Review | “Unlimited clients review and comment without an account, on every plan”; version approval12 | A shareable link to a finished episode that opens a real page; no account needed on the other end |

| AI images | 250 a month on Solo, 1,000 on Pro, 2,000 on Team, 3,000 on Agency1 | The first 15 reference images on an account are free; video is priced by the second |

| Plans | Solo $39/month ($26 billed yearly), one user; Pro $75 ($50), up to 5 users; Team $125 ($85), up to 10; Agency $250 ($165), up to 301 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Free | 5 active storyboards, 3 active scripts and 8 one-off sample AI image credits; no card1 | 50 free credits and 15 free reference images for new accounts |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Veil at Dawn”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Boords if…

- You need clients to review and sign off a board — comments on the exact frame and an Approved version12.

- You are planning a live-action shoot or an edit in another tool — animatic export to Premiere, Resolve or After Effects2.

- You work in an agency team — plans for up to 30 users and an enterprise tier with SSO1.

- You want storyboards and scripts without generating video — unlimited storyboards on paid plans1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

- You want the storyboard to become the video. Each storyboard segment is rendered in place instead of exported to another tool. AI storyboard generator

## Where Boords lands in SceneMixer

| In Boords | Closest step in SceneMixer |

| Script to frames | The script read into an episode outline, then a storyboard per episode. Upload a script and read it |

| A frame | A numbered shot with seconds, who is in it and the spoken line. How to write one shot |

| Animatic with voiceover | Each segment rendered with its voices and score, then composited into one episode. Composite and share an episode |

| Client review link | A link that opens the finished episode on a real page. Composite and share an episode |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; frame descriptions you wrote can become shots in the storyboard.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Boords alternative?

For turning a storyboard into rendered video episodes, yes. For client review, sign-off and animatics for a shoot or an edit elsewhere, Boords is built around that.

### How much does Boords cost?

On November 2, 2026, its pricing page listed Solo at $39 a month ($26 billed yearly), Pro at $75 ($50), Team at $125 ($85) and Agency at $250 ($165), with a free plan and a 30-day money-back guarantee on paid plans. Check boords.com/pricing for current prices.

### Is Boords free?

Its pricing page, checked November 2, 2026, listed a free plan with 5 active storyboards, 3 active scripts and 8 one-off sample AI image credits, no card required.

### Can clients comment on a SceneMixer episode?

A finished episode can be shared with a link that opens a real page, with no account needed on the other end. Composite and share an episode.

## Sources and corrections

- Pricing – Boords (boords.com/pricing) — plans, seats, AI images, free plan, guarantee
- Storyboard & Pre-Production Software for Video Agencies – Boords (boords.com) — what it is, review, animatics, export

Facts about Boords come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Boords's site before you buy. Boords is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Canva AI Video Alternative for Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/canva/
Canva's Create a Video Clip makes 8-second clips inside a design suite; SceneMixer turns a script into an episodic drama. How they compare, and when Canva fits.

Alternatives

# A Canva AI Video Alternative for Drama Series

Updated November 2, 2026

Quick answer: Canva is a design suite, and its AI video generator, Create a Video Clip, makes a 16:9 video of up to eight seconds from text with synchronized audio including dialogue and sound effects, to drop into a design; it also offers talking-head avatars in over 40 languages. SceneMixer makes a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Canva for clips and videos inside presentations, social posts and brand templates; choose SceneMixer for a series told in episodes.

## A clip in a design, or an episode in a series

Canva describes Create a Video Clip as powered by Google's Veo-3, one video per prompt, which you then add to a Canva design for further customization2. SceneMixer is organised the other way round: the series comes first, and each clip is a segment of a storyboard.

## At a glance

Canva facts as stated on its own pages, checked November 2, 2026

| | Canva | SceneMixer |

| Built around | A design suite for presentations, social posts and video; Create a Video Clip generates a clip from text12 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Clip length | A 16:9 video of up to eight seconds with synchronized audio (dialogue, sound design, music); one video per prompt2 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Talking heads | A photo or selfie turned into a talking head, or an available AI avatar, delivering a script in over 40 languages2 | Lines are performed in the generated video itself, in any of 15 languages |

| Who can use it | Pro, Business, Enterprise and Nonprofit users get a limited number of clips a month2 | 50 free credits for new accounts, spent on whatever you generate |

| Plans | Free; Pro US$144 a year; Business US$250 a year per person; Enterprise on request. AI tools share a monthly allowance that varies by plan; an AI Pass add-on adds more1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Commercial use | May be used for personal or commercial projects under Canva's AI Product Terms and Terms of Use2 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Glass Floor”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Canva if…

- You design more than video — presentations, social posts, print and brand templates in one place1.

- You need an eight-second clip inside a design — generate it and add it straight to the page you are building2.

- You want a presenter in over 40 languages from a photo or an available avatar2.

- You work in a team with brand controls — its plans list collaboration and brand management1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Canva lands in SceneMixer

| In Canva | Closest step in SceneMixer |

| Create a Video Clip | A storyboard segment, rendered on a video tier. How to write one shot |

| Talking head / AI avatar | A character performing lines in a scene. AI video with talking characters |

| Editing the clip in the design | The episode editor, with one card per segment. The episode editor, a tour |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; the prompts you wrote for single clips can become shots in the storyboard.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Canva alternative?

For turning a written story into a multi-episode drama, yes. For presentations, social posts and design work with video inside, Canva is built around that.

### How much does Canva cost?

On November 2, 2026, its US pricing page listed Canva Free, Pro at US$144 a year and Business at US$250 a year per person, with Enterprise on request. Check canva.com/pricing for your region.

### Is Canva's AI video generator free?

Its page, checked November 2, 2026, said Pro, Business, Enterprise and Nonprofit users can create a limited number of clips a month.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15. Native dialogue vs dubbing.

## Sources and corrections

- Plans and pricing – Canva (canva.com/pricing) — plans, prices in US dollars, shared AI allowance
- AI Video Generator – Canva (canva.com/features/ai-video-generator) — Create a Video Clip: length, audio, avatars, who can use it, commercial use

Facts about Canva come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Canva's site before you buy. Canva is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# CapCut AI Video Alternative for Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/capcut/
CapCut is a video editor with an AI video maker; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when CapCut fits.

Alternatives

# A CapCut AI Video Alternative for Drama Series

Updated November 3, 2026

Quick answer: CapCut is a video editor on desktop, web and mobile, and its AI video maker starts from a script or reference image, a style, an aspect ratio, a voiceover and a duration, then lets you edit visuals, pacing, music, subtitles and effects before export. SceneMixer is not an editor: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose CapCut to edit footage and make short videos with templates and effects; choose SceneMixer to produce a series whose characters stay the same.

## An editor with AI, or a series pipeline

CapCut's pricing page lists a free plan with free templates and basic editing, Pro with 850 monthly AI credits and 1 TB of storage, and Ultra with a larger credit pool1. Its AI video page describes a guided workflow: start from a topic or script, adjust the settings, preview the AI output, then refine the video before export2. SceneMixer's workflow is fixed around the story: script, cast, storyboard, video, composite.

## At a glance

CapCut facts as stated on its own pages, checked November 2, 2026

| | CapCut | SceneMixer |

| Built around | A video editor (desktop, online, mobile) with an AI video maker, templates, captions and effects12 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| AI video inputs | A script or reference image; styles such as realistic film, cartoon 3D or cinematic movie; aspect ratio, voiceover and duration settings2 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Characters across scenes | Digital avatars with voiceover choices; how a character is kept the same across scenes is not stated on the page2 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Plans | Free; Pro 850 monthly AI credits and 1 TB of storage; Ultra 12,179 monthly AI credits; Teams per seat. Shown in local currency (AUD 37.99 and 111.99 a month for Pro and Ultra when we checked)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Seconds of AI video | Pro: up to 100 seconds a month of its Seedance 2.5 video at 720p; Ultra: up to 235 seconds; first-month prices from $0.12 and $0.18 a second12 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| After generating | Adjust visuals, pacing, background music, subtitles and effects before export2 | The episode editor: one card per segment; trim, re-dub, AI-edit or enhance a clip; subtitles and a built-in score per episode |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rust City Knockout”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with CapCut if…

- You edit video — trimming, captions, templates and effects across desktop, web and mobile1.

- You make short social or marketing videos from a topic with a guided workflow and your own footage beside the AI clips2.

- You want to start free — the free plan lists free templates and basic editing tools1.

- You prefer one app for editing and generation with a plan that bundles storage and AI credits1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where CapCut lands in SceneMixer

| In CapCut | Closest step in SceneMixer |

| Instant AI video (script to video) | The script read into cast, locations and an episode outline, then a storyboard per episode. Upload a script and read it |

| Digital avatar and voiceover | Character sheets drawn once and a voice per character, performed in the shot. Pick character voices |

| Captions and music | Subtitles and a score per episode. Add subtitles · Background music |

| Timeline editing | The episode editor, with one card per segment on the timeline. The episode editor, a tour |

## Moving a story over

- Bring the script you would have pasted into the AI video maker — SceneMixer's read returns the cast, locations and an episode outline.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a CapCut alternative?

For turning a written story into a multi-episode drama with a fixed cast, yes. For editing footage you shot, templates and effects, CapCut is built around that.

### How much does CapCut cost?

On November 2, 2026, its pricing page listed a free plan, Pro with 850 monthly AI credits and Ultra with 12,179, with prices shown in local currency (AUD 37.99 and 111.99 a month for us). Check capcut.com/pricing for your region.

### Is SceneMixer a video editor like CapCut?

No. It generates the episode and lets you trim, re-dub, AI-edit and enhance each clip, but it is not a multi-track editor. The episode editor, a tour.

### Can I take a SceneMixer episode into CapCut?

The finished episode can be downloaded at full quality, and you own the results. Composite and share an episode.

## Sources and corrections

- CapCut pricing (capcut.com/pricing) — plans, monthly AI credits, per-second video prices
- AI Video Generator – CapCut (capcut.com/tools/ai-video-generator) — what the AI video maker takes, steps, seconds per plan

Facts about CapCut come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check CapCut's site before you buy. CapCut is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Dashtoon Studio Alternative for AI Story Video | SceneMixer

URL: https://scenemixer.com/alternatives/dashtoon/
Dashtoon's creation studio shut down on August 15, 2026; its site still hosts comics. If you made stories there, here is how SceneMixer picks the work up.

Alternatives

# A Dashtoon Studio Alternative for AI Story Video

Updated November 2, 2026

Quick answer: Dashtoon's own notice says Dashtoon Studio shut down on 15 August 2026 and is no longer available, and that creation continues on Frameo, which it calls its new platform for AI video and short drama; dashtoon.com itself still offers comics, manga and manhwa to read. If you used Dashtoon Studio to make story content, SceneMixer covers that work from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Dashtoon to read comics; choose SceneMixer to make a story into episodes.

## What Dashtoon says now

The page that used to be the creator studio now reads “Dashtoon Studio has closed”, with a pointer to Frameo and a contact address for questions about your account or work1. The home page offers a reading app for comics, manga and manhwa2. We did not compare Frameo here; check its own pages.

## At a glance

Dashtoon facts as stated on its own pages, checked November 2, 2026

| | Dashtoon | SceneMixer |

| Creation | Dashtoon Studio shut down on 15 August 2026 and is no longer available; creation continues on Frameo, its new platform for AI video and short drama1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Reading | Comics, manga and manhwa to read online, with an app2 | Not a reading platform: SceneMixer makes video episodes |

| Your account and work | Questions: contact@dashtoon.com1 | Bring the story as text — a script, a document, or one sentence — and, on a paid plan, character art you own |

| Price | Not stated on the pages we read: dashtoon.com/pricing returned “page could not be found”2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rust City Knockout”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Dashtoon if…

- You want to read comics, manga and manhwa — Dashtoon's site and app are a reading platform2.

- You want its creator tools as they continue — Dashtoon points creators to Frameo; read Frameo's own pages before deciding1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Dashtoon lands in SceneMixer

| In Dashtoon | Closest step in SceneMixer |

| A story you were drawing | A script, read into a cast, locations and an episode outline. Upload a script and read it |

| A recurring character | A character reference sheet used in every segment; paid plans can upload their own art. Regenerate or replace one sheet |

| A scene or page | A storyboard shot: where it happens, who is there and what they say. How to write one shot |

## Moving a story over

- Start from the story. Paste or upload the script or the text of your comic, or describe the premise in one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is Dashtoon Studio still available?

Dashtoon's page, checked November 2, 2026, says Dashtoon Studio shut down on 15 August 2026 and is no longer available, and that creation continues on Frameo.

### Is SceneMixer a Dashtoon alternative?

For making a story into animated episodes with a fixed cast and spoken dialogue, yes. For reading comics, Dashtoon's site and app are built for that.

### What happened to my Dashtoon work?

The closure notice says to write to contact@dashtoon.com with questions about your account or your work; SceneMixer has no access to it.

### Can SceneMixer turn a comic into video?

It works from the story as text rather than from drawn pages: the script is read into a cast and an episode outline, and each panel's action and lines become shots. Make an AI animated short film.

## Sources and corrections

- Dashtoon Studio has closed (dashtoon.com/create) — the shutdown notice and what Dashtoon says about where creation continues
- Dashtoon (dashtoon.com) — what the site offers now

Facts about Dashtoon come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Dashtoon's site before you buy. Dashtoon is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# DramaSo Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/dramaso/
DramaSo breaks down viral clips and previews dialogue as audio first; SceneMixer turns a script into episodes with a fixed cast. How they compare.

Alternatives

# A DramaSo Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: DramaSo and SceneMixer both make AI short dramas, but they start in different places. DramaSo starts from what already went viral: it breaks a clip into a beat sheet, lets you hear the dialogue on an audio-first timeline, then renders native 9:16 video with burned-in subtitles. SceneMixer starts from your script or novel: it reads it into a cast, an episode outline and a storyboard per episode, renders every shot against the same character sheets, and has the characters perform their lines on screen in 15 languages, landscape or vertical. Choose DramaSo to reverse-engineer hooks and test pacing cheaply; choose SceneMixer to produce a series from your own story.

## Decode a hit, or produce your story

DramaSo describes an audio-first workflow — text, then playable dialogue, then dialogue with storyboard stills, then full vertical video — so credits go to motion only when an episode sounds right1. SceneMixer's order is script, cast sheets, storyboard, video: the look of every character is fixed before the first shot is rendered, and each segment can be regenerated on its own.

## At a glance

DramaSo facts as stated on its own pages, checked November 2, 2026

| | DramaSo | SceneMixer |

| Starting point | Break down a trending clip into a beat sheet, or paste your own script1 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Workflow | Audio-first: dialogue you can play, then storyboard stills on the timeline, then motion1 | Script read, reference sheets, storyboard per episode, video per segment, composite |

| Output | Native 9:16 with captions and safe zones; burned-in subtitles; 720p or 1080p export1 | 16:9 by default, 9:16 available; dialogue performed on screen; captions added after export |

| Characters | Storyboard frames lock the look per line1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Price to start | Free: 80 welcome credits once; Pro $39/month with 3,000 credits (“≈ 55s of finished film” on the default model)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Unused credits | A rolling 30-day monthly pool is spent first; wallet credits never expire1 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Glass Floor”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with DramaSo if…

- You study viral hits before writing. DramaSo breaks clips into hook timing, conflict beats and cliffhanger markers1.

- You want to hear an episode before paying for video. Its audio-first timeline is built for that1.

- You only make vertical episodes with burned-in captions for TikTok, Reels and Shorts1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You need landscape as well as vertical. 16:9 is the default and 9:16 is one setting away. Vertical and landscape projects.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

## Where DramaSo lands in SceneMixer

| In DramaSo | Closest step in SceneMixer |

| Beat sheet | The episode outline, with a conflict arc and an end-of-episode hook per episode. Upload a script and read it |

| Audio-first dialogue | Lines written into the storyboard and performed in the video. Lines, voice-over and on-screen text |

| Storyboard stills on the timeline | The storyboard per episode, shot by shot. How to write one shot |

## Moving a story over

- Bring the script or the beat sheet as text — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a DramaSo alternative?

For producing a series from your own script with a fixed cast and on-screen dialogue in many languages, yes. For breaking down viral clips and testing pacing as audio first, DramaSo is built around that.

### How much does DramaSo cost?

On November 2, 2026, its site listed a free tier with 80 welcome credits, Pro at $39 a month (3,000 credits) and Studio at $149 a month (12,000 credits), plus credit packs from $20 and a $9.9 first-episode pack of 2,000 credits for new customers. Check drama.so for current prices.

### Does SceneMixer burn in subtitles?

No. Short on-screen text can be written into a shot; platform captions are added after export.

### Can SceneMixer make vertical dramas?

Yes. Set the project to 9:16 in step 1 and every shot is generated vertical. Vertical and landscape projects.

## Sources and corrections

- AI Short Drama Generator · DramaSo (drama.so) — workflow, export format and pricing

Facts about DramaSo come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check DramaSo's site before you buy. DramaSo is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Dramily Alternative for AI Short Drama | SceneMixer

URL: https://scenemixer.com/alternatives/dramily/
Dramily takes a story idea to scripts, locked assets, storyboards and video prompts for short dramas; SceneMixer renders the episodes. How they compare.

Alternatives

# A Dramily Alternative for AI Short Drama

Updated November 2, 2026

Quick answer: Dramily is an AI short-drama platform: it connects scripts, reusable assets (characters, scenes, props, covers), reviewed storyboards and video prompts in one workspace, with projects of up to 120 episodes, a choice of video models in supported workflows and monthly credit plans from $49. SceneMixer covers the same ground from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages, and credits that do not expire at the end of the month. Choose Dramily for its asset-locking and review-before-video workflow; choose SceneMixer to go from one script to finished episodes in a single project.

## Two pipelines for the same job

Dramily describes its workflow as one SaaS that turns “an idea into scripts, reusable assets, storyboards, video prompts, previews, covers, and export-ready episodes”1. SceneMixer does the same job with a different centre of gravity: the script is read once into a cast and an episode outline, and every segment is rendered against the same character sheets.

## At a glance

Dramily facts as stated on its own pages, checked November 2, 2026

| | Dramily | SceneMixer |

| Built around | One SaaS for short-drama production: scripts, reusable assets, storyboards, video prompts, previews, covers and export-ready episodes1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Episodes | Serialized dramas of up to 120 episodes per project1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Assets | Create, upload, compare and lock reusable character, scene, prop and cover assets; each storyboard shot is linked to its assets1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Review before video | Review scripts, shots, prompts and linked assets together before spending video credits1 | Each segment's storyboard is text you can read and edit before you generate it; prices are shown before each stage runs |

| Video models | Seedance 2.0 and 2.5, Kling V3, Wan 2.7, HappyHorse 1.0, Grok Imagine Video 1.5, Veo 3.1 and Hailuo 2.3 in supported workflows1 | Tiers by resolution and price per second; the model behind each tier is managed for you |

| Voices | Generate or upload voice samples per character and reuse them across storyboard videos1 | Lines are performed in the generated video itself, in any of 15 languages |

| Plans | Free: 50 evaluation credits. Creator $49/month with 1,250 credits; Studio $99 with 2,600; Producer $199 with 5,300; yearly saves 10%1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Credits | Monthly credits expire after 30 days1 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

| Commercial use | Included on paid plans; the free plan is for non-commercial evaluation1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Neon Debt”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Dramily if…

- You want to approve every scene before any video credits are spent — Dramily puts a review step between storyboard and video1.

- You want to compare and lock asset versions — it lists variants you approve and keep attached to every downstream shot1.

- You want to choose among many video models yourself — it lists eight model families in supported workflows1.

- You plan a series of up to 120 episodes — its page states that limit per project1.

## SceneMixer fits if…

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

- You need landscape as well as vertical. 16:9 is the default and 9:16 is one setting away. Vertical and landscape projects.

## Where Dramily lands in SceneMixer

| In Dramily | Closest step in SceneMixer |

| AI script generation | AI Create drafts a script from one sentence, and the script read returns the cast and an episode outline. Upload a script and read it |

| Asset workspace | The asset library: characters, locations and props, each with its reference image. Regenerate or replace one sheet |

| Storyboard planning and review | A storyboard per segment, edited as text before you generate it. How to write one shot |

| Per-character voices | A voice per character chosen from a library. Pick character voices |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; character, scene and prop images you already have can be uploaded as references on a paid plan.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Dramily alternative?

For turning a story into finished short-drama episodes with a fixed cast and spoken dialogue, yes. Dramily is built around the same job, with its own asset-locking and review steps.

### How much does Dramily cost?

On November 2, 2026, its pricing section listed a Free plan with 50 evaluation credits, Creator at $49 a month (1,250 credits), Studio at $99 (2,600) and Producer at $199 (5,300), with monthly credits expiring after 30 days. Check dramily.com for current prices.

### Is Dramily free?

Its pricing section, checked November 2, 2026, listed a Free plan with 50 evaluation credits for non-commercial evaluation use.

### Do SceneMixer credits expire?

Plan credits and bought credits sit in one balance and do not expire at the end of the month. Do credits expire?

## Sources and corrections

- Dramily – Free AI Short Drama Generator (dramily.com) — what the platform does, plans, credits, models, commercial use and voices (its FAQ)

Facts about Dramily come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Dramily's site before you buy. Dramily is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Freepik (Now Magnific) Alternative for AI Drama | SceneMixer

URL: https://scenemixer.com/alternatives/freepik/
Freepik now trades as Magnific: stock content plus many AI video models on yearly credits. SceneMixer turns a script into an episodic drama. How they compare.

Alternatives

# A Freepik (Magnific) Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Freepik now trades as Magnific — its pages are titled “Magnific (formerly Freepik)” — a creative platform with 250M+ stock assets and AI image, video and audio tools across many models, sold as yearly credit allowances. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Magnific for stock content plus a wide menu of video and image models; choose SceneMixer to produce a series with a fixed cast.

## A stock library with models, or a series pipeline

Magnific's pricing page lists Premium with 240K credits a year, Premium+ with 600K and Pro Starter with 1.5M, with credits valid for a year rather than resetting monthly, and shows what each video model costs in credits1. SceneMixer prices video by the finished second and shows the price before each stage runs.

## At a glance

Freepik facts as stated on its own pages, checked November 2, 2026

| | Freepik | SceneMixer |

| Built around | A creative platform: 250M+ stock photos, videos, vectors and PSDs plus AI image, video and audio tools across many models1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Video inputs | An image to open on, optionally a last frame; some models make sound with the picture or several shots in one clip (up to 15 seconds on the longest)2 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Video models | Lists Kling, Seedance, Google Veo 3.1, Gemini Omni, MiniMax and others, each with its own credit cost per clip12 | Tiers by resolution and price per second; the model behind each tier is managed for you |

| Characters | A “Train AI characters” perk listed at 3,000 credits a character, up to 200 characters1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Plans | Premium 240K credits a year; Premium+ 600K, with unlimited generations on some models; Pro Starter 1.5M; credits valid for 1 year; shown in local currency (A$18.50, A$45 and A$109.50 a month billed annually when we checked)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Licence | Commercial AI license on paid plans; Premium+ adds music rights1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Neon Debt”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Freepik if…

- You also need stock content — 250M+ photos, videos, vectors and PSDs alongside the AI tools1.

- You want many video and image models on one plan — each clip names the model it was made with12.

- You like yearly credits that do not reset monthly — its credits are valid for a year1.

- You need upscalers and editing tools around the clip — Magnific and Topaz upscalers are listed on its plans1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Freepik lands in SceneMixer

| In Freepik | Closest step in SceneMixer |

| Start from a frame, add a last frame | The next segment opens where the last one ended. Shot-to-shot continuity |

| Train AI characters | Character sheets generated once and attached to every segment. Regenerate or replace one sheet |

| Video Upscaler | Enhance on a finished clip, one resolution step up. Enhance a clip |

| Spaces (workflow canvas) | The fixed project pipeline: read, sheets, storyboard, video, composite |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; storyboard panels you drew can become shots, and character art you own can be uploaded on a paid plan.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Freepik alternative?

For turning a written story into a multi-episode drama with a fixed cast, yes. For stock content and a wide menu of models in one subscription, Magnific (formerly Freepik) is built around that.

### Is Freepik now Magnific?

On November 2, 2026, freepik.com/pricing led to magnific.com/pricing, whose page title reads “Magnific (formerly Freepik)”. Check Magnific's own pages for the current name and plans.

### How much does Freepik cost?

On November 2, 2026, Magnific's pricing page listed Premium with 240K credits a year, Premium+ with 600K and Pro Starter with 1.5M, priced in local currency (A$18.50, A$45 and A$109.50 a month billed annually for us). Check magnific.com/pricing for your region.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15. Native dialogue vs dubbing.

## Sources and corrections

- Pricing plans – Magnific, formerly Freepik (magnific.com/pricing) — plans, yearly credits, video models and credit costs, licences
- AI Video Generator – Magnific, formerly Freepik (magnific.com/ai/video-generator) — how a clip is made, models, first and last frame

Facts about Freepik come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Freepik's site before you buy. Freepik is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Gemini Omni Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/gemini-omni/
Gemini Omni Flash makes and edits short clips in conversation; SceneMixer turns a script into an episodic drama. What Google states, and when Omni fits.

Alternatives

# A Gemini Omni Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Gemini Omni Flash is Google's video model for fast generation and conversational editing: Google's release notes describe 3–10 second clips at 720p from text or a still image, edited by talking to it, with extension and first-and-last-frame transitions added at general release on August 27, 2026. SceneMixer starts one level up: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Omni to make and refine single clips with Google's model; choose SceneMixer to produce a series from a written story.

## A clip you talk to, or a series you write

Google lists Gemini Omni Flash as a model for high-speed video generation and conversational video editing: describe a clip, then refine it in follow-up turns1. SceneMixer is organised around the story rather than the clip: the outline decides the episodes, every segment is rendered against the same character sheets, and a change is made by editing the storyboard text of one segment.

## At a glance

Gemini Omni facts as stated on its own pages, checked November 2, 2026

| | Gemini Omni | SceneMixer |

| Built around | A video model for high-speed generation and conversational editing, used through the Gemini API1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Clip length | 3–10 second videos; at general release, extension continues a clip from its end1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Resolution | 720p by default; 360p, 1080p and 4K options, with 1080p and 4K generated by upscaling1 | 480P, 720P and 1080P video tiers, chosen per segment, each with its own price per second |

| Inputs | Text, or still images to animate; up to 2 images for a first-and-last-frame transition1 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Price | Not stated in the release notes (checked November 2, 2026)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Violin on Mercer St”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Gemini Omni if…

- You want Google's own video model through the Gemini API, and you build your own tools around it1.

- You refine one clip at a time by describing changes in conversation1.

- You need 4K output from the same model, accepting that 1080p and 4K are upscaled1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Gemini Omni lands in SceneMixer

| In Gemini Omni | Closest step in SceneMixer |

| A conversational edit | Edit the storyboard text of one segment and regenerate it. How to write one shot |

| Extend a clip | The next segment of the storyboard, which opens where the last one ended |

| First and last frame | Character, location and prop sheets attached to every segment. Regenerate or replace one sheet |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; prompts you wrote for single clips can become shots in the storyboard.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Gemini Omni alternative?

For turning a written story into a multi-episode drama with a fixed cast, yes. For Google's own model used clip by clip, or through the Gemini API, Omni is the direct route.

### How long can a Gemini Omni Flash video be?

Google's release notes, checked November 2, 2026, describe 3-10 second videos, with an extension option added at general release that continues a clip from its end.

### What resolution does Gemini Omni Flash make?

720p by default. The general-release notes list 360p, 1080p and 4K as options and say 1080p and 4K are generated by upscaling (checked November 2, 2026).

### How much does Gemini Omni cost?

The release notes we checked on November 2, 2026 do not state a price. Check Google's own pricing pages for the Gemini API.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15.

### Does SceneMixer use Gemini models?

Two of SceneMixer's image tiers for drawing reference sheets are Google's Nano Banana models: Nano Banana Pro and Nano Banana 2. See Nano Banana on SceneMixer. The video tiers are listed on the pricing page by resolution and price per second.

## Sources and corrections

- Gemini API release notes (ai.google.dev/gemini/docs/changelog) — the June 30, 2026 preview and August 27, 2026 general release of the Gemini Omni Flash video model

Facts about Gemini Omni come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Gemini Omni's site before you buy. Gemini Omni is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Google Flow Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/google-flow/
Google Flow is Google's AI creative studio with Veo and Nano Banana; SceneMixer turns a script into an episodic drama. How they compare, and when Flow fits.

Alternatives

# A Google Flow Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Google Flow is Google's AI creative studio built on its own models — Veo 3.1 with native audio, Nano Banana for images, Gemini Omni for video creation and editing — with an agent and user-built tools such as a Storyboard Studio, free to try with a Google account. SceneMixer turns a script into a drama series: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Flow to work shot by shot with Google's models; choose SceneMixer to produce a series from a written story.

## Google's studio, or a series pipeline

Flow lists creating and editing images and video with natural language, an agent, and tools people build inside it — text overlays, a video reframer, an image editor, a Storyboard Studio for scripts, casts and boards — with features varying by Google AI subscription tier, platform and region1. SceneMixer starts from the script instead of the shot: the outline decides the episodes, and every segment is rendered against the same character sheets.

## At a glance

Google Flow facts as stated on its own pages, checked November 2, 2026

| | Google Flow | SceneMixer |

| Built around | Google's models: Gemini Omni, Nano Banana and Veo 3.1 (native audio); an agent and user-built tools such as a Storyboard Studio1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Free tier | Free with a Google account; no credit figure stated on the page1 | 50 free credits and 15 free reference images for new accounts |

| Paid | Google AI Plus $4.99/month, Google AI Pro $19.99, Google AI Ultra $99.99 and $199.99; prices vary by market1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Access | Features vary by subscription tier, platform and region; 18+1 | Web app; the same features in every region the site serves |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Neon Debt”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Google Flow if…

- You want Google's models — Veo 3.1 with native audio, Nano Banana, Gemini Omni — straight from Google1.

- You already pay for Google AI and its monthly Flow credits cover your shots1.

- You build scenes shot by shot with Google's models and the tools built inside Flow1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Google Flow lands in SceneMixer

| In Google Flow | Closest step in SceneMixer |

| Storyboard Studio tool | The storyboard per episode, joined into one file. How to write one shot |

| Reference images for a character | Character sheets generated once and attached to every segment. Regenerate or replace one sheet |

| Agent | AI Create and the script read, which plan the whole series |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Google Flow alternative?

For turning a written story into a multi-episode drama, yes. For Google's own models used shot by shot, Flow is the direct route.

### How much does Google Flow cost?

On November 2, 2026, its page listed free use with a Google account, Google AI Plus at $4.99 a month, Google AI Pro at $19.99 and Google AI Ultra at $99.99 or $199.99, with prices varying by market. Check Google's page for current prices in your region.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15.

## Sources and corrections

- Google Flow – AI Creative Studio for Video, Images & Custom Tools (flow.google.com/about) — models, tools and plans

Facts about Google Flow come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Google Flow's site before you buy. Google Flow is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Grok Imagine Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/grok-imagine/
Grok Imagine generates and edits videos up to 15 seconds; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when Grok fits.

Alternatives

# A Grok Imagine Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Grok Imagine is xAI's image and video generator. xAI's developer documentation describes videos of up to 15 seconds from text or a still image, guidance from up to 14 reference images and 3 voice references, video editing from a prompt and extension from the last frame, priced per second by duration and resolution. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Grok Imagine for single clips inside xAI's tools; choose SceneMixer for a series.

## Clips from a prompt, or a series from a script

xAI documents Grok Imagine video as text-to-video and image-to-video with configurable duration, aspect ratio and resolution, plus reference-to-video, editing and extension1. SceneMixer keeps references inside the project instead of in each request: every character, location and prop has one sheet, and every segment that needs it uses it.

## At a glance

Grok Imagine facts as stated on its own pages, checked November 2, 2026

| | Grok Imagine | SceneMixer |

| Built around | Video generation in xAI's Grok, with an API for developers1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Clip length | Configurable duration up to 15 seconds; extension continues a video from its last frame1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| References | Up to 14 reference images and 3 voice references on grok-imagine-video-1.51 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Editing | Edit an existing video with a text prompt1 | Edit one segment's storyboard text and regenerate it; trim, re-dub or AI-edit a finished clip |

| Price | Per second, by duration and resolution; rates on xAI's pricing page, not on the documentation page (checked November 2, 2026)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Glass Floor”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Grok Imagine if…

- You already use Grok and want video in the same place1.

- You build on an API with references, editing and extension in each request1.

- You make one-off clips up to 15 seconds1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Grok Imagine lands in SceneMixer

| In Grok Imagine | Closest step in SceneMixer |

| Reference images | Character, location and prop sheets in step 2. Regenerate or replace one sheet |

| Video extension | The next segment of the storyboard |

| Video editing | Edit this clip: trim, AI visuals or a re-dub of one finished segment. How to write one shot |

## Moving a story over

- Bring the story — a script, a document, or one sentence for AI Create.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Grok Imagine alternative?

For a drama in episodes with a fixed cast and spoken dialogue, yes. For single clips inside Grok, or video through xAI's API, Grok Imagine is built for that.

### How long can a Grok Imagine video be?

xAI's documentation, checked November 2, 2026, says video generation supports configurable duration up to 15 seconds, and extension continues a video from its last frame.

### Is Grok Imagine free?

xAI's video documentation does not state free-tier limits (checked November 2, 2026). Check Grok's own plans page for what your account includes.

### How is Grok Imagine video priced?

Per second; xAI's documentation says both duration and resolution affect the total and points to its pricing page for rates (checked November 2, 2026).

### Can SceneMixer keep the same character across episodes?

Yes. Each character is drawn once as a reference sheet and attached to every segment they appear in, across every episode. How consistency works.

### Can I use SceneMixer results commercially?

Yes. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Sources and corrections

- Video generation (docs.x.ai/docs/guides/video-generations) — xAI's developer documentation for Grok Imagine video: duration, inputs, references, editing, extension and pricing basis

Facts about Grok Imagine come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Grok Imagine's site before you buy. Grok Imagine is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Hailuo AI Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/hailuo/
Hailuo AI makes video clips with MiniMax models from $14.99 a month; SceneMixer turns a script into an episodic drama. How they compare, and when Hailuo fits.

Alternatives

# A Hailuo AI Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Hailuo AI is MiniMax's video and image generator: its home page announces MiniMax H3 with “native multimodal generation, precise multimodal editing, production-ready content creation”, and its plans sell monthly credits for clips at 480p, 768p and 2K, with other labs' video and image models on paid plans. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Hailuo for fast single clips and a multi-model toolbox; choose SceneMixer for a series with a fixed cast.

## Clips from a model maker, or a series from a script

Hailuo's plans are priced by the second for each model and resolution, with discounts shown next to list prices and yearly billing lower than monthly1. SceneMixer prices video by the second too, but by resolution tier, and the outline, cast and storyboard around the video are part of the same project.

## At a glance

Hailuo AI facts as stated on its own pages, checked November 2, 2026

| | Hailuo AI | SceneMixer |

| Built around | A video and image generator from MiniMax: “native multimodal generation, precise multimodal editing, production-ready content creation”2 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Clip length | MiniMax H3 clips of 4–15 seconds at 768p or 2K; H3 Max 5–15 seconds at 480p or 768p1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Other models | Paid plans list access to Veo 3.1 beside Hailuo 2.3; image models such as Nano Banana Pro and Seedream1 | Tiers by resolution and price per second; the model behind each tier is managed for you |

| Plans | Free: one-time trial credits, Hailuo models only. Standard $14.99 with 1,000 credits a month; Pro $54.99 with 4,500; Master $119.99 with 10,500; Max $184 billed yearly with 27,000; discounts shown on the page1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Video price | MiniMax H3 at 2K from $0.081 a second and at 768p from $0.047 (Pro and above, billed yearly); Standard $0.101 and $0.059; higher when billed monthly1 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| Commercial use | “Any videos you can download without watermarks are yours”; paid plans remove watermarks1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

| Failed generations | “No credits will be charged for failed generations”1 | Credits are held while a task runs, charged on success and returned on failure. Do credits expire? |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Crimson Vow”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Hailuo AI if…

- You want fast single clips from MiniMax's own models — MiniMax H3 up to 2K and 15 seconds1.

- You want several video models on one subscription — the plans list Hailuo 2.3 and Veo 3.1 beside MiniMax's own models1.

- You want images, lip sync, music and upscaling beside the video — its plan table lists them as agent tools and image models1.

- You want a large monthly credit pool — up to 27,000 credits a month on its Max plan1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Hailuo AI lands in SceneMixer

| In Hailuo AI | Closest step in SceneMixer |

| References (Omni Reference) | Character, location and prop sheets attached to every segment. Regenerate or replace one sheet |

| Lip Sync tool | Lip sync or re-perform on a finished clip. Lip sync or re-perform |

| Music Generation | A built-in score written per episode. Background music |

| Video Upscale | Enhance on a finished clip, one resolution step up. Enhance a clip |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; prompts you wrote for single clips can become shots in the storyboard.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Hailuo AI alternative?

For a drama in episodes with a fixed cast and spoken dialogue, yes. For quick single clips and a multi-model toolbox, Hailuo is built around that.

### How much does Hailuo AI cost?

On November 2, 2026, its plans page listed a Free plan with one-time trial credits, Standard at $14.99 a month (1,000 credits), Pro at $54.99 (4,500), Master at $119.99 (10,500) and Max at $184 billed yearly (27,000), with discounts shown on each. Check hailuoai.video/subscribe for current prices.

### Is Hailuo AI free?

Its plans page, checked November 2, 2026, listed a Free plan with one-time trial credits for Hailuo models only.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15. Native dialogue vs dubbing.

## Sources and corrections

- Hailuo AI plans (hailuoai.video/subscribe) — plans, monthly credits, per-second prices, models, commercial use and watermark answers
- Hailuo AI (hailuoai.video) — what the site says it is for

Facts about Hailuo AI come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Hailuo AI's site before you buy. Hailuo AI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Hedra Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/hedra/
Hedra offers visual models and the infrastructure to run them; SceneMixer turns a script into an episodic drama. How they compare, and when Hedra fits.

Alternatives

# A Hedra Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Hedra describes itself as real-world AI from research to runtime: inference, sandboxes and model training for developers, billed per use through its API. SceneMixer is an application for one job: turning a script into a drama series, with an episode outline, a cast drawn once, a storyboard per episode and characters performing lines on screen in 15 languages. Choose Hedra to build on models; choose SceneMixer to produce episodes.

## A model platform, or a production line

Hedra's site leads with inference, sandboxes and model training for developers1; it is billed per use through its API, with no subscription, seats or minimum2. SceneMixer hides the models behind a pipeline: you choose a video tier by resolution and price, and the project handles the rest.

## At a glance

Hedra facts as stated on its own pages, checked November 2, 2026

| | Hedra | SceneMixer |

| Built around | “Real-world AI, from research to runtime”: inference, sandboxes and model training for developers1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Price to start | Pay-as-you-go API billing, “no subscription, no seats, no minimum”: video generation from 1.22¢ a second, image generation from 0.59¢ a generation2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Last Envelope”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Hedra if…

- You are building a product on visual models and need inference, fine-tuning or sandboxes1.

- You want direct access to many models through one account or API1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Hedra lands in SceneMixer

| In Hedra | Closest step in SceneMixer |

| Models and inference | Video tiers chosen per project or per segment, by resolution and price |

| Talking characters | Lines performed in the video in the project's language. Native dialogue vs dubbing |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Hedra alternative?

If you want finished drama episodes rather than models to build on, yes. If you are building your own product on visual models, Hedra's platform is built for that.

### How much does Hedra cost?

On November 2, 2026, its pricing page listed pay-as-you-go API billing with no subscription, no seats and no minimum: video generation from 1.22¢ a second and image generation from 0.59¢ a generation, prepaid from a dollar balance. Check hedra.com/pricing for current rates.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15.

## Sources and corrections

- Hedra — Visual Inference Models and Infrastructure (hedra.com) — positioning
- Hedra Pricing (hedra.com/pricing) — plans and credits

Facts about Hedra come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Hedra's site before you buy. Hedra is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# HeyGen Alternative for AI Story Video | SceneMixer

URL: https://scenemixer.com/alternatives/heygen/
HeyGen makes avatar presenter videos and translations; SceneMixer turns a script into an episodic drama with a fixed cast. When to pick which.

Alternatives

# A HeyGen Alternative for AI Story Video

Updated November 2, 2026

Quick answer: HeyGen is built around avatars and presenters: it says it helps businesses “create, localize, scale, and collaborate on video, no camera or crew needed”, with ready-made and custom video avatars, voices, translation into 175+ languages and dialects, and team workspaces. SceneMixer is built around stories: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with several characters acting and speaking in 15 languages. Choose HeyGen for a presenter speaking to camera or a video translated for another market; choose SceneMixer for scenes and episodes with characters who act.

## A presenter on camera, or a cast in scenes

HeyGen's plans are sized in credits, video length, export resolution and avatar counts: its Free plan lists 3 watermarked videos a month of up to a minute, and its Pro plan lists videos of up to 30 minutes and 4K export1. SceneMixer does not make a presenter-to-camera video; it makes a scene in which characters talk to each other.

## At a glance

HeyGen facts as stated on its own pages, checked November 2, 2026

| | HeyGen | SceneMixer |

| Built around | “Create, localize, scale, and collaborate on video, no camera or crew needed”: avatars, voices, translation and team tools1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Presenters | 500+ ready-made presenters, a custom video avatar with your own voice and likeness, and photo avatars1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Languages | 30+ languages on Free; 175+ languages and dialects from Creator1 | Lines are performed in the generated video itself, in any of 15 languages |

| Video length | Up to 1 minute each on Free; up to 30 minutes each on Pro1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Plans | Free: 3 watermarked videos a month. Creator $29/month ($24 billed yearly) with 600 credits and 1080p. Pro $49 ($41 billed yearly) with 1,000 credits and 4K. Business $149 plus $20 a seat1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Video generation | Creator lists HeyGen Video 1 (text, image and reference to video); Pro lists Seedance 2.5, Veo 3.1 and Nano Banana1 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| Credits | Roll over one month; annual plans accumulate credits until renewal1 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Last Envelope”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with HeyGen if…

- You need a presenter speaking to camera — one of 500+ ready-made presenters or an avatar of yourself delivering a script1.

- You translate videos for other markets — 175+ languages and dialects, with translation scripts you can edit and proofread1.

- You make training or onboarding video — its plan comparison lists interactive video, quizzes, branching and SCORM export1.

- You need one video of up to 30 minutes on Pro, or a team workspace with brand kits and role controls1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- Your story needs scenes, not a presenter. Characters act, move and speak to each other in locations drawn for the story. AI video with talking characters.

## Where HeyGen lands in SceneMixer

| In HeyGen | Closest step in SceneMixer |

| Avatar / presenter | A character sheet, drawn once; the character acts in the shot rather than speaking to camera. Regenerate or replace one sheet |

| Script to video | The script read into cast, locations and an episode outline. Upload a script and read it |

| Voices | A voice per character chosen from a library. Pick character voices |

| Translation languages | The dialogue language is set once in step 1 and the lines are performed in it. Choose the dialogue language |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; a presenter script can become dialogue lines in a scene.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a HeyGen alternative?

For stories told in episodes by several characters, yes. For avatar presenters, video translation and training video, HeyGen is built around those.

### How much does HeyGen cost?

On November 2, 2026, its pricing page listed a Free plan ($0, 3 watermarked videos a month), Creator at $29 a month ($24 billed yearly), Pro at $49 ($41 billed yearly) and Business at $149 plus $20 a seat. Check heygen.com/pricing for current prices.

### Does SceneMixer make avatar presenter videos?

No. It makes scenes in which characters act and speak to each other; there is no presenter-to-camera mode. Native dialogue vs dubbing.

### How many languages can SceneMixer characters speak?

Fifteen. Pick one in step 1 and the lines are performed in it. Choose the dialogue language.

## Sources and corrections

- Pricing Plans for Creators and Marketers – HeyGen (heygen.com/pricing) — what HeyGen is for, plans, credits, languages, video length, models

Facts about HeyGen come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check HeyGen's site before you buy. HeyGen is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Higgsfield Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/higgsfield/
Making a series? How SceneMixer compares with Higgsfield on workflow, consistent characters, dialogue and price, and when Higgsfield is the better fit.

Alternatives

# A Higgsfield Alternative for AI Drama Series

Updated November 3, 2026

Quick answer: If you are making a multi-episode story with the same cast, SceneMixer is built for that job: start from a script, a novel or one sentence, and it produces the cast, an episode outline, a storyboard for every episode and the video segment by segment, then joins each episode into one file. Higgsfield describes itself as an AI-native creative suite with 35+ models and tools, which makes it the closer fit if you want to choose a model for each shot, train a character identity from photos or produce ads. SceneMixer video starts at about $0.06 a second on pay-as-you-go credits, and plan credits do not expire at the end of the month.

## Two different centres of gravity

Higgsfield describes itself as a creative suite: 35+ models and tools in one account, with studios built on top of them for camera and lens control, character identity, lip sync, ads, and an agent that runs production from a single chat1. SceneMixer is organised around a project: one script becomes one show, and every later step — reference art, storyboards, video, compositing — inherits the same cast and locations. Neither approach is better in general; they suit different work.

## At a glance

Higgsfield facts as stated on its own pages, checked November 2, 2026

| | Higgsfield | SceneMixer |

| Built around | An AI-native creative suite with 35+ models and tools1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Where you start | Supercomputer, its agent, plans and runs production from a single chat; Cinema Studio 4.0 takes up to 50 references1 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Keeping a character | Soul ID trains a persistent identity from 20+ reference photos1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Camera | Cinema Studio 4.0: genre, lens, aperture and lighting controls, 30+ camera movement presets1 | The camera move is written into each storyboard shot as text you can edit, then that one segment is regenerated |

| Length per generation | Cinema Studio clips up to 30 seconds1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Dialogue | Some models, such as Veo 3.1, generate audio with the video; Seed Audio 1.0 and Lipsync Studio cover speech and lip sync1 | Lines are performed in the generated video itself, in any of 15 languages |

| Price to start | Free plan with limited use; Starter $19/month billed yearly (270 credits a month); Plus $59/month or $47 billed yearly (1,200); Ultra $129 or $99 billed yearly (3,000)2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Video price | Credits per generation vary by model (its table lists about 7 credits per 5 seconds of Kling 3.0 at 720p and about 22 for Seedance 2.0 at 720p); plans include 7-day or 30-day “unlimited” windows on some models, on the website only2 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| Unused plan credits | “Individual-plan credits reset each month”1 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Starfall Command”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Higgsfield if…

- You want to choose the model for each shot. Higgsfield puts 35+ models and tools in one account1; SceneMixer offers a smaller set of video tiers that differ by resolution, segment length and price.

- You need an identity trained from photos of a real person. That is what Soul ID does1.

- Your work is mostly ads, product shots or influencer content. Marketing Studio builds ads from a product URL and Ads Studio makes image ads1.

- You want lens-level camera control on a single clip. Cinema Studio exposes genre, lens, aperture and camera presets directly1.

- You generate in high volume and the unlimited model windows in its plans cover the models you use2.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Higgsfield lands in SceneMixer

| In Higgsfield | Closest step in SceneMixer |

| Soul ID | Character sheets in step 2, drawn from the script's description; on a paid plan you can upload an image instead. Regenerate or replace one sheet |

| Cinema Studio camera settings | The camera line of each storyboard shot. How to write one shot |

| Supercomputer (production from a chat) | AI Create drafts a script from one sentence; the project then runs step by step and you can stop and edit at each one |

| Lipsync Studio and Seed Audio | Dialogue performed in the video itself, in the project's language. Lines, voice-over and on-screen text |

| Marketing Studio / Ads Studio | The ad video mode, a separate project type. AI ad video generator |

## Moving a story over

- Bring the script. Paste it, upload .txt, .docx or .md, or start from one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Higgsfield alternative?

For story work, yes. If you are making a multi-episode drama with a fixed cast from a script or novel, SceneMixer runs that whole pipeline in one project. For choosing a model per shot, training an identity from photos or producing ads, Higgsfield's tools are closer to the job.

### How much does Higgsfield cost?

On November 2, 2026, Higgsfield's pricing page listed a Free plan with limited use, Starter at $19 a month billed yearly (270 credits), Plus at $59 a month or $47 billed yearly (1,200) and Ultra at $129 or $99 billed yearly (3,000). Check higgsfield.ai/pricing for current prices.

### Is there a free Higgsfield plan?

Its pricing page, checked on November 2, 2026, lists a Free plan described as limited use next to the paid plans. SceneMixer gives new accounts 50 free credits and 15 free reference images.

### Do Higgsfield credits roll over?

Higgsfield's own comparison page says individual-plan credits reset each month. In SceneMixer, plan credits go into the same balance as bought credits and do not expire at the end of the month.

### How much does SceneMixer video cost per second?

From about $0.06 per second at 480P up to $0.47 on the most expensive tier, at the lowest credit price. Credit packs start at $1.49.

### Can I use characters I made in Higgsfield?

Yes, if you have the rights to the images: on a paid plan you can upload your own image to SceneMixer in place of a generated character sheet, and every segment that features that character uses it as the reference.

### Can I choose the video model in SceneMixer?

You choose a video tier. Tiers differ in resolution, maximum segment length and price per second, and you can regenerate any single segment on another tier.

## Sources and corrections

- Higgsfield AI vs Runway 2026: Plans and Pricing Compared (higgsfield.ai blog) — Higgsfield's own description of its tools and credit rules
- Pricing plans (higgsfield.ai/pricing) — plan prices; where the blog's plan prices differ, this page is used

Facts about Higgsfield come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Higgsfield's site before you buy. Higgsfield is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# invideo AI Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/invideo/
invideo AI pairs agents with an editor and stock footage; SceneMixer generates an episodic drama from a script. How they compare, and when invideo fits better.

Alternatives

# An invideo AI Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: invideo AI combines AI agents and generative models with a full editing timeline, stock footage, your own footage and 4K export — a general video creation platform. SceneMixer generates every frame of a drama from a script: cast, episode outline, storyboards and video, with the characters performing their lines on screen in 15 languages. Choose invideo if you edit with footage and stock; choose SceneMixer when the story itself must be generated as a series.

## Edit a video, or generate a show

invideo's plans include an editing timeline with colour, audio and 3D tools, your own footage with 4K export, premium stock clips and multiplayer editing seats1. SceneMixer has no stock library and does not edit uploaded footage; it renders new video from the storyboard, so every shot belongs to the same cast and the same world.

## At a glance

invideo AI facts as stated on its own pages, checked November 2, 2026

| | invideo AI | SceneMixer |

| Built around | AI agents and generative models plus an editing timeline, stock footage and your own footage1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Editing | Full editing timeline; colour, audio and 3D tools; multiplayer editing seats1 | Edit the storyboard text and regenerate one segment; composite the episode; finishing edits happen after export |

| Footage | Your own footage with unlimited export; premium stock clips (100 on Starter); 4K via Seedance 2.0 on Plus1 | Every shot is generated; no stock library |

| Voice | An AI voice generator and voice cloning offered as separate tools; not listed in the plan comparison1 | Lines are performed in the generated video itself, in any of 15 languages |

| Price to start | Starter $20 per seat a month billed monthly ($10 billed annually), 400 credits; Plus $60 ($30), 2,000 credits1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Unused credits | “Unused credits don't roll over to the next month”1 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Last Envelope”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with invideo AI if…

- You edit with your own footage or stock. invideo includes a timeline editor, stock clips and 4K export of your own footage1.

- Several people edit together. Plans include multiplayer editing seats1.

- You make explainers, ads or social videos from existing material rather than fully generated scenes1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where invideo AI lands in SceneMixer

| In invideo AI | Closest step in SceneMixer |

| AI agents | AI Create and the script read: outline, cast and storyboard from your text |

| Editing timeline | Segment list along the bottom of the episode editor; regenerate one, then composite |

| Voice tools | Lines performed in the video in the project's language. Lines, voice-over and on-screen text |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer an invideo AI alternative?

For generating a drama series with a fixed cast from a script, yes. For editing your own footage, using stock clips or team editing, invideo is built around those.

### How much does invideo AI cost?

On November 2, 2026, its pricing page listed Starter at $20 per seat a month billed monthly or $10 billed annually (400 credits), Plus at $60 or $30 (2,000 credits) and Max at $250 (8,000 credits). Check invideo.io/pricing for current prices.

### Do invideo credits roll over?

Its pricing page says unused credits don't roll over to the next month. In SceneMixer, plan credits and bought credits do not expire at the end of the month.

### Can I upload my own footage to SceneMixer?

Not to edit it. You can upload a clip to Video Remake, which reads it shot by shot and rebuilds it with your own cast; the footage itself is never handed to the video model. Remake a video with AI.

### Does SceneMixer use stock footage?

No. Every shot is generated from the storyboard, so it matches the cast and locations of the project.

## Sources and corrections

- Pricing | invideo AI (invideo.io/pricing) — plans, credits, editor features, models and credit rules

Facts about invideo AI come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check invideo AI's site before you buy. invideo AI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# JoggAI Alternative for Story Video With Characters | SceneMixer

URL: https://scenemixer.com/alternatives/jogg/
JoggAI makes avatar-led videos for marketing, training and podcasts, with a free plan; SceneMixer makes dramas with a cast that acts. When to pick which.

Alternatives

# A JoggAI Alternative for Story Video With Characters

Updated November 1, 2026

Quick answer: JoggAI calls itself an “AI Video Generator with Lifelike AI Avatars”: a presenter speaks your script to camera for marketing, training, podcasts and product videos, with talking photos, translation and product-ad tools around it. SceneMixer makes dramas: a cast of characters drawn once who act out scenes, with the lines performed in the shot, in 15 languages. Choose JoggAI when one presenter explaining something is the video; choose SceneMixer when the video is a story with several people in a place.

## A presenter, or a cast

JoggAI's free plan lists 200 credits valid for 31 days, about one minute of video, a 3-minute maximum length and one custom avatar; paid plans raise the maximum to 30 and 60 minutes and include watermark removal2. SceneMixer has no avatar: the people on screen are characters from your script, and the length is built from segments.

## At a glance

JoggAI facts as stated on its own pages, checked November 1, 2026

| | JoggAI | SceneMixer |

| Built around | AI avatars presenting a script: marketing, training, podcast and product videos; talking photos; translation1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Who speaks | A stock, custom or photo-based avatar; paid plans include voice clones1 | Lines are performed in the generated video itself, in any of 15 languages |

| Length | Free plan 3 minutes per video; Starter and Creator 30 minutes; Team 60 minutes2 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Plans | Free $0 (200 credits, 31 days); Starter from $24/month billed yearly (144,000 credits a year); Creator $40.66/month; Team $215/month for 3 seats2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Watermark | Removal listed from Starter upward; not included on Free2 | Members download without the badge; shared links keep it |

| Free start | “No credit card required”; talking photo, translation and product tools need a paid plan12 | 50 free credits and 15 free reference images, no card |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Ikrar Jakarta”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with JoggAI if…

- One presenter to camera is the whole video — training, onboarding, a podcast1.

- You need translation of an existing video or a talking photo — JoggAI lists both as tools1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where JoggAI lands in SceneMixer

| In JoggAI | Closest step in SceneMixer |

| Avatar reading a script | Characters acting the scene; narration where you mark it as voice-over |

| Custom avatar from a photo | A character sheet; paid plans can upload a reference image. Regenerate or replace one sheet |

| Video translation | Dialogue language set per project; the lines are performed in it. Choose the dialogue language |

## Moving a story over

- Turn the script into a scene — who is in the room, what they say to each other, not to the camera.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a JoggAI alternative?

For story video with a cast acting in a place, yes. For avatar presenters, translation and talking photos, JoggAI is built around that.

### How much does JoggAI cost?

On November 1, 2026 its pricing page listed a free plan (200 credits, 31 days, 3-minute videos), Starter from $24 a month billed yearly, Creator at $40.66 and Team at $215 for three seats; its FAQ quoted a different Starter figure, so check jogg.ai/pricing.

### Can SceneMixer make a presenter video?

A single character can address the camera in a scene, but there is no avatar library; the character is drawn from your description.

## Sources and corrections

- JoggAI (jogg.ai) — what the product is, features, who it is for
- Pricing – JoggAI (jogg.ai/pricing) — plans, credits, video length, watermark

Facts about JoggAI come only from the pages listed above, as checked on November 1, 2026. Prices and plans change often, so check JoggAI's site before you buy. JoggAI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Katalist Alternative for Script to Storyboard to Video | SceneMixer

URL: https://scenemixer.com/alternatives/katalist/
Katalist turns a script into a storyboard and then into ad videos, from $19 a month; SceneMixer does the same for episodic drama. When to pick which.

Alternatives

# A Katalist Alternative for Script to Storyboard to Video

Updated November 1, 2026

Quick answer: Katalist is “an AI creative studio for performance teams”: it breaks a script into shots, generates a storyboard with consistent characters and scenes, turns shots into video, and can swap your product into a winning creative. SceneMixer runs the same script-to-storyboard-to-video road for drama: a cast drawn once, an episode outline, a storyboard per episode, video segment by segment with the dialogue performed. Choose Katalist for ad creative at agency volume, with Premiere Pro export and product swaps; choose SceneMixer for a story told in episodes.

## Storyboards for ads, or storyboards for episodes

Katalist sizes its plans in image and video credits: Essential lists 200 image credits and 700 video credits a month with 5 projects and 1 custom character; Pro 700 and 1,500 with 20 projects and 10 custom characters2. SceneMixer has no project cap and no character cap; the first reference images on an account are free and the rest are priced per image.

## At a glance

Katalist facts as stated on its own pages, checked November 1, 2026

| | Katalist | SceneMixer |

| Built around | Script to storyboard to video for marketers and agencies; product swaps into winning creatives; UGC and promo videos1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Characters across shots | “Consistent across every shot”; a character cast from a reference keeps the same face through the storyboard and video1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Video models | Runway Gen-4, Google Veo 3.2, Kling, Pixverse v4, WAN, Sora 2, LTX-2 listed2 | Tiers by resolution and price per second; the model behind each tier is managed for you |

| Plans | Essential $19/month; Pro $39/month; Unlimited $99/month (2 seats); Enterprise on request; an annual block lists different figures2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Exports | Premiere Pro integration on Pro; ZIP and .ppt on Unlimited; video and FCP export on Enterprise2 | Video files per episode; storyboard text copies out as plain text |

| Free start | “Sign up free and get free credits to generate your first storyboard, no credit card needed”; a 7-day trial is also mentioned12 | 50 free credits, 15 free reference images and a free first-episode storyboard, no card |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Veil at Dawn”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Katalist if…

- You make ad creative for clients and need product swaps, variants and Premiere Pro export1.

- You want to pick the video model per shot from a listed set2.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Katalist lands in SceneMixer

| In Katalist | Closest step in SceneMixer |

| Script to storyboard | The read plus the storyboard: shots with seconds, framing, who is in frame and the lines. Upload a script and read it |

| Custom characters (1 / 10 / unlimited) | Character sheets, as many as the script has; paid plans can upload reference art. Regenerate or replace one sheet |

| Product swap | An Ad Video project from one to four product photos. AI ad video generator |

## Moving a story over

- Bring the screenplay as written — scene headings and character cues are read.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Katalist alternative?

For script-to-storyboard-to-video on a story in episodes, yes. For ad creative at agency volume with Premiere Pro export and product swaps, Katalist is built around that.

### How much does Katalist cost?

On November 1, 2026 its pricing page listed Essential at $19 a month, Pro at $39 and Unlimited at $99 for two seats, with an annual block showing different figures; check katalist.ai/pricing.

### Can SceneMixer export to Premiere Pro?

No. You download finished video files per episode and segment; the storyboard is text you can copy.

## Sources and corrections

- Katalist (katalist.ai) — what the product does, inputs, consistency, free trial
- Pricing – Katalist (katalist.ai/pricing) — plans, credits, projects, characters, exports

Facts about Katalist come only from the pages listed above, as checked on November 1, 2026. Prices and plans change often, so check Katalist's site before you buy. Katalist is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Kling AI Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/kling/
Kling AI makes clips with its own models; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when Kling fits better.

Alternatives

# A Kling AI Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Kling AI is the home of the Kling models: Kling 3.0 and 3.0 Omni with synchronized audio-video generation, intelligent storyboarding and element reference, plus motion control and native 4K. SceneMixer starts one level up, from a script: it writes the episode outline, draws the cast once, writes a storyboard for every episode and renders the video segment by segment, with lines in 15 languages. Choose Kling for its models and clip-level control; choose SceneMixer to produce a series.

## A model maker's studio, or a series pipeline

Kling's own plans centre on its models — multi-shot generation with Element AI, motion control from captured movement, 1080p and 4K output1. SceneMixer centres on the story: the script decides the episodes, the episodes decide the shots, and every shot is rendered against the same character sheets.

## Kling 4.0 and Kling 4.0 Flash: what was announced

The next generation was announced on September 28, 2026: Kling 4.0, with a lighter Kling 4.0 Flash beside it. The announced specification moves the ceilings again —
takes of 3–30 seconds in one generation (Flash: 3–20), resolution up to 4K (Flash: 720p only, 8-bit SDR), first and last frame plus up to 10 keyframes anywhere on the timeline,
and Omni Reference with up to 15 references per task (10 images, 5 videos up to 30 seconds combined, 7 subjects, plus a voice). Audio is announced as stereo with tighter lip sync and 9 languages for dialogue and on-screen text,
and an editing pass can take up to 5 input clips and rewrite a single expression, camera, background or line2.

What has not happened is a release. Kling 4.0 Flash opened on announcement day to a limited group of creators. The full model was announced for October with no specific date; as of our check on November 2, 2026, we found no report of it going live, no open API and no published pricing for either model.
Extension to a 2-minute maximum and 10-bit HDR are announced as coming soon. Everything in this section is the officially announced specification, not a measurement — nobody outside Kling has run a released version.

Kling 4.0 and 4.0 Flash, as announced on September 28, 20262

| | Kling 4.0 Flash | Kling 4.0 |

| Take length | 3–20 seconds | 3–30 seconds |

| Resolution | 720p, 8-bit SDR | 720p, 1080p, 4K; 10-bit HDR coming soon |

| Control | First-frame image-to-video; no keyframe control in the announcement | First and last frame, up to 10 keyframes, editing over up to 5 clips |

| Status | Limited creator preview from announcement day | Live: listed on Kling's membership page with Omni reference, native 30-second generation and multiple keyframes (checked November 2, 2026) |

Kling 4.0 in detail · Kling 4.0 Flash in detail · what the 4.0 announcement changes.

## At a glance

Kling AI facts as stated on its own pages, checked November 2, 2026

| | Kling AI | SceneMixer |

| Built around | Kling 4.0 (“dynamic motions, Omni reference, native 30s-generation and multiple keyframes”) and Kling 3.0 / 3.0 Omni (“synchronized audio-video generation, intelligent storyboarding and element reference”)1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Characters | Element creation (30 on Basic, up to 500 on Ultra+) and Element AI Multi-Shot1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Motion | Motion Control recreates movement and facial expression from motion capture1 | Actions and staging written into each storyboard shot |

| Price to start | Free Basic plan; Standard $10/month list, $6.99 for the first month then $8.80 a month, 660 credits a month1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Commercial use | “Generated content is for commercial use” on paid plans1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Neon Debt”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Kling AI if…

- You want Kling's models straight from their maker, including native 4K1.

- You animate from captured movement. Motion Control is built for that1.

- You work clip by clip and want daily free credits as a subscriber1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Kling AI lands in SceneMixer

| In Kling AI | Closest step in SceneMixer |

| Elements | Character, scene and prop sheets in step 2. Regenerate or replace one sheet |

| Intelligent storyboarding | A storyboard written for every episode from the script. How to write one shot |

| Synchronized audio | Lines performed in the video in the project's language. Native dialogue vs dubbing |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Kling AI alternative?

For turning a script into a multi-episode drama, yes. For Kling's own models, motion control or native 4K, Kling AI is the direct route.

### How much does Kling AI cost?

On November 2, 2026, its membership page listed a free Basic plan and Standard at $10 a month (first month $6.99, then $8.80 a month on the plan shown) with 660 credits, then Pro at $37 (3,000 credits), Premier at $207 (18,000) and Ultra+ at $619 (57,000), each with a 30% first-month offer. Check Kling's site for current prices.

### Can I use SceneMixer results commercially?

Yes. Our terms give you ownership of the results, including commercial use, on free allowances too.

### Do SceneMixer credits expire?

Plan credits and bought credits sit in one balance and do not expire at the end of the month.

## Sources and corrections

- Kling AI: Membership Plans, Benefits & Account Options (kling.ai) — plans, credits and the features of Kling 3.0
- Kling: 4.0 model release page (kling.ai) — the September 28, 2026 announcement of Kling 4.0 and 4.0 Flash
- Runware: Kling 4 specification table — a restatement of the announced 4.0 and Flash specifications

Facts about Kling AI come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Kling AI's site before you buy. Kling AI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Kling 4.0: Specs, Release Status and an Alternative | SceneMixer

URL: https://scenemixer.com/alternatives/kling-4-0/
Kling 4.0 was announced on September 28, 2026: 3-30 second takes, 4K, up to 10 keyframes. What was announced, what is missing, and where a series pipeline fits.

Alternatives

# Kling 4.0: What Was Announced, and Where SceneMixer Fits

Updated November 2, 2026

Quick answer: Kling 4.0 is the next generation of Kling’s video model, announced on September 28, 2026 alongside a lighter Kling 4.0 Flash: takes of 3–30 seconds, resolution up to 4K, first and last frame plus up to 10 keyframes, Omni Reference with up to 15 references, stereo audio in 9 languages, and an editing pass over up to 5 clips. The full model was announced for October with no specific date, and as of November 2, 2026 it had not gone live, published no pricing and opened no API. This page covers the announcement, and where an episodic pipeline fits around it.

## The announced specifications

This is the full Kling 4.0 specification as announced on September 28, 2026, restated from the release material. Where an item was announced as later work, it says so.

| Announced | Kling 4.0 |

| Take length | 3–30 seconds in one generation; extension to 2 minutes marked coming soon |

| Resolution | 720p, 1080p and 4K; 10-bit HDR at 1080p and 4K marked coming soon |

| Aspect ratios | 21:9, 16:9, 1:1, 9:16, Auto |

| Frame control | First and last frame, plus up to 10 keyframes anywhere on the timeline |

| References | Omni Reference, up to 15 per task: 10 images, 5 videos (30 seconds combined), 7 subjects (up to 3 video subjects), plus a voice |

| Audio | Stereo, tighter lip sync; dialogue and on-screen text in 9 languages with accents and dialects |

| Editing | Up to 5 input clips (one main), 30 seconds in total; single expression, camera, background or line re-edits |

| Prompt | Up to 8,000 tokens |

The number that matters most for series work is the take length. A beat that used to be two or three generations can be one, and the joins between takes — where framing, light and faces drift — get fewer.

## Announced for later, and simply unknown

Three items in the announcement are explicitly later work: extension to a 2-minute maximum, 10-bit HDR at 1080p and 4K, and the full model's release itself (announced for October, no date).

Two things are unknown rather than promised: pricing (nothing published for either model) and the API (not open at the time of writing). Any per-second figure you see for Kling 4.0 today is not an official number.

## Kling 4.0 or Kling 4.0 Flash?

Flash is the 720p, 3–20 second, 8-bit SDR model positioned for high-frequency creation, with first-frame image-to-video and Omni Reference (limits not announced) and no keyframe or editing layer in the announcement. Kling 4.0 Flash in detail.

If you need takes past 20 seconds, output above 720p, keyframes to pin a beat, or the editing pass, the announced answer is the full model. The two are compared line by line in Kling 4.0 vs Kling 4.0 Flash.

## Where SceneMixer fits

This site starts one level up from a single clip: a script is read into an episode outline, the cast is drawn once as reference images, each episode gets a storyboard of numbered shots with their seconds and their lines, and those shots render segment by segment into one file per episode.
The video models available here — Wan 3.0, Seedance 2.5, Seedance 2.0 and MiniMax H3 — are chosen per segment, and each carries its own price per second on the pricing page.

That layer is the part a new model plugs into rather than replaces: cast sheets keep a face across episodes, the storyboard decides what each shot contains, and the composite joins whatever the segments produced.
How to make an AI short drama walks the whole pipeline.

## If you are deciding today

- Wait for the full model if your shots need pinned beats, 4K or long takes — and you can hold the work until the October release actually ships.

- Consider Flash when it opens more widely if your destination is a phone feed and your takes fit inside 20 seconds.

- Do not plan around prices yet. Neither model has published one, and no official per-second rate exists to compare.

- Start the story now. Script, cast and storyboard are the parts that survive whichever model renders the shots.

## Frequently asked questions

### Is Kling 4.0 released?

Not as far as we could establish by November 2, 2026. It was announced on September 28, 2026 for October, with no specific date; no pricing and no API access had been published. Kling 4.0 Flash opened the same day to a limited group of creators.

### What are Kling 4.0's announced specs?

Takes of 3-30 seconds, resolution to 4K, first and last frame plus up to 10 keyframes, Omni Reference with up to 15 references per task, stereo audio in 9 languages, and an editing pass over up to 5 clips. A 2-minute extension and 10-bit HDR are announced as coming soon.

### How much will Kling 4.0 cost?

Unknown. No pricing was published with the announcement, for either version.

### What is the difference between Kling 4.0 and 4.0 Flash?

Flash is the 720p, 20-second, 8-bit SDR version for high-volume work, without the keyframe and editing layers in the announcement. The full model goes to 4K and 30 seconds with 10 keyframes.

### Is SceneMixer a Kling 4.0 alternative?

For turning a script into a multi-episode series, yes - that is what this site is for, whatever renders the segments. For the announced 4.0 capabilities themselves, Kling is the direct route once it ships.

### Can I use SceneMixer results commercially?

Yes. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Sources and corrections

- Kling: 4.0 model release page (kling.ai) — the September 28, 2026 announcement and the officially announced specifications
- Runware: Kling 4 specification table — a restatement of the announced specs for Kling 4.0 and 4.0 Flash
- OrcaRouter: Kling 4.0 release-date roundup (October 4, 2026) — release status: full version still to come, Flash in limited preview

Facts about Kling 4.0 come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Kling's site before you buy. Kling AI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Kling 4.0 Flash: 720p, 20 Seconds, Who It Suits | SceneMixer

URL: https://scenemixer.com/alternatives/kling-4-0-flash/
Kling 4.0 Flash is the announced 720p, 20-second model built for volume. Its specs, what it leaves out, and who should wait for the full Kling 4.0 instead.

Alternatives

# Kling 4.0 Flash: The Announced 720p, 20-Second Version

Updated November 2, 2026

Quick answer: Kling 4.0 Flash is the lighter model announced on September 28, 2026 beside the full Kling 4.0, positioned in the release material as the version built for high-frequency creation: takes of 3–20 seconds, 720p output, 8-bit SDR, text-to-video, first-frame image-to-video and Omni Reference with unpublished limits, and no first-and-last-frame or keyframe control in the announcement. It opened to a limited group of creators on announcement day; pricing and general availability have not been published.

## What was announced for Flash

Kling 4.0 Flash was announced on September 28, 2026 alongside the full Kling 4.0, and it is positioned in the release material as the version built for high-frequency creation — many takes, quickly, at a size you can review on a phone.

| Announced | Kling 4.0 Flash |

| Take length | 3–20 seconds |

| Resolution | 720p only |

| Colour | 8-bit SDR |

| Modes | Text-to-video, first-frame image-to-video, Omni Reference (limits not announced) |

| Frame control | No first-and-last-frame and no multi-keyframe control in the announcement |

| Status | Early access for a limited group of creators from September 28, 2026 |

Audio and editing are not stated either way in the announcement, and the Omni Reference limits were not published for Flash. Silence is not a “no”, but it is also not something to plan around until release notes state it.

## Why a lighter model exists

The reasoning is in the phrase itself: high-frequency creation. Storyboard previews, timing checks, animatics before the expensive pass, and a steady output of cuts that will only ever be watched on a phone.

Most of the cost of a series is in scenes you shoot more than once, so a cheap, fast take you can actually look at is worth more than a beautiful one you were afraid to try. What Flash cannot do, per the announcement, is finish a shot that needs to land a specific beat inside the take or survive a crop to a larger screen.

## What it leaves out, versus the full Kling 4.0

- Ten more seconds per take. 20 against 30, which in a dialogue scene is the difference between three or four takes and five or more.

- Any resolution above 720p. The full model is announced at 720p, 1080p and 4K, with 10-bit HDR coming.

- Keyframes. Up to 10 pinned moments inside a take are announced for the full model only.

- The editing pass. Up to 5 input clips and single-line or single-expression re-edits: announced for the full model.

The full comparison, item by item, is in Kling 4.0 vs Kling 4.0 Flash.

## Where SceneMixer fits

This site is built for the volume side of the same problem: one read of a script returns the cast, locations, props and an episode-by-episode outline; each episode gets a storyboard of numbered shots with their seconds and lines; segments render one at a time and composite into one file per episode.
The video models in the editor — Wan 3.0, Seedance 2.5, Seedance 2.0 and MiniMax H3 — are picked per segment, and the price of the segment is shown before anything is charged.

Drafting cheap and finishing the keepers is the working pattern on this site, at whatever resolutions the tiers offer: Draft at 480P, finish at 1080P.

## Who should wait for the full model

- Large-screen deliverables. 720p does not survive a television, a projector or a crop.

- Shots that must land on a beat. Without keyframes, the way to fix a take is to regenerate it and hope.

- Work you expect to re-edit. The editing pass is announced for the full model.

- Anyone planning a budget. Neither model has published a price, so “Flash will be cheaper” is an expectation, not a number.

## Frequently asked questions

### Is Kling 4.0 Flash available?

It opened on September 28, 2026 as an early-access preview for a limited group of creators. Whether it has widened since is a question for Kling's own release page; announcement-day material does not settle it.

### What are Kling 4.0 Flash's specs?

3-20 second takes, 720p output, 8-bit SDR, text-to-video, first-frame image-to-video and Omni Reference with unpublished limits. No first-and-last-frame or multi-keyframe control in the announcement.

### Does Flash have audio?

The announcement does not say. Treat it as unknown until release notes state it.

### Can Flash do keyframes or editing?

Neither is in the announcement. Keyframes (up to 10) and the 5-clip editing pass are announced for the full Kling 4.0.

### Will Flash be cheaper than the full model?

It is positioned as the lighter, high-volume version, but no prices were published for either model, so the size of any gap is unknown.

### Is SceneMixer a Kling 4.0 Flash alternative?

For episodic work, yes - a script becomes a cast, a storyboard per episode and segments composited into one file, whatever model renders them. For Flash's own announced capabilities, Kling is the direct route.

### Can I use SceneMixer results commercially?

Yes. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Sources and corrections

- Kling: 4.0 model release page (kling.ai) — the September 28, 2026 announcement and the officially announced specifications
- Runware: Kling 4 specification table — a restatement of the announced specs for Kling 4.0 and 4.0 Flash
- OrcaRouter: Kling 4.0 release-date roundup (October 4, 2026) — release status: full version still to come, Flash in limited preview

Facts about Kling 4.0 Flash come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Kling's site before you buy. Kling AI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Krea Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/krea/
Krea is a creative suite for images, video, 3D and LoRA training; SceneMixer turns a script into an episodic drama. How they compare, and when Krea fits.

Alternatives

# A Krea Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Krea is a creative suite: real-time image generation, all major image, video, 3D and lip-sync models, LoRA training on your own characters and styles, Nodes workflows and an Agent that plans and iterates across models. SceneMixer is a pipeline for one kind of work: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with lines performed on screen in 15 languages. Choose Krea for visual exploration and custom styles; choose SceneMixer for a drama series.

## A visual playground, or a series pipeline

Krea's plans give compute units across its models, with real-time generation, LoRA training and workflow tools on every tier1. SceneMixer keeps the choices narrower so a whole series can run from one script: the look is set once in step 1, and every sheet and shot follows it.

## At a glance

Krea facts as stated on its own pages, checked November 2, 2026

| | Krea | SceneMixer |

| Built around | Image, video, 3D and lip-sync models with real-time generation, Nodes workflows and the Krea Agent1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Custom looks | LoRA training on your own characters, styles or objects (50 images; 2,000 on Max)1 | Visual style chosen in step 1; your own character image can replace a sheet on a paid plan |

| Licence | Commercial licence on every plan, including Free1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

| Price to start | Free: 100 units a day; Pro $35/month or $21 billed yearly with 20,000 units a month1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Top-ups | One-time compute packs, valid for 90 days1 | Credit packs from $1.49; credits do not expire at the end of the month |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rocket Pup”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Krea if…

- You explore looks in real time or train LoRAs on your own characters and styles1.

- You build node-based workflows or reusable apps1.

- You work across image, video and 3D in one account1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Krea lands in SceneMixer

| In Krea | Closest step in SceneMixer |

| LoRA training | Character sheets generated once per character; upload your own on a paid plan. Regenerate or replace one sheet |

| Krea Agent | AI Create and the script read, which plan the whole series |

| Nodes | The fixed five-step project pipeline |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Krea alternative?

For producing a drama series from a script, yes. For real-time image work, LoRA training, 3D or node-based workflows, Krea is built around those.

### How much does Krea cost?

On November 2, 2026, its pricing page listed Free with 100 units a day, Basic at $9 a month or $5 billed yearly (5,000 units), Pro at $35 or $21 (20,000) and Max at $105 or $63 (60,000). Check krea.ai/pricing for current prices.

### Do SceneMixer credits expire?

Plan credits and bought credits sit in one balance and do not expire at the end of the month.

## Sources and corrections

- Krea Pricing | AI Creative Tools for Everyone (krea.ai/pricing) — plans, units, features and compute packs

Facts about Krea come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Krea's site before you buy. Krea is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Leonardo AI Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/leonardo/
Leonardo.Ai is an image-first platform with video and animation on token plans; SceneMixer turns a script into an episodic drama. How they compare.

Alternatives

# A Leonardo AI Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Leonardo.Ai is an image-first creative platform that also makes video: text-to-video and image-to-video with start and end frames, animation workflows and Blueprints (pre-built workflows), sold as monthly token allowances across 80+ models. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Leonardo for images, animation and stills brought to life; choose SceneMixer to produce a series with a fixed cast.

## Images that move, or a story that plays

Leonardo describes its video tool as one that “puts you in control of motion”: prompts or stills in, short sequences out, with start and end frames to anchor the scene2. SceneMixer does not start from a still; each segment is rendered against character sheets drawn once for the whole series.

## At a glance

Leonardo.Ai facts as stated on its own pages, checked November 2, 2026

| | Leonardo.Ai | SceneMixer |

| Built around | “Create across image, video, design, and motion”: one subscription, 80+ models1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Video tools | Text-to-video and image-to-video, start and end frames, animation workflows, Blueprints and a choice of video models2 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Consistency | “Maintain visual consistency across shots and iterations”; personal AI models you train: 10 on Essential, 20 on Premium, 50 on Ultimate12 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Plans | Free: 150 fast tokens a day, Auto model only, public creations. Essential $12 a month (8,500 tokens), Premium $30 (25,000), Ultimate $60 (60,000); yearly billing up to 20% off; prices ex. tax1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Video price | Not stated as a per-second price on the pricing page; Ultimate lists “Video Generation for select models at a relaxed pace”1 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| Teams and API | Team Starter $72 a month ($24 a seat), Growth $144 ($48 a seat); pay-as-you-go API whose credit does not expire1 | One account; plan credits and bought credits sit in one balance and do not expire at the end of the month |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Veil at Dawn”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Leonardo.Ai if…

- You work image-first — 80+ models and personal AI models you train yourself1.

- You animate stills with start and end frames and ready-made Blueprints2.

- You build on an API with pay-as-you-go credit that does not expire1.

- You work in a team with shared tokens and private team generations1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Leonardo.Ai lands in SceneMixer

| In Leonardo.Ai | Closest step in SceneMixer |

| Start frame and end frame | The next segment opens where the last one ended. Shot-to-shot continuity |

| Personal AI models | Character sheets generated once and attached to every segment; paid plans can upload their own art. Regenerate or replace one sheet |

| Blueprints | AI Create and the script read, which plan the whole series. Upload a script and read it |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; stills you made can go in as character art on a paid plan.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Leonardo alternative?

For turning a written story into a multi-episode drama with a fixed cast, yes. For image-first work and animating stills, Leonardo is built around that.

### How much does Leonardo AI cost?

On November 2, 2026, its pricing page listed Free (150 fast tokens a day), Essential at $12 a month (8,500 tokens), Premium at $30 (25,000) and Ultimate at $60 (60,000), excluding tax, with yearly billing up to 20% off. Check leonardo.ai/pricing for current prices.

### Is Leonardo AI free?

Its pricing page, checked November 2, 2026, listed a Free plan with 150 fast tokens a day, the Auto model only and public creations.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15. Native dialogue vs dubbing.

## Sources and corrections

- Leonardo.Ai Pricing (leonardo.ai/pricing) — plans, tokens, models, personal AI models, team and API plans
- AI Video Generator – Leonardo.Ai (leonardo.ai/ai-video-generator) — text and image to video, start and end frames, Blueprints

Facts about Leonardo.Ai come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Leonardo.Ai's site before you buy. Leonardo.Ai is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# LlamaGen Alternative for Comic Stories as Video | SceneMixer

URL: https://scenemixer.com/alternatives/llamagen/
LlamaGen's comic-to-video mode animates finished comic panels; SceneMixer turns a comic script into an acted animated episode. When to pick which.

Alternatives

# A LlamaGen Alternative for Comic Stories as Video

Updated November 1, 2026

Quick answer: LlamaGen calls itself “the leading AI comic generator”: it makes comic strips, manga and anime art, and its Comic to Video mode turns finished panels into a short vertical animated clip that follows panel order and dialogue timing. SceneMixer works from the comic's script rather than its pages: it draws the cast once in a comic look, writes each panel as a shot with seconds and spoken lines, and generates the episode with voices and music. Choose LlamaGen to make the comic itself and give its pages motion; choose SceneMixer to turn the story into acted animation.

## Animate the pages, or act the story

LlamaGen's comic-to-video page describes the output as a “short animated preview” in 9:16, warns that motion can alter faces, hands and lettering, and that lip sync is not guaranteed2. SceneMixer never animates a page: each panel becomes a new shot with the character sheet drawn once, and the lines are performed by a voice per character.

## At a glance

LlamaGen facts as stated on its own pages, checked November 1, 2026

| | LlamaGen | SceneMixer |

| Built around | AI comic, manga and anime art generation; a Comic to Video mode for finished panels12 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Comic to video input | Finished comic art: panels, strips, manga pages, webtoon art; split dense pages into shots2 | The script or panel text; members can add a character reference image |

| Output | A short 9:16 animated clip, MP4 depending on mode and plan; length and resolution vary by mode2 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Dialogue | Dialogue cues and sound can be planned around speech bubbles; exact lip sync and preserved lettering not guaranteed2 | Lines are performed in the generated video itself, in any of 15 languages |

| Plans | Standard $18/month (15,000 credits); Pro $30 (45,000); Unlimited $96 (135,000); Production Unlimited $200 (270,000); annual saves up to 40%1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Free limits | 1 custom character, 1 storyboard project, 1 video, 4 panels per page; Comic to Video is described as a paid-plan mode12 | 50 free credits, 15 free reference images, a free first-episode storyboard |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rust City Knockout”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with LlamaGen if…

- You want to make the comic itself — strips, manga pages, anime art and a layer splitter1.

- You have finished pages and want a teaser that keeps your art — comic-to-video animates the panels you drew2.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where LlamaGen lands in SceneMixer

| In LlamaGen | Closest step in SceneMixer |

| Finished panels | Panel text as a storyboard: each panel a shot with seconds, framing and the line |

| Custom comic character | A character sheet in American Comic, Korean Webcomic or Painted Manhua; paid plans can upload your design. Regenerate or replace one sheet |

| Comic to Video mode | An episode: the shots rendered with voices and music and joined into one file |

## Moving a story over

- Paste the chapter — the panel text with who is in each and what is said.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a LlamaGen alternative?

For turning a comic story into acted animation with spoken lines, yes. For making the comic pages themselves and animating them as drawn, LlamaGen is built around that.

### How much does LlamaGen cost?

On November 1, 2026 its pricing page listed Standard at $18 a month (15,000 credits), Pro at $30, Unlimited at $96 and Production Unlimited at $200, with up to 40% off annually; check llamagen.ai/pricing.

### Does SceneMixer keep my drawn panels?

No. It works from the text; members can upload a character reference so the drawn cast follows your design.

## Sources and corrections

- Pricing – LlamaGen (llamagen.ai/pricing) — plans, credits, free limits, commercial terms
- Comic to Video – LlamaGen (llamagen.ai/comic-to-video) — what the mode does, inputs, limits

Facts about LlamaGen come only from the pages listed above, as checked on November 1, 2026. Prices and plans change often, so check LlamaGen's site before you buy. LlamaGen is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# LTX Studio Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/ltx-studio/
LTX Studio is a suite for storyboards, pitch decks and video; SceneMixer turns a script into episodes. How they compare, and when LTX Studio fits better.

Alternatives

# An LTX Studio Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: LTX Studio is a creative suite for professional creators, marketers and filmmakers: storyboards, pitch decks, camera controls, saved consistent “Elements”, node-based Flows and a wide set of image and video models. SceneMixer is narrower and more automatic: give it a script or a novel and it produces the episode outline, the cast, a storyboard per episode and the video, with dialogue performed on screen in 15 languages. Choose LTX Studio for pre-production, pitching and shot-level control; choose SceneMixer to turn a written story into a series.

## A studio toolkit, or a production line

LTX Studio's plans add tools as you go up: AI storyboards, pitch decks and consistent Elements from Standard, collaborators per project on Pro, custom model training on Enterprise1. SceneMixer is one pipeline that every project runs through, so the storyboard, the cast and the video are always connected to the script that produced them.

## At a glance

LTX Studio facts as stated on its own pages, checked November 2, 2026

| | LTX Studio | SceneMixer |

| Built for | “Professional creator, marketer or filmmaker” plans with storyboards, pitch decks and node-based Flows1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Storyboards | Blank storyboards on Free; AI Storyboards from Standard1 | A storyboard is written for every episode from the script, shot by shot, and stays editable |

| Keeping a character | “Save and create consistent Elements” from Standard1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Camera | Advanced camera controls on every plan1 | The camera move is written into each storyboard shot as text you can edit |

| Dialogue | Dubbing and captions are listed in the plan comparison1 | Lines are performed in the generated video itself, in any of 15 languages |

| Licence | Free: personal use licence; commercial use licence from Standard1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

| Price to start | Free: 800 one-time credits; Lite $15/month ($12 billed yearly) with 8,000 credits; Standard $35/month ($28 billed yearly) with 28,0001 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Eagle's Vow”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with LTX Studio if…

- You are pitching or pre-visualising. Pitch decks and AI storyboards are part of LTX Studio's plans1.

- You want node-based automation. Flows lets you wire generation steps together1.

- You want specific named models such as Veo 3.1 or Kling 3.0, which its plans list1.

- Several people work on one project. Pro includes 3 collaborators per project; Enterprise has unlimited collaborators1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where LTX Studio lands in SceneMixer

| In LTX Studio | Closest step in SceneMixer |

| Elements | Character, scene and prop sheets in step 2. Regenerate or replace one sheet |

| AI Storyboards | The storyboard written for every episode. How to write one shot |

| Camera controls | The camera line of each shot in the storyboard |

| Dubbing | Dialogue performed in the video in the project's language. Native dialogue vs dubbing |

## Moving a story over

- Bring the script you storyboarded — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer an LTX Studio alternative?

For turning a written story into a multi-episode drama, yes. For pitch decks, node-based automation, team collaboration or choosing specific named models, LTX Studio has more tools.

### How much does LTX Studio cost?

On November 2, 2026, its pricing page listed Free (800 one-time credits), Lite at $15 a month or $12 billed yearly (8,000 credits), Standard at $35 or $28 billed yearly (28,000 credits) and Pro at $125 or $100 billed yearly (110,000 credits). Check its pricing page for current prices.

### Can I use LTX Studio's free plan commercially?

Its pricing page lists a personal use licence on Free and a commercial use licence from Standard. SceneMixer's terms give you ownership of results, including commercial use, on free allowances too.

### Does SceneMixer write the storyboard for me?

Yes. Every episode gets a shot-by-shot storyboard written from the script, with durations, the cast in each shot and the lines; you can edit any of it. How to write one shot.

### Do SceneMixer credits expire?

Plan credits and bought credits sit in one balance and do not expire at the end of the month.

## Sources and corrections

- LTX Studio pricing (ltx.io/studio/pricing) — plans, credits, features, models, licences and FAQ

Facts about LTX Studio come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check LTX Studio's site before you buy. LTX Studio is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Luma Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/luma/
Luma pairs its Ray models and Luma Agents with third-party models; SceneMixer turns a script into an episodic drama. How they compare, and when Luma fits.

Alternatives

# A Luma Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Luma offers its own Ray models — text-, image- and video-to-video, reframe and extend, with HDR and EXR output — plus third-party models and Luma Agents, on credit plans with commercial use. SceneMixer is a pipeline for drama: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with lines performed on screen in 15 languages. Choose Luma for high-end shots and finishing formats; choose SceneMixer to produce a series.

## Shot craft, or series production

Luma's pricing lists credits per generation for each model and resolution, from drafts up to 1080p, with HDR and EXR multipliers on Ray3.21. SceneMixer prices by the second per tier, and organises the work as episodes rather than individual generations.

## At a glance

Luma facts as stated on its own pages, checked November 2, 2026

| | Luma | SceneMixer |

| Built around | Its Ray models plus third-party image and video models, and Luma Agents1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Finishing | HDR and HDR+EXR output and reframe on Ray3.2; extend on Ray3.141 | One composited file per episode; resolution follows the tier |

| Collaboration | Edit access for guest collaborators; Team plans with shared credits1 | One account per project owner |

| Price to start | Plus $30/month ($25 billed yearly) with 10,000 credits and commercial use; Pro $90 ($75) with 40,000; no free plan listed1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Starfall Command”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Luma if…

- You need HDR or EXR deliverables. Ray3.2 offers them1.

- You reframe, extend or transform existing video with video-to-video tools1.

- A team shares credits on a Team plan1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Luma lands in SceneMixer

| In Luma | Closest step in SceneMixer |

| Luma Agents | AI Create and the script read, which plan the whole series |

| Image-to-video | Segments rendered against the character sheets of the project |

| Extend | Longer scenes split into segments in the storyboard, joined in the composite |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Luma alternative?

For producing a drama series from a script, yes. For HDR or EXR deliverables, video-to-video work or team credit pools, Luma has more.

### How much does Luma cost?

On November 2, 2026, its pricing page listed Plus at $30 a month or $25 billed yearly (10,000 credits), Pro at $90 or $75 (40,000) and Ultra at $300 or $250 (150,000), with Team and Enterprise plans on request. Check lumalabs.ai/pricing for current prices.

### Can I start SceneMixer for free?

Yes. New accounts get 50 free credits and 15 free reference images, and credit packs start at $1.49.

## Sources and corrections

- Plans & Pricing | Luma (lumalabs.ai/pricing) — plans, credits, models and per-generation costs

Facts about Luma come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Luma's site before you buy. Luma is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# MagicLight Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/magiclight/
MagicLight makes long story videos; SceneMixer makes episodic dramas with a fixed cast. How they compare on format, voice and price, and when MagicLight fits.

Alternatives

# A MagicLight Alternative for AI Drama Series

Updated November 3, 2026

Quick answer: MagicLight and SceneMixer both start from a story, but they aim at different films. MagicLight's pitch is long form — stories “up to 50 minutes with steady characters, scenes, and tone” — with templates for kids stories, comedy skits, faith lessons and explainers. SceneMixer is built for drama in episodes: a cast that performs its lines on screen in one of 15 languages, an outline and a storyboard for every episode, and video rendered segment by segment. Choose MagicLight for one long story video at a low monthly price; choose SceneMixer when the story is a series with dialogue.

## One long film, or a show in episodes

MagicLight describes an all-in-one flow “from script to visuals, voice, and export” and positions itself as built for long form, with plans that list up to 50-minute video generation1. SceneMixer splits a story into episodes, gives each one a storyboard you can edit line by line, and renders the shots against a fixed set of character sheets, so a series keeps its look from the first episode to the last.

## At a glance

MagicLight facts as stated on its own pages, checked November 2, 2026

| | MagicLight | SceneMixer |

| Built for | Long-form story videos, “up to 50 minutes with steady characters, scenes, and tone”1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Where you start | Your story or script, or templates for kids stories, comedy skits, faith lessons, historical explainers and science topics1 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Your own characters and voice | “AI Voice Cloning” is listed as a tool; every paid plan lists free character and image generation1 | Paid plans can upload their own character image; voices are matched to each character from a preset library |

| Dialogue | Script to visuals and voice in one flow1 | Lines are performed in the generated video itself, in any of 15 languages |

| What a plan lists | Standard: 8,000 credits a month, “up to 130 min video generation” and “up to 290 s image-to-video generation”1 | Pro: 700 credits a month; video from 10 credits a second, so the seconds you get depend on the tier you pick |

| Price to start | A free plan with limited credits, no card; Standard $7.50/month billed yearly ($90) with 8,000 credits a month, shown next to a struck-through $151 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Export | 1080P HD video with no watermark on every paid plan1 | One composited file per episode; resolution follows the video tier, 480P to 1080P and above |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rocket Pup”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with MagicLight if…

- You want one long story video rather than episodes. MagicLight's plans list up to 50-minute video generation1.

- You make kids stories, faith lessons or explainers. It has ready-made templates for exactly those1.

- You want the most minutes for a small monthly fee. Its yearly Standard plan lists up to 130 minutes of video generation for $7.50 a month1.

- You want to clone your own voice. Voice cloning is on its paid plans1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- You need landscape as well as vertical. 16:9 is the default and 9:16 is one setting away. Vertical and landscape projects.

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

- You want live-action looks as well as animation. Pick a realistic style or a 2D, 3D or stop-motion one in step 1; the look is drawn into every sheet and shot. Make an AI animated short film.

## Where MagicLight lands in SceneMixer

| In MagicLight | Closest step in SceneMixer |

| Templates (kids stories, comedy skits, explainers) | AI Create drafts a script from one sentence; the genre guides show how to set up each kind of story |

| Steady characters | Character sheets generated once and attached to every segment. How consistency works |

| Voice | Lines performed on screen in the project's language, with a voice matched to each character. Native dialogue vs dubbing |

| Long-form story | A Film-type project with no episode cap, or a series split into episodes with an outline |

## Moving a story over

- Bring the story text. Paste it or upload .txt, .docx or .md; a long story is split into an episode outline you can edit.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a MagicLight alternative?

If your story is a drama in episodes with characters who speak on screen, yes. If you want one long story video or a template-based kids story, MagicLight is built around that.

### How much does MagicLight cost?

On November 2, 2026, MagicLight's pricing page listed Standard at $7.50 a month billed yearly ($90; 8,000 credits a month), Plus at $14 ($168), Pro at $22 ($264), Ultra at $45 ($540) and Ultimate at $75 ($900), each next to a higher struck-through price. Check magiclight.ai/pricing for current prices.

### Is MagicLight free?

Its pricing FAQ says you can start free with limited credits and no card. SceneMixer also starts free: new accounts get 50 credits and 15 free reference images.

### Can SceneMixer make animated stories?

Yes. Choose a 2D, 3D or stop-motion style in step 1 and every character sheet, location and shot is drawn in it. Make an AI animated short film.

### Can I use my own character designs?

Yes, on a paid plan: upload your own image to SceneMixer in place of a generated character sheet, and every segment with that character uses it as the reference.

### Do SceneMixer credits expire?

Plan credits and bought credits sit in one balance and do not expire at the end of the month.

## Sources and corrections

- Pricing – AI Video Generator Plans (magiclight.ai/pricing) — plans, credits, long-form claim, templates and FAQ

Facts about MagicLight come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check MagicLight's site before you buy. MagicLight is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Medeo Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/medeo/
Medeo creates videos through chat with an editor and stock footage; SceneMixer turns a script into an episodic drama. How they compare, and when Medeo fits.

Alternatives

# A Medeo Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Medeo creates videos through a chat, with a video editor, collaborative editing and stock video alongside generative models. SceneMixer turns a script into a drama series: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Medeo for chat-driven videos that mix stock and AI footage; choose SceneMixer for a series.

## Chat-made videos, or a drama in episodes

Medeo's plans state minutes two ways — Pro lists about 3 minutes of full AI video or about 30 minutes with stock video1. SceneMixer generates every shot, so the cast and locations stay the same across the whole project.

## At a glance

Medeo facts as stated on its own pages, checked November 2, 2026

| | Medeo | SceneMixer |

| Built around | Creating videos using chat, with a video editor and collaborative editing tools1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Footage | AI generative features on every plan; Pro lists about 30 minutes with stock video1 | Every shot is generated; no stock library |

| Price to start | Free plan; Pro $39/month ($28 billed annually) with 700 credits (≈ 3 min of full AI video)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Crimson Vow”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Medeo if…

- You want to create by chatting and edit the result in the same place1.

- You mix stock footage with AI shots to stretch your credits1.

- Several people edit together. Collaborative editing is on its plans1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Medeo lands in SceneMixer

| In Medeo | Closest step in SceneMixer |

| Chat | AI Create from one sentence, then the step-by-step project |

| Video editor | Storyboard text and per-segment regeneration; composite the episode |

## Moving a story over

- Bring the story — paste a script, upload a document, or start from one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Medeo alternative?

For a drama series with a fixed cast and on-screen dialogue, yes. For chat-driven videos that mix stock footage and AI shots, Medeo is built around that.

### How much does Medeo cost?

On November 2, 2026, its pricing page listed a free plan, Pro at $39 a month or $28 billed annually (700 credits), Insider at $106 or $76 (2,400) and Elite at $328 or $236 (9,000). Check medeo.app/pricing for current prices.

### Does SceneMixer use stock footage?

No. Every shot is generated from the storyboard, so it matches the cast and locations of the project.

## Sources and corrections

- Pricing - Medeo (medeo.app/pricing) — plans, credits, features and model prices

Facts about Medeo come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Medeo's site before you buy. Medeo is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Midjourney Video Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/midjourney-video/
Midjourney animates its images into 5-second clips you can extend; SceneMixer turns a script into an episodic drama. How they compare, and when Midjourney fits.

Alternatives

# A Midjourney Video Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Midjourney Video animates an image: Midjourney's announcement describes pressing Animate on a Midjourney image, or marking an uploaded image as a start frame, to get four 5-second videos per job, each extendable about 4 seconds at a time, four times. SceneMixer starts from a script instead of an image: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Midjourney to bring its images to life; choose SceneMixer to produce a series from a story.

## Animate a picture, or shoot a story

Midjourney positions video as image-to-video: you start from an image, choose automatic or manual animation and high or low motion, and extend the result1. SceneMixer starts from the written story: the script decides the episodes and shots, and each shot is rendered against the same character sheets.

## At a glance

Midjourney Video facts as stated on its own pages, checked November 2, 2026

| | Midjourney Video | SceneMixer |

| Built around | Animating Midjourney images, or an uploaded image used as a start frame, on the web1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Clip length | Four 5-second videos per job; extend about 4 seconds at a time, four times total1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Motion control | Automatic or manual animation prompt; high or low motion1 | Each shot's action, camera and lines are written in the storyboard |

| Dialogue and sound | Not mentioned in the announcement (checked November 2, 2026)1 | Lines are performed in the generated video itself, in any of 15 languages |

| Cost | A video job costs about 8 times an image job, per the announcement; usage comes from the Midjourney plan1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rust City Knockout”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Midjourney Video if…

- Your look comes from Midjourney images and you want them to move1.

- You already have a Midjourney plan and its time covers your clips1.

- You make short loops and moving stills rather than scenes with dialogue1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want a painted or animated look across a whole series. Pick one of the visual styles in step 1 and every sheet and shot is drawn in it. Make an AI animated short film.

## Where Midjourney Video lands in SceneMixer

| In Midjourney Video | Closest step in SceneMixer |

| Animate an image | Generate a segment from the storyboard; the character sheets are the references. How to write one shot |

| Start frame | Upload your own image in place of a generated sheet, on a paid plan. Regenerate or replace one sheet |

| Extend | The next segment of the storyboard |

## Moving a story over

- Bring the story — a script or a document; on a paid plan, upload character art you own in place of generated sheets.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Midjourney video alternative?

For a story told in episodes with a fixed cast and spoken lines, yes. For animating Midjourney images you already have, Midjourney is the direct route.

### How long are Midjourney videos?

Midjourney's announcement, checked November 2, 2026, describes four 5-second videos per job, each extendable about 4 seconds at a time, four times.

### Does Midjourney video have sound?

The announcement does not mention audio (checked November 2, 2026). Check Midjourney's own documentation for the current state.

### Can I use my Midjourney images in SceneMixer?

On a paid plan you can upload your own image in place of a generated character, location or prop sheet, as long as you have the rights to it.

### Can SceneMixer make anime or painterly series?

Yes. Choose a style in step 1 and every sheet and shot is drawn in it. Make an AI animated short film.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15.

## Sources and corrections

- Introducing Our V1 Video Model (updates.midjourney.com) — Midjourney's announcement of its video model: how it works, clip length, extension and cost relative to images

Facts about Midjourney Video come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Midjourney Video's site before you buy. Midjourney Video is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Mootion Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/mootion/
Mootion turns ideas and scripts into voiced visual stories; SceneMixer turns scripts into episodic dramas. How they compare, and when Mootion fits better.

Alternatives

# A Mootion Alternative for AI Drama Series

Updated November 3, 2026

Quick answer: Mootion turns text, ideas, images and scripts into visual stories with voiceovers, and it has an education plan — a good fit for lessons, explainers and narrated shorts. SceneMixer is built for drama in episodes: an outline with end-of-episode hooks, a fixed cast drawn once, a storyboard per episode and characters who perform their lines on screen in 15 languages. Choose Mootion for narrated stories and teaching; choose SceneMixer for a series.

## Visual storytelling, or episodic drama

Mootion describes generating “expressive visual scenes with synced motion, audio, and narrative flow” and lists 10+ supported languages1; its plans add voiceover quality, scenes per project and idea-to-video length as you go up2. SceneMixer organises the same raw material — a script — into episodes, each with its own storyboard and segments, all sharing one cast.

## At a glance

Mootion facts as stated on its own pages, checked November 2, 2026

| | Mootion | SceneMixer |

| Built for | Turning text, ideas, images and scripts into cinematic visual stories1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Voice | Realistic voiceovers on Standard; ultra-realistic voiceovers from Plus2 | Lines are performed in the generated video itself, in any of 15 languages |

| Project size | Max 64 scenes per project on Plus; 50,000 characters for Idea-to-Video on Pro2 | Episodes from an outline; each episode as many segments as its storyboard needs |

| Price to start | Standard $15/month billed monthly ($12 billed yearly), 1,500 credits; EDU plan $6/month ($5 billed yearly)2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Veil at Dawn”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Mootion if…

- You teach or explain. Mootion has an EDU plan and voiceover-led storytelling2.

- You want narrated visual stories from an idea or a script in a few steps1.

- You want premium voiceovers included as you go up its plans2.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Mootion lands in SceneMixer

| In Mootion | Closest step in SceneMixer |

| AI script editing | Edit every storyboard line (episode outlines are read-only). How to write one shot |

| Voiceovers | Lines performed on screen by the characters, or a voice-over line where you mark one. Lines, voice-over and on-screen text |

| Scenes per project | Episodes and segments; the outline decides how many |

## Moving a story over

- Bring the script or idea — paste it, upload a document, or start from one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Mootion alternative?

For episodic drama with characters who speak on screen, yes. For narrated visual stories, lessons and explainers, Mootion is built around those.

### How much does Mootion cost?

On November 2, 2026, its pricing page listed Standard at $15 a month billed monthly or $12 billed yearly (1,500 credits), Plus at $25 or $20 (3,000), Pro at $50 or $40 (8,000), Max at $200 or $160 (32,000) and an EDU plan at $6 a month or $5 billed yearly. Check its pricing page for current prices.

### Can SceneMixer do voice-over?

Yes. Mark a line as voice-over in the storyboard and it is spoken over the shot; other lines are performed by the characters on screen. Lines, voice-over and on-screen text.

### Which languages does SceneMixer support?

15 languages for the interface, storyboards and on-screen dialogue.

## Sources and corrections

- Mootion – AI Video Generator for Cinematic Visual Storytelling (mootion.com) — positioning and languages
- Mootion AI Video Generator Pricing (mootion.com/pricing.html) — plans, credits and per-plan features

Facts about Mootion come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Mootion's site before you buy. Mootion is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Morphic Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/morphic/
Morphic gives you models and workflows to bring stories to life; SceneMixer turns a script into an episodic drama. How they compare, and when Morphic fits.

Alternatives

# A Morphic Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Morphic offers all its image and video models with workflows, team seats and credit packs, aimed at creatives who want to “bring your stories to life”. SceneMixer is narrower and more automatic: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Morphic to build your own workflows across models; choose SceneMixer to produce a series from a script.

## Build the workflow, or use one

Morphic's plans give monthly credits across all models with workflows on every tier, and shared credits plus extra seats on Pro and Max1. SceneMixer's workflow is fixed — script, sheets, storyboard, video, composite — so a whole series runs from one project.

## At a glance

Morphic facts as stated on its own pages, checked November 2, 2026

| | Morphic | SceneMixer |

| Built around | All models plus workflows; shared credits and extra seats on Pro and Max1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Free tier | Up to 20 credits, limited models1 | 50 free credits and 15 free reference images for new accounts |

| Price to start | Basic $19/month with 2,400 credits; Standard $24 with 3,6251 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Top-ups | Credit packs valid for 180 days, from $25 for 2,500 credits1 | Credit packs from $1.49; credits do not expire at the end of the month |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rocket Pup”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Morphic if…

- You design your own workflows across many models1.

- A small team shares credits with extra seats1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Morphic lands in SceneMixer

| In Morphic | Closest step in SceneMixer |

| Workflows | The fixed five-step project pipeline: read, sheets, storyboard, video, composite |

| Shared credits | One balance per account; plan credits and bought credits do not expire at the end of the month |

## Moving a story over

- Bring the story — a script, a document, or one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Morphic alternative?

For producing a drama series from a script with a fixed cast, yes. For building your own workflows across many models, Morphic is built around that.

### How much does Morphic cost?

On November 2, 2026, its pricing page listed a free tier (up to 20 credits), Basic at $19 a month (2,400 credits), Standard at $24 (3,625), Pro at $45 and Max at $170, with credit packs valid for 180 days. Check morphic.com/pricing for current prices.

### Do SceneMixer credits expire?

Plan credits and bought credits sit in one balance and do not expire at the end of the month.

## Sources and corrections

- Morphic Pricing: AI Video & Image Plans and Credits (morphic.com/pricing) — plans, credits, credit packs and features

Facts about Morphic come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Morphic's site before you buy. Morphic is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# MovieFlow Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/movieflow/
MovieFlow makes short AI videos from a line of text with an agent and several models; SceneMixer turns a script into an episodic drama. How they compare.

Alternatives

# A MovieFlow Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: MovieFlow (movieflow.ai, marked Beta) is an AI movie and video generator with a video agent: “Your idea. A movie. In minutes.” Its home page advertises its own model at USD 0.008 a second and five free 5-second videos for new users, and its plans list MovieFlow-2.0, Seedance 2.5, MiniMax H3, Kling 3.0 and Vidu. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose MovieFlow for fast short clips from a line of text; choose SceneMixer for a series with a fixed cast.

## Clips in seconds, or episodes from a script

MovieFlow sizes its plans in credits, quoting up to 100 five-second clips on Starter and up to 1,430 on Ultra2. The search “movie flow ai” can also mean Google's Flow; this page is about movieflow.ai, and Google Flow has its own page. SceneMixer's unit is the project: every sheet, storyboard and segment belongs to one story.

## At a glance

MovieFlow facts as stated on its own pages, checked November 2, 2026

| | MovieFlow | SceneMixer |

| Built around | A video agent and several models: “Your idea. A movie. In minutes.” (Beta)1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Models | MovieFlow-2.0, Seedance 2.5, MiniMax H3, Kling 3.0, Vidu and Nano Banana Pro on its plans2 | Tiers by resolution and price per second; the model behind each tier is managed for you |

| Plans | Standard $15.83/month billed yearly ($19 monthly) with 1,800 credits; Pro $32.50 ($39) with 3,880; Ultra $124.92 ($142) with 14,300; Starter $5.75 billed yearly with 1,0002 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Video price | Its own model: USD 0.008 a second; one 5-second video $0.04, as advertised on its home page1 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| Free | New users: 5 free 5-second videos, no card1 | 50 free credits and 15 free reference images for new accounts |

| Credit validity | Membership credits valid 31 days; purchased credits valid one year2 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

| Output | 1080P; watermark-free downloads after buying any plan2 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Last Envelope”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with MovieFlow if…

- You want very short clips at a very low price per second — its home page advertises USD 0.008 a second on its own model1.

- You want several video models in one place — its plans list five video and image models2.

- You want to try it first — new users get 5 free 5-second videos1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where MovieFlow lands in SceneMixer

| In MovieFlow | Closest step in SceneMixer |

| Video Agent (one line to a video) | AI Create drafts a script from one sentence; the script read returns the cast and an episode outline. Upload a script and read it |

| Choosing a model | A video tier chosen per segment, by resolution and price per second |

| Credit packs | Credit packs from $1.49; plan credits and bought credits sit in one balance. Do credits expire? |

## Moving a story over

- Bring the story — a script, a document, or one sentence for AI Create.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a MovieFlow alternative?

For a drama in episodes with a fixed cast and spoken dialogue, yes. For quick short clips from a line of text, MovieFlow is built around that.

### Is MovieFlow the same as Google Flow?

No. MovieFlow is movieflow.ai; Google Flow is Google's studio at flow.google.com. See Google Flow for that one.

### How much does MovieFlow cost?

On November 2, 2026, its pricing page listed Starter at $5.75 a month billed yearly (1,000 credits), Standard at $15.83 (1,800), Pro at $32.50 (3,880) and Ultra at $124.92 (14,300), and its home page advertised USD 0.008 a second on its own model. Check movieflow.ai/pricing for current prices.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15. Native dialogue vs dubbing.

## Sources and corrections

- MovieFlow (movieflow.ai) — what it is, its own model's per-second price, free videos for new users
- AI Movie Pricing | Plans | MovieFlow (movieflow.ai/pricing) — plans, credits, models, watermark and credit validity

Facts about MovieFlow come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check MovieFlow's site before you buy. MovieFlow is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Neolemon Alternative for Animated Story Video | SceneMixer

URL: https://scenemixer.com/alternatives/neolemon/
Neolemon makes consistent cartoon characters and print-ready picture books; SceneMixer turns a story into animated episodes that speak. When to pick which.

Alternatives

# A Neolemon Alternative for Animated Story Video

Updated November 2, 2026

Quick answer: Neolemon is built for illustrated children's books: it makes cartoon characters that stay consistent across scenes, multi-character scenes, speech bubbles and covers, and exports a print-ready PDF for Amazon KDP or a flipbook. SceneMixer makes the story move: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters speaking their lines on screen in 15 languages. Choose Neolemon for still illustrated pages and a printed book; choose SceneMixer for animated episodes.

## Pictures on a page, or a story that plays

Neolemon's home page walks through four steps — illustrate your character, create consistent variations, compose your scenes, assemble and publish2 — and sells one plan, with 600 credits a month1. SceneMixer starts from the same kind of story but produces video: the character sheet is the start, not the end.

## At a glance

Neolemon facts as stated on its own pages, checked November 2, 2026

| | Neolemon | SceneMixer |

| Built around | Illustrated children's books: consistent cartoon characters, a multi-character scene creator, speech bubbles, covers and print-ready export2 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Output | Book pages as print-ready PDFs (300 DPI, trim and bleed) for Amazon KDP, or a flipbook; video is not described on the pages we read2 | One video file per episode, with the lines spoken on screen |

| Consistency | Change pose, outfit, expression or background with editors while the character stays consistent; photo-to-cartoon avatars12 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Plans | Creator plan $29 a month, billed monthly: 600 credits (up to 150 generations), commercial use included, credits do not roll over1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Free | Sign in free with 20 credits for image generation; 5 illustrations free, no card1 | 50 free credits and 15 free reference images for new accounts |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rocket Pup”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Neolemon if…

- You are making a printed picture book — print-ready PDFs with trim and bleed, KDP upload and a flipbook2.

- You need still illustrations, not video — characters, scenes and covers for each page2.

- You want cartoon versions of people or pets — its photo-to-cartoon avatar creator is listed on the plan1.

- You want one simple plan with commercial use — one price, monthly billing only1.

## SceneMixer fits if…

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

- You want the story to move and speak. A children's story can become animated episodes with lines spoken on screen. AI kids' story video generator · Make an AI animated short film

## Where Neolemon lands in SceneMixer

| In Neolemon | Closest step in SceneMixer |

| Illustrate your character | A character sheet in the style you pick, drawn once. Regenerate or replace one sheet |

| Editors for consistency | Change one detail of a sheet with words. AI edit a sheet: outfit, hair, expression |

| Compose your scenes | A storyboard shot: where it happens, who is there and what they say. How to write one shot |

| Speech bubbles | Spoken lines performed in the video, with optional subtitles. Add subtitles |

## Moving a story over

- Bring the story — paste the text of your book or one sentence for AI Create; character art you made can be uploaded as a reference on a paid plan.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Neolemon alternative?

For turning a children's story into animated episodes, yes. For still illustrated pages and a printed book, Neolemon is built around that.

### How much does Neolemon cost?

On November 2, 2026, its pricing page listed one Creator plan at $29 a month, billed monthly, with 600 credits (up to 150 generations) and commercial use included. Check neolemon.com/pricing for current prices.

### Is Neolemon free?

Its pricing page, checked November 2, 2026, said you can sign in free with no card and get 20 credits for image generation, and that you can create 5 illustrations for free.

### Does SceneMixer make picture books?

No. It makes video episodes, not printed pages.

## Sources and corrections

- Pricing – Neolemon (neolemon.com/pricing) — plan, credits, free allowance, commercial use
- Neolemon (neolemon.com) — what the tool makes and how

Facts about Neolemon come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Neolemon's site before you buy. Neolemon is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# OpenArt Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/openart/
OpenArt is an image and video studio with Director and consistent characters; SceneMixer turns a script into episodes. How they compare, and when OpenArt fits.

Alternatives

# An OpenArt Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: OpenArt is an AI creator studio built around images and video: access to 100+ premium image, video and audio models, consistent characters and Director on every plan. SceneMixer does one thing end to end: it turns a script or novel into a series, with an episode outline, a fixed cast, a storyboard per episode and dialogue performed on screen in 15 languages. Choose OpenArt if images and quick stories are most of your work; choose SceneMixer for a drama in episodes.

## A creator studio, or a series pipeline

OpenArt's plans are measured in credits and consistent characters — Starter lists 4,000 credits a month and up to ~13 consistent characters, Plus 12,000 credits and ~40 characters1. SceneMixer is measured in a project: one script becomes one show, and every sheet, storyboard and segment belongs to it.

## At a glance

OpenArt facts as stated on its own pages, checked November 2, 2026

| | OpenArt | SceneMixer |

| Built around | An AI creator studio for images and video, with 100+ image, video and audio models1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Stories | Director on every plan1 | A full episode outline from the script, then a storyboard per episode you can edit |

| Keeping a character | Consistent characters, up to ~13 on Starter and ~40 on Plus1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Images | Access to 100+ premium image, video and audio models1 | Character sheets, location plates and props generated for the project; replace any one |

| Licence | Commercial use rights listed from Plus1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

| Price to start | Starter $14 per seat a month billed monthly ($13 billed annually), 4,000 credits1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rust City Knockout”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with OpenArt if…

- Most of your work is images. OpenArt pairs 100+ models with an image editing suite and personalised models1.

- You want quick one-click stories rather than a full episode workflow1.

- You want to choose among many named models for each generation1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where OpenArt lands in SceneMixer

| In OpenArt | Closest step in SceneMixer |

| Consistent characters | Character sheets generated once and attached to every segment. Regenerate or replace one sheet |

| Director | AI Create from one sentence, then the episode outline and storyboards |

| Image editing suite | Regenerate a sheet with a new description, or upload your own image on a paid plan |

## Moving a story over

- Bring the story. Paste a script, upload a document, or describe the premise in one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer an OpenArt alternative?

For multi-episode drama with dialogue, yes. For image-heavy work, personalised models or quick one-click stories, OpenArt is built around those.

### How much does OpenArt cost?

On November 2, 2026, its pricing page listed Starter at $14 per seat a month billed monthly ($13 billed annually) with 4,000 credits, Plus at $34 ($27), Pro at $56 ($44) and Wonder at $240 ($175). Check openart.ai/pricing for current prices.

### Do OpenArt credits roll over?

Its pricing page has a FAQ entry on whether monthly credits roll over; check it for the current answer. In SceneMixer, plan credits and bought credits sit in one balance and do not expire at the end of the month.

### Can I keep a character consistent in SceneMixer?

Yes. Each character gets a reference sheet once, and every segment that features them is rendered against it. How consistency works.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15.

## Sources and corrections

- OpenArt Pricing – AI Creator Studio for Video & Images (openart.ai/pricing) — plans, credits, Director, consistent characters and credit rules

Facts about OpenArt come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check OpenArt's site before you buy. OpenArt is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Pika Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/pika/
Pika gives you many models for quick clips; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when Pika fits better.

Alternatives

# A Pika Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Pika gives you a set of image, video and audio models on credit plans and packs, with parallel generations for fast iteration on clips. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with lines performed on screen in 15 languages. Choose Pika for quick clips across models; choose SceneMixer for a series with a story.

## Clips, or a story in episodes

Pika's plans are sized by credits and parallel generations, and its packs sell credits at a flat rate1. SceneMixer is sized by the project: the script decides the episodes and shots, and every shot is rendered against the same character sheets.

## At a glance

Pika facts as stated on its own pages, checked November 2, 2026

| | Pika | SceneMixer |

| Built around | All models on every plan; parallel video, image and audio generations1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Licence | Commercial licence from Starter ($10/month); not included on Free1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

| Price to start | Free: credit packs only; Starter $10/month with 900 credits; Creator $35/month with 3,1501 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Top-ups | Packs from $25 for 1,500 credits (“60 credits per $1 on every plan”)1 | Credit packs from $1.49 |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Starfall Command”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Pika if…

- You iterate on short clips and want many generations running in parallel1.

- You want to try many models from one account1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Pika lands in SceneMixer

| In Pika | Closest step in SceneMixer |

| Clips | Segments generated from a storyboard, joined into an episode |

| Credit packs | Credit packs from $1.49; see pricing |

## Moving a story over

- Bring the story — a script, a document, or one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Pika alternative?

For producing a drama in episodes from a script, yes. For quick clips across many models, Pika is built around that.

### How much does Pika cost?

On November 2, 2026, its pricing page listed Free (credit packs only), Starter at $10 a month (900 credits), Creator at $35 (3,150) and Fancy from $95, with packs from $25 for 1,500 credits. Check pika.art/pricing for current prices.

### Can I use Pika for free commercially?

Its pricing page lists a commercial licence from Creator, not on Free or Starter. SceneMixer's terms give you ownership, including commercial use, on free allowances too.

### Does SceneMixer keep characters consistent?

Yes. Each character gets a reference sheet once, and every segment that features them is rendered against it. How consistency works.

## Sources and corrections

- Pricing — Pika (pika.art/pricing) — plans, credits, credit packs, licences and model costs

Facts about Pika come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Pika's site before you buy. Pika is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Pippit Alternative for AI Short Drama | SceneMixer

URL: https://scenemixer.com/alternatives/pippit/
Pippit's drama tool makes videos from a script or prompt with a choice of models; SceneMixer builds an episodic drama from one script. When to pick which.

Alternatives

# A Pippit Alternative for AI Short Drama

Updated November 1, 2026

Quick answer: Pippit is a broad AI video suite; its AI Drama Generator turns a script or a prompt into drama videos, with a choice of video models, adjustable aspect ratio, length and language, and a Story Studio that organises a script into scenes or episodes. SceneMixer does one thing: from a script it produces a cast drawn once, an episode outline, a storyboard per episode and video segment by segment, with the lines performed on screen in 15 languages. Choose Pippit for a toolbox that also makes avatars, talking photos and product videos; choose SceneMixer for a series whose faces, rooms and voices must hold across episodes.

## A suite with a drama tool, or a drama pipeline

Pippit describes its drama tool as a way to “turn your scripts into powerful, viral drama videos”, with a prompt box, uploaded images or clips, a model picker, and a Story Studio that can organise a script “into scenes or episodes”1. In SceneMixer the episode structure, the cast and the storyboard are the product, not a setting.

## At a glance

Pippit facts as stated on its own pages, checked November 1, 2026

| | Pippit | SceneMixer |

| Built around | A video suite: AI Drama Generator, Story Studio, avatars, talking photos, product videos1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Drama inputs | A script, a text prompt, uploaded images or clips; a model picker and settings for aspect ratio, length and language1 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Characters across shots | Guidance to keep character references and scene details consistent; generated characters can be reviewed and regenerated1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Length | “A short drama can range from several minutes to about one hour”, adjustable per generation1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Plans | Free plan with daily credits and no card; paid plans list 1,575, 3,885 and 35,500 credits a month and “Remove watermark”; prices shown in local currency2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Video price | Not stated as a per-second figure (credits per generation vary by model)2 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Çantasındaki Sözleşme”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Pippit if…

- You want one subscription for many video jobs — avatars, talking photos, product videos and drama in one place2.

- You like choosing the video model yourself per generation, with a prompt box and uploads1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Pippit lands in SceneMixer

| In Pippit | Closest step in SceneMixer |

| Story Studio (scenes or episodes) | One read of the script returns the episode outline, cast, locations and props |

| Prompt box + model picker | The storyboard: numbered shots with seconds and lines, rendered on the tier you choose |

| Review and regenerate characters | Character sheets, drawn once and attached to every segment. Regenerate or replace one sheet |

## Moving a story over

- Bring the script — paste it, upload .txt / .docx / .md, or write one sentence in AI Create.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Pippit alternative?

For a drama in episodes with a fixed cast and performed dialogue, yes. For a general video suite with avatars and product tools, Pippit covers more ground.

### How much does Pippit cost?

On November 1, 2026 its pricing page listed a free plan with daily credits and paid plans with 1,575, 3,885 and 35,500 monthly credits, priced in local currency; check pippit.ai/pricing for current figures.

### Does SceneMixer let me pick the video model?

You pick a tier by resolution and price per second; the model behind each tier is managed for you.

## Sources and corrections

- AI Drama Generator – Pippit (pippit.ai/create/short-dramas) — what the drama tool takes and returns, steps, length
- Pricing – Pippit (pippit.ai/pricing) — plans, credits, watermark removal

Facts about Pippit come only from the pages listed above, as checked on November 1, 2026. Prices and plans change often, so check Pippit's site before you buy. Pippit is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# PixVerse Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/pixverse/
PixVerse offers its own models plus many others, with daily free credits; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare.

Alternatives

# A PixVerse Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: PixVerse is a generation app with its own models and access to many others, daily free credits, off-peak discounts and batch creation. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with lines performed on screen in 15 languages. Choose PixVerse for quick clips and daily free generations; choose SceneMixer for a series.

## Daily clips, or a series in episodes

PixVerse's plans combine daily credits with a monthly pool that resets every 30 days, and raise resolution, concurrency and model access as you go up1. SceneMixer's balance does not reset at the end of the month, and every generation belongs to a project.

## At a glance

PixVerse facts as stated on its own pages, checked November 2, 2026

| | PixVerse | SceneMixer |

| Built around | PixVerse models plus many other image and video models; batch creation and off-peak modes1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Daily credits | 60 a day for subscribers, reset daily, on top of the monthly pool1 | 50 free credits and 15 free reference images for new accounts |

| Price to start | Standard $8/month on the yearly plan ($96), 1,200 credits every 30 days, up to 720P1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Unused credits | The monthly pool resets every 30 days1 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Veil at Dawn”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with PixVerse if…

- You want daily free credits to experiment every day1.

- You batch many clips and use off-peak discounts1.

- You want many models in one app1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where PixVerse lands in SceneMixer

| In PixVerse | Closest step in SceneMixer |

| A clip | A storyboard segment, joined into an episode |

| Batch creation | Generate every segment of an episode, then regenerate only the ones that miss |

## Moving a story over

- Bring the story — a script, a document, or one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a PixVerse alternative?

For a drama series with a fixed cast and spoken dialogue, yes. For daily free clips across many models, PixVerse is built around that.

### How much does PixVerse cost?

On November 2, 2026, its subscription page listed Standard at $8 a month on the yearly plan ($10 monthly; 1,200 credits every 30 days), Pro at $24 ($30; 6,000), Premium at $48 ($60; 15,000) and Ultra at $149 ($249; 25,000), each with 60 daily credits; a free tier is not shown on that page. Check PixVerse's site for current prices.

### Do SceneMixer credits reset?

No. Plan credits and bought credits sit in one balance and do not expire at the end of the month.

## Sources and corrections

- PixVerse Pricing: Credits, Plans & AI Video Costs (app.pixverse.ai/subscribe) — plans, credits, resolutions and models

Facts about PixVerse come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check PixVerse's site before you buy. PixVerse is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Pollo AI Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/pollo/
Pollo AI gives you its own and third-party models for quick image and video generations; SceneMixer turns a script into an episodic drama. How they compare.

Alternatives

# A Pollo AI Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Pollo AI is a generation platform with its own Pollo models alongside third-party ones, sized by credits and parallel tasks for quick images and videos. SceneMixer is built for a story: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with lines performed on screen in 15 languages. Choose Pollo for fast, varied generations; choose SceneMixer for a drama series.

## Many quick generations, or one connected show

Pollo's plans list credits a month and how many videos or images that buys on its models, with more parallel tasks on higher plans1. SceneMixer's unit is the project: every sheet, storyboard and segment belongs to one story.

## At a glance

Pollo AI facts as stated on its own pages, checked November 2, 2026

| | Pollo AI | SceneMixer |

| Built around | Its own Pollo models plus third-party video and image models1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Plans | Free, Lite (400 credits a month), Pro (800) and Ultra (5,000), with 1–6 parallel tasks1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Prices | Shown in the visitor's local currency on its pricing page1 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Eagle's Vow”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Pollo AI if…

- You want quick generations across many models, including Pollo's own1.

- You run several tasks in parallel on higher plans1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Pollo AI lands in SceneMixer

| In Pollo AI | Closest step in SceneMixer |

| A generation | A storyboard segment, rendered against the project's character sheets |

| Parallel tasks | Generate all segments of an episode, then regenerate only the ones that miss |

## Moving a story over

- Bring the story — a script, a document, or one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Pollo AI alternative?

For drama in episodes with a fixed cast, yes. For quick generations across many models, Pollo is built around that.

### How much does Pollo AI cost?

Its pricing page lists Free, Lite, Pro and Ultra plans with 400, 800 and 5,000 monthly credits on the paid ones, priced in the visitor's local currency. Check pollo.ai/pricing for the price where you are.

### How much does SceneMixer video cost?

From about $0.06 per second at 480P up to $0.47 on the most expensive tier, at the lowest credit price.

## Sources and corrections

- Plans & Pricing | Pollo AI (pollo.ai/pricing) — plans, credits and models; prices are shown in the visitor's local currency

Facts about Pollo AI come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Pollo AI's site before you buy. Pollo AI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Revid AI Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/revid/
Revid AI creates, publishes and remixes short-form videos for growth; SceneMixer turns a script into an episodic drama. How they compare, and when Revid fits.

Alternatives

# A Revid AI Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Revid AI is built around a growth loop for short-form creators: turn text or links into videos, remix from millions of viral videos, publish to TikTok, Instagram and YouTube, and run Auto-Mode workers. SceneMixer is built around a story: a script becomes an episode outline, a fixed cast, a storyboard per episode and video segment by segment, with characters performing lines on screen in 15 languages. Choose Revid to grow a channel with volume; choose SceneMixer to produce a drama series.

## A channel engine, or a series pipeline

Revid's plans bundle rendering, publishing, post stats and automation — Auto-Mode workers, API, MCP and CLI access1. SceneMixer does not publish for you; it focuses on making the episodes themselves, with every shot rendered against the same character sheets.

## At a glance

Revid AI facts as stated on its own pages, checked November 2, 2026

| | Revid AI | SceneMixer |

| Built around | Create, publish, learn, repeat: rendering, publishing, performance stats and Auto-Mode1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Input | Text or links; 3M+ viral videos to remix1 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Voice | 70+ language voiceovers; create and clone your own voices on Ultra1 | Lines are performed in the generated video itself, in any of 15 languages |

| Publishing | Publish to TikTok, Instagram, YouTube and X1 | Download the composited episode and publish it yourself |

| Price to start | Growth $39/month ($32 billed yearly) with 2,000 AI credits; Ultra $199/month with 12,0001 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Neon Debt”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Revid AI if…

- You run faceless or short-form channels at volume. Revid includes publishing, post stats and Auto-Mode workers1.

- You remix trends. It offers millions of viral videos to remix1.

- You need avatars, face swaps or podcasts alongside your videos1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Revid AI lands in SceneMixer

| In Revid AI | Closest step in SceneMixer |

| Idea or link to video | AI Create from one sentence, or a full script read into episodes |

| Voiceovers | Lines performed by the characters on screen; voice-over lines where you mark them. Lines, voice-over and on-screen text |

| Remix | Video Remake reads a clip you upload and rebuilds it with your own cast. Remake a video with AI |

## Moving a story over

- Bring the story — a script, a document, or one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Revid AI alternative?

For drama series with a fixed cast and spoken dialogue, yes. For publishing, trend remixing and automated short-form channels, Revid is built around those.

### How much does Revid AI cost?

On November 2, 2026, its pricing page said Growth is $39 a month or $32 billed yearly with 2,000 AI credits, and Ultra is $199 a month or $166 billed yearly with 12,000. Check revid.ai/pricing for current prices.

### Does SceneMixer post to TikTok for me?

No. You download each composited episode and publish it wherever you want.

### Can SceneMixer remake a viral clip?

Video Remake reads a clip you upload — its shots, people and lines — and rebuilds it with characters from your own sheets. Remake a video with AI.

## Sources and corrections

- Revid AI Pricing: Plans, Credits & Features (revid.ai/pricing) — plans, credits and what each plan includes

Facts about Revid AI come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Revid AI's site before you buy. Revid AI is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Runway Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/runway/
Runway is a generation and editing platform; SceneMixer turns a script into an episodic drama with a fixed cast. How they compare, and when Runway fits better.

Alternatives

# A Runway Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Runway is a generation and editing platform: its own Gen-4.5 model alongside other video and image models, an agentic creative collaborator, music, voice and sound-effect generation, 4K upscaling and studio formats on its top plan. SceneMixer is a production line for drama in episodes: from a script it produces the outline, the cast, a storyboard per episode and the video, with lines performed on screen in 15 languages. Choose Runway for shot-level craft and editing; choose SceneMixer to turn a written story into a series.

## Shot-level craft, or a whole show

Runway's plans are sized in seconds of its models — Standard's 625 monthly credits equal 52 seconds of Gen-4.5 — and add projects, brand kits, custom voice clones and studio-grade formats as you go up1. SceneMixer is sized in projects: one script, one cast, every episode rendered and composited in the same place.

## At a glance

Runway facts as stated on its own pages, checked November 2, 2026

| | Runway | SceneMixer |

| Built around | Its Gen-4.5 model plus other video and image models, an agentic creative collaborator and editing1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Audio | Music, voice and sound-effect generation; custom voice clones from Pro1 | Lines are performed in the generated video itself, in any of 15 languages |

| Finishing | 4K upscaling from Standard (unlimited on Pro and Max); HDR, ProRes and image sequences on Max1 | One composited file per episode; resolution follows the tier |

| Price to start | Free: 125 one-time credits; Standard $15/month ($12 billed annually), 625 credits = 52s of Gen-4.51 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Unused credits | Standard and Pro monthly credits don't roll over; Max rolls over up to one month; purchased credits never expire1 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Glass Floor”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Runway if…

- You are editing or finishing shots. Runway pairs generation with editing tools, 4K upscaling and studio formats1.

- You want Runway's own model, Gen-4.51.

- You need music and sound effects generated alongside the picture1.

- You work under brand kits across projects1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

## Where Runway lands in SceneMixer

| In Runway | Closest step in SceneMixer |

| Agentic creative collaborator | AI Create and the script read, which plan the whole series |

| Voice generation and custom voices | A voice matched to each character, performing lines in the project's language. Native dialogue vs dubbing |

| Generations and upscales | Segments rendered per tier; regenerate a segment at a higher resolution when it matters |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Runway alternative?

For producing a multi-episode drama from a script, yes. For shot-level editing, Runway's own model, music and sound effects or studio formats, Runway has more.

### How much does Runway cost?

On November 2, 2026, Runway's pricing page listed Free with 125 one-time credits, Standard at $15 a month or $12 billed annually (625 credits), Pro at $35 or $28 (2,250) and Max at $95 or $76 (9,500). Check its pricing page for current prices.

### Do Runway credits roll over?

Its pricing FAQ says Standard and Pro monthly credits don't roll over, Max rolls over up to one month, and purchased credits never expire. In SceneMixer, plan credits and bought credits do not expire at the end of the month.

### Does SceneMixer generate sound?

The dialogue is performed in the generated video, in the project's language, with the ambient sound you write into each shot.

## Sources and corrections

- Choose the Right Plan for You (runway.com/pricing) — plans, credits, credit rules, models and features

Facts about Runway come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Runway's site before you buy. Runway is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Showrunner Alternative for Making AI Shows | SceneMixer

URL: https://scenemixer.com/alternatives/showrunner/
Showrunner lets you watch AI shows and make scenes inside them; SceneMixer builds your own series from a script with a fixed cast. When to pick which.

Alternatives

# A Showrunner Alternative for Making AI Shows

Updated November 1, 2026

Quick answer: Showrunner, by Fable Studio, is a place to “watch AI shows” and “make your own”: you can create a show and build its world, create a show from a YouTube link or a video, direct a scene, upload yourself to star in a show, combine scenes and remix other creators' shows, in genres from comedy to horror. SceneMixer starts from your script: it reads it into a cast, locations and episodes, draws the cast once, storyboards each episode and renders it segment by segment with the dialogue performed. Choose Showrunner for a community of shows to watch, remix and appear in; choose SceneMixer for a series that is yours from the first line, downloaded as files.

## A platform of shows, or your own series

Showrunner's homepage is a feed of shows with creator credits and “Remix” buttons, alongside “Create Show”, “Create Scene” and “Upload yourself”1. SceneMixer has no public feed of your work: projects are private, and the output is a video file per episode.

## At a glance

Showrunner facts as stated on its own pages, checked November 1, 2026

| | Showrunner | SceneMixer |

| Built around | Watching AI shows and making scenes within them; remixing; starring in a show by uploading yourself1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| How you start | Create Show (build a story world), from a YouTube link, from a video; Create Scene; Remix an existing show1 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Where it lives | A web app; a Discord community1 | A web app and apps; projects stay private to your account |

| Pricing | Not stated on the homepage (Sign in / Create Show)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Ownership and commercial use | Not stated on the homepage1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

| Episode length | Not stated1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rocket Pup”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Showrunner if…

- You want to watch, remix and be part of other people's shows — Showrunner is a community as much as a tool1.

- You want to star in a show yourself — “Upload yourself” is a feature there1; SceneMixer does not generate real people's faces.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

- You need landscape as well as vertical. 16:9 is the default and 9:16 is one setting away. Vertical and landscape projects.

## Where Showrunner lands in SceneMixer

| In Showrunner | Closest step in SceneMixer |

| Create Show / build your story world | One read of the script returns cast, locations, props and an episode outline. Upload a script and read it |

| Create Scene / direct the shot | A storyboard segment: numbered shots with seconds, framing and lines, edited as text. How to write one shot |

| Remix a show | Reuse your own characters, scenes and voices across your projects |

## Moving a story over

- Write the show down — a premise in one sentence, a script, or a chapter.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Showrunner alternative?

For making your own series from a script and downloading the episodes, yes. For watching and remixing a community of AI shows, Showrunner is built around that.

### How much does Showrunner cost?

Its homepage did not state pricing on November 1, 2026; it shows Sign in and Create Show buttons. Check showrunnerstudio.com.

### Can I appear in my own SceneMixer drama?

No. Real people's faces are not generated; characters are drawn from descriptions or from reference art you own.

## Sources and corrections

- Showrunner (showrunnerstudio.com) — what the product is, how shows and scenes are created, who is behind it

Facts about Showrunner come only from the pages listed above, as checked on November 1, 2026. Prices and plans change often, so check Showrunner's site before you buy. Showrunner is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Sora Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/sora/
OpenAI discontinued the Sora app in April 2026 and the Sora API in September 2026. If you made stories with Sora, here is how SceneMixer picks up the work.

Alternatives

# A Sora Alternative for AI Drama Series, After the Shutdown

Updated November 2, 2026

Quick answer: OpenAI discontinued the Sora web and app experiences on April 26, 2026 and the Sora API on September 24, 2026. If you used Sora to tell stories, SceneMixer covers that work as a pipeline: from a script, a novel or one sentence it produces the cast, an episode outline, a storyboard per episode and the video segment by segment, with characters performing their lines on screen in 15 languages. Clips you exported from Sora can also be read by Video Remake and rebuilt with your own cast.

## What happened to Sora

OpenAI's help center says the Sora web and app experiences were discontinued on April 26, 2026, the API on September 24, 2026, and that data is permanently deleted after any final export window1. There is no Sora to switch back to, so the question is what to use for the same work.

## At a glance

Sora facts as stated on its own pages, checked November 2, 2026

| | Sora | SceneMixer |

| Status | App discontinued April 26, 2026; API discontinued September 24, 20261 | Available on the web; new projects start from a script, a novel or one sentence |

| Your old content | Export was offered at sora.chatgpt.com/sunset; data is deleted after any final export window1 | Upload an exported clip of up to 300 seconds to Video Remake: it reads the shots, cast and lines and rebuilds them with your characters |

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “The Last Envelope”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Sora lands in SceneMixer

| In Sora | Closest step in SceneMixer |

| A Sora prompt | A storyboard shot: who is there, what they do, the camera and the line. How to write one shot |

| A clip you exported | Video Remake reads it and rebuilds it with your own cast. Remake a video with AI |

| A recurring character | A character reference sheet used in every segment. Regenerate or replace one sheet |

## Moving a story over

- Start from the story, not the clip. Paste or upload the script, or describe the premise in one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### When did Sora shut down?

OpenAI's help center says the Sora web and app experiences were discontinued on April 26, 2026 and the Sora API on September 24, 2026.

### Can I still get my Sora videos?

OpenAI said content could be exported from sora.chatgpt.com/sunset and that data is permanently deleted after any final export window. Check OpenAI's help center for what is still possible.

### Can SceneMixer reuse my exported Sora clips?

Video Remake can read a clip of up to 300 seconds — its cuts, people, actions and lines — and rebuild it with characters from your own sheets. The clip itself is never handed to the video model. Remake a video with AI.

### Is SceneMixer a Sora alternative for stories?

For stories told in episodes with a fixed cast and spoken dialogue, yes: that is the whole pipeline SceneMixer runs. If you mainly made single clips, a general video generator may suit you better; see the other pages in this section.

### How much does SceneMixer cost?

Credit packs start at $1.49, the Pro plan is $7.99 a month, and video starts at about $0.06 a second at 480P at the lowest credit price. New accounts get 50 free credits.

## Sources and corrections

- What to know about the Sora discontinuation (OpenAI Help Center) — shutdown dates, export and data deletion

Facts about Sora come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Sora's site before you buy. Sora is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# StoryShort Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/storyshort/
StoryShort automates faceless short videos and auto-posts them; SceneMixer turns a script into an episodic drama. How they compare, and when to stay.

Alternatives

# A StoryShort Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: StoryShort automates short videos: pick a series, and it generates voiced videos with AI images or video, background music and captions, and can auto-post them. SceneMixer makes drama: from a script it produces an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose StoryShort for automated faceless channels; choose SceneMixer for a story with recurring characters.

## Automated shorts, or a drama with a cast

StoryShort sizes its plans in videos — Starter lists about 26 videos of 30 seconds a month — and in series, its word for an automated channel feed1. In SceneMixer a series means a drama: episodes from one outline, with the same characters throughout.

## At a glance

StoryShort facts as stated on its own pages, checked November 2, 2026

| | StoryShort | SceneMixer |

| Built around | Automated short videos with voiceovers, background music and auto-posting1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| What “series” means | An automated feed of videos (1 series on Starter, 4 on Ultra)1 | A drama in episodes, written from one outline |

| Per-video cost | About 15 credits for a 30-second video with AI images; about 110 for an “AI movie”1 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| Price to start | Starter $39/month ($23/month billed yearly) with 400 credits1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rust City Knockout”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with StoryShort if…

- You run a faceless channel and want videos generated and posted on a schedule1.

- You make many short formats such as ASMR, UGC ads or talking-avatar videos1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where StoryShort lands in SceneMixer

| In StoryShort | Closest step in SceneMixer |

| Series (automated feed) | A project with episodes from one outline |

| Voiceovers | Characters performing their lines, plus voice-over where you mark it. Lines, voice-over and on-screen text |

| AI movie | An episode composited from storyboard segments |

## Moving a story over

- Write the story down — a script, a premise in one sentence, or a novel chapter.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a StoryShort alternative?

For a drama with recurring characters and dialogue, yes. For automated faceless shorts that post themselves, StoryShort is built around that.

### How much does StoryShort cost?

On November 2, 2026, its pricing page listed Starter at $39 a month or $23 billed yearly (400 credits), Growth at $69 or $42, Influencer at $129 or $75 and Ultra at $199 or $117. Check storyshort.ai/pricing for current prices.

### Does SceneMixer auto-post?

No. You download the composited episode and publish it yourself.

## Sources and corrections

- Pricing – StoryShort AI (storyshort.ai/pricing) — plans, credits, per-video costs and features

Facts about StoryShort come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check StoryShort's site before you buy. StoryShort is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Topview Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/topview/
Topview Drama Studio and SceneMixer both turn scripts into episodes. How they differ on languages, credits and price, and when Topview is the better fit.

Alternatives

# A Topview Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Topview's Drama Studio and SceneMixer now do the same core job: both take a story idea, a script or a novel and plan characters, episodes and storyboards before generating video. The differences are in the details. In SceneMixer the cast speaks any of 15 languages on screen, a project can be relocated to another market, video can be landscape or vertical, and credits are pay-as-you-go from $1.49 with plan credits that do not expire at the end of the month. Topview is the closer fit if you also make product ads and avatar videos, or want the unlimited generation windows on its annual plans.

## The same idea, two different workshops

Topview describes Drama Studio as planning a drama with an agent, reviewing story assets in a workbench, creating storyboards and moving approved beats into episode production1. SceneMixer runs the same arc as five steps in one project — read the script, draw the reference art, write a storyboard per episode, render each segment, composite the episode — and every step can be edited before the next one starts. If you have used one, the other will feel familiar; the table below is where they part ways.

## At a glance

Topview facts as stated on its own pages, checked November 2, 2026

| | Topview | SceneMixer |

| What you can start from | “A one-line story idea, a short plot outline, a full drama script, a novel chapter, or a serialized story concept”1 | One sentence (AI Create), a pasted script, or an uploaded script or novel (.txt, .docx, .md) |

| Planning | An agent plans the drama; story assets are reviewed in a workbench; Single Episode and Series modes1 | One read returns cast, locations, props and an episode outline; Type is Short Drama, TV Series or Film |

| Output format | Recommends vertical 9:16 as “usually the best format” for short drama platforms1 | 16:9 by default, 9:16 available, set per project |

| Voice and on-screen text | An AI voiceover workflow and auto captions are listed for Topview; Drama Studio says available options depend on the production flow and models1 | Lines are performed in the video itself; no automatic subtitle track (short on-screen text can be written into a shot, platform captions are added after export) |

| Dialogue languages | Not stated on the Drama Studio page1 | 15 languages |

| Also on the same plan | AI avatars (5,000+ public avatars), product avatars, voice cloning (5 free clones on Plus, 30 on Business, 60 on Ultra), TikTok Ad Library browsing, API and MCP access2 | Ad video and video remake modes alongside drama projects |

| Price to start | Plus $36/month billed annually (100 credits a month); Business $44/month billed annually against a $75 list price (3,000 credits a year, delivered up front); Ultra $50/month billed annually against $150 ($599.9 a year; 500 credits a month); Team $56 a seat2 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Video price | Its own comparison table (dated August 7, 2026) lists $0.12 per second of 720P video on the Ultra Annual plan ($599.9 billed annually)2 | From about $0.06 per second at 480P, $0.12 at 720P and $0.24 at 1080P (lowest tier at each resolution, at the lowest credit price). |

| Unused credits | On a monthly plan “credits reset every month” and don't roll over; top-up credits are valid for 24 months2 | Plan credits and bought credits sit in one balance and do not expire at the end of the month. |

| Unlimited generation | Credit-free “Unlimited” windows on listed models: image models only on Plus, video models too from Business; on the website only, not through the API2 | None; every generation is priced in credits |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Crimson Vow”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Topview if…

- You also make product ads, avatar or UGC-style videos. Topview puts avatars, product avatars, voice cloning and TikTok Ad Library browsing on the same plan as Drama Studio2.

- You generate at volume and its unlimited windows on annual plans cover the models and resolutions you use2.

- You need API or MCP access, which its plans list2.

- You want captions generated for you as part of the drama workflow1.

## SceneMixer fits if…

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You need landscape as well as vertical. 16:9 is the default and 9:16 is one setting away. Vertical and landscape projects.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want to try it before paying. New accounts get 50 free credits and 15 free reference images.

## Where Topview lands in SceneMixer

| In Topview | Closest step in SceneMixer |

| Agent planning | AI Create (from one sentence) or the script read, which returns the cast, locations, props and an episode outline you edit. Upload a script and read it |

| Workbench story assets | Step 2: characters, scenes and props, each with a reference sheet. Regenerate or replace one sheet |

| Storyboards | A storyboard per episode, shot by shot, with editable text and versions. How to write one shot |

| Episode production | Video per segment; regenerate one segment without re-billing the episode; composite the episode into one file |

| Voiceover and captions | Dialogue performed in the video; captions added after export. Lines, voice-over and on-screen text |

## Moving a story over

- Copy the script or outline out of Drama Studio and paste or upload it.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Topview alternative?

For script-to-episode drama work, yes: both plan characters, episodes and storyboards before generating video. SceneMixer differs in on-screen dialogue in 15 languages, market relocation, landscape as well as vertical output, and pay-as-you-go credits. For avatar videos, product ads or unlimited generation on an annual plan, Topview covers more.

### How much does Topview cost?

On November 2, 2026, Topview's pricing page listed Plus at $36 a month billed annually (100 credits), Business at $44 billed annually (list price $75; 3,000 credits a year up front), Ultra at $50 billed annually (list price $150; $599.9 a year, 500 credits a month) and Team at $56 a seat, and its FAQ said Pro is no longer sold. Check topview.ai/pricing for current prices.

### Do Topview credits roll over?

Its pricing FAQ says that on a monthly plan credits reset every month and unused credits don't roll over, while top-up credits are valid for 24 months. In SceneMixer, plan credits and bought credits sit in one balance and do not expire at the end of the month.

### Which languages can SceneMixer characters speak?

Any of 15 languages. The language is set per project, and the video performs the lines in that language as it renders.

### Does SceneMixer have an unlimited plan?

No. Every generation is priced in credits, from a single segment upward, so you pay for what you generate and nothing expires at the end of the month.

### Does SceneMixer add captions?

Not as an automatic subtitle track. Short on-screen text can be written into a shot, and platform captions are added after you export the episode.

### Can I bring my Topview project over?

Bring the script or outline as text and, on a paid plan, upload character images you have the rights to in place of generated sheets. Episodes, storyboards and video are then generated in SceneMixer.

## Sources and corrections

- AI Drama Generator for Short & Micro Dramas (topview.ai/drama-studio) — Drama Studio inputs, workflow and output format
- Pricing (topview.ai/pricing) — plans, credits, credit rules, unlimited windows and the tools included on each plan

Facts about Topview come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Topview's site before you buy. Topview is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# VEED Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/veed/
VEED is a browser video editor with an AI studio of many models; SceneMixer turns a script into an episodic drama. How they compare, and when VEED fits.

Alternatives

# A VEED Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: VEED is a browser video editor with a generative studio: its AI video generator lets you switch between several models — it lists Seedance 2.5, Veo 3, Kling, Hailuo 2.3 and its own Fabric 1.0 for talking videos up to 60 seconds — then add voiceovers, captions, dubbing and your branding in the same editor. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose VEED to generate clips and finish them in an editor with captions, avatars and translation; choose SceneMixer to produce a series with a fixed cast.

## An editor with a model menu, or a series pipeline

VEED's plans are sized in credits — its Creator plan lists 6,000 credits (about 1,500 AI videos) and its Pro plan 30,000 — and each model uses a different number of credits12. SceneMixer's unit is the project: every sheet, storyboard and segment belongs to one story.

## At a glance

VEED facts as stated on its own pages, checked November 2, 2026

| | VEED | SceneMixer |

| Built around | A browser video editor with a generative studio: AI video, avatars, dubbing, subtitles and translation12 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| AI video models | Seedance 2.5, Veo 3, Kling, Hailuo 2.3 and Fabric 1.0 (talking videos up to 60 seconds), among others, chosen per shot2 | Tiers by resolution and price per second; the model behind each tier is managed for you |

| Inputs | Text, images, product photos, scripts, or an existing clip for video-to-video models2 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| After generating | Click “AI-generated Media” on a scene, add captions, voiceovers, music and brand assets, dub into other languages2 | The episode editor: one card per segment; trim, re-dub, AI-edit or enhance a clip; subtitles and a built-in score per episode |

| Plans | Creator 6,000 credits, Pro 30,000, Studio 180,000, a user a month billed yearly; shown in local currency (AU$17, AU$35 and AU$56 when we checked)1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Free | Free AI credits at sign-up; each model uses a different number of credits2 | 50 free credits and 15 free reference images for new accounts |

| Languages | Pro lists translation into 50+ languages1 | Lines are performed in the generated video itself, in any of 15 languages |

| Commercial use | “You retain rights to videos you create with VEED and can use them commercially”, subject to its terms2 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rocket Pup”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with VEED if…

- You edit video in the browser — captions, trimming, brand kits and AI generation in one editor12.

- You want to switch between video models shot by shot inside the same project2.

- You make talking-head or UGC-style videos — Fabric 1.0 and AI avatars are built for them2.

- You translate or dub existing videos — Pro lists 50+ languages and the AI video page lists dubbing12.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where VEED lands in SceneMixer

| In VEED | Closest step in SceneMixer |

| A scene from the AI studio | A storyboard segment, rendered on a video tier. How to write one shot |

| AI avatars and Fabric talking videos | A character performing lines in a scene. AI video with talking characters |

| Dubbing AI | Lip sync or re-perform on a finished clip. Lip sync or re-perform |

| AI-generated Media on a scene | Edit this clip, AI visuals: change one thing in a finished clip. What AI visuals can change |

| Auto subtitles | Subtitles burned in or exported as SRT. Add subtitles |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; prompts you wrote for single clips can become shots in the storyboard.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a VEED alternative?

For turning a written story into a multi-episode drama with a fixed cast, yes. For editing, captions, dubbing and avatar videos in the browser, VEED is built around that.

### How much does VEED cost?

On November 2, 2026, its pricing page listed Creator with 6,000 credits, Pro with 30,000 and Studio with 180,000 a user a month billed yearly, priced in local currency (AU$17, AU$35 and AU$56 for us). Check veed.io/pricing for your region.

### Is VEED's AI video generator free?

Its page, checked November 2, 2026, said you get free AI credits when you sign up, and each model uses a different number of credits.

### Is SceneMixer a video editor like VEED?

No. It generates the episode and lets you trim, re-dub, AI-edit and enhance each clip, but it is not a multi-track editor. The episode editor, a tour.

## Sources and corrections

- Pricing – VEED (veed.io/pricing) — plans, credits, translation languages, AI voice hours
- AI Video Generator – VEED (veed.io/tools/ai-video) — models, inputs, editing, free credits, rights

Facts about VEED come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check VEED's site before you buy. VEED is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Veo Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/veo/
Veo is Google's video model with native audio; SceneMixer turns a script into an episodic drama. How they compare, and when Veo is the better fit.

Alternatives

# A Veo Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Veo is Google DeepMind's video generation model: Veo 3.1 makes video with sound effects, ambient noise and dialogue generated natively, from text, images, up to three reference images or a first and last frame, in clips of 4 to 8 seconds that can be extended. Google's pages now say Gemini Omni replaces Veo in the Gemini app and the Veo 3.1 preview in the Gemini API shut down on October 22, 2026, while Flow still lists Veo 3.1. SceneMixer starts one level up: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Veo for single clips from Google's model; choose SceneMixer to produce a series from a written story.

## A model from Google, or a series pipeline

Google presents Veo 3.1 as “Video, meet audio”: a model with reference images, first-and-last-frame transitions, camera controls and scene extension1. Where you can reach it has been changing — the Gemini app page says Omni replaces Veo there3, and the API documentation says the Veo 3.1 preview shut down on October 22, 20262 — so check Google's pages for today's state. SceneMixer starts from the script instead of the clip: the outline decides the episodes, and every segment is rendered against the same character sheets.

## At a glance

Veo facts as stated on its own pages, checked November 2, 2026

| | Veo | SceneMixer |

| Built around | A video generation model, Veo 3.1: sound effects, ambient noise and dialogue generated natively with the picture1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Inputs | Text, images, up to three reference images, a first and a last frame, and an earlier Veo clip to extend12 | Upload or paste a script or novel (.txt, .docx, .md), or draft one from a single sentence with AI Create |

| Clip length | 4, 6 or 8 seconds per generation; extension adds 7 seconds, up to 20 times, in the API documentation2 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Resolution | 720p, 1080p or 4k, landscape 16:9 or portrait 9:16, at 24 fps (1080p and 4k at 8 seconds)2 | 480P, 720P and 1080P video tiers, chosen per segment, each with its own price per second |

| Where to use it | Gemini app: Omni replaces Veo. Gemini API: Veo 3.1 preview shut down October 22, 2026. Flow: Veo 3.1 listed234 | At scenemixer.com in a browser, with the same features in every region the site serves |

| Price | Not stated on Google's Veo pages. Flow, which lists Veo 3.1, is free with a Google account or part of Google AI Plus ($4.99/month), Pro ($19.99) and Ultra ($99.99 and $199.99); prices vary by market4 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Starfall Command”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Veo if…

- You want Google's own model with sound made alongside the picture — Veo 3.1 is built for single clips with generated audio1.

- You need fine camera control on one clip — Google lists camera moves such as move back, zoom in, move up and move right1.

- You work on existing footage — outpainting and adding or removing objects are listed among Veo's capabilities1.

- You build on Google's platform and follow where Google points Veo users next: Omni in the Gemini app, Flow for the studio34.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

- You want Google's models compared one to one. See Gemini Omni for the model Google now points Gemini users to, and Google Flow for the studio that lists Veo 3.1.

## Where Veo lands in SceneMixer

| In Veo | Closest step in SceneMixer |

| Reference images (ingredients) | Character, location and prop sheets attached to every segment. Regenerate or replace one sheet |

| Extend your scene | The next segment of the storyboard, which opens where the last one ended. Shot-to-shot continuity |

| Camera controls | Camera and framing written per shot in the storyboard. Write camera moves for AI video |

| Native audio and dialogue | Speech generated with the picture, in the project's language. Native dialogue vs dubbing |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; prompts you wrote for single clips can become shots in the storyboard.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Veo alternative?

For producing a drama series from a written story, yes. For single clips straight from Google's model, Veo is built for that.

### Is Veo still available?

On November 2, 2026, Google's pages said Gemini Omni replaces Veo in the Gemini app, the Veo 3.1 preview in the Gemini API shut down on October 22, 2026, and Flow still listed Veo 3.1. Google's pages change, so check them for the current state.

### How much does Veo cost?

Google's Veo pages list no price. On November 2, 2026, Flow's page listed free use with a Google account, and Google AI Plus at $4.99 a month, Google AI Pro at $19.99 and Google AI Ultra at $99.99 or $199.99, with prices varying by market.

### Can SceneMixer characters talk?

Yes. Lines are written into the storyboard and performed in the video in the project's language, one of 15. Native dialogue vs dubbing.

### Does SceneMixer keep the same character from clip to clip?

Yes. Each character is drawn once as a reference sheet and attached to every segment they appear in, across every episode. How consistency works.

## Sources and corrections

- Veo – Google DeepMind (deepmind.google/models/veo) — what Veo 3.1 does: native audio, reference images, first and last frame, extension, camera controls
- Generate videos with Veo 3.1 in Gemini API (ai.google.dev/gemini-api/docs/veo) — durations, resolutions, extension limits and the October 22, 2026 shutdown notice
- Gemini Omni (gemini.google/overview/video-generation) — what Google says about Veo in the Gemini app
- Google Flow (flow.google.com/about) — Veo 3.1 listed in Flow; plans and prices

Facts about Veo come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Veo's site before you buy. Veo is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Vidnoz Alternative for AI Story Video | SceneMixer

URL: https://scenemixer.com/alternatives/vidnoz/
Vidnoz makes avatar presenter videos from templates, with a free daily allowance; SceneMixer turns a script into an episodic drama. When to pick which.

Alternatives

# A Vidnoz Alternative for AI Story Video

Updated November 2, 2026

Quick answer: Vidnoz is an AI avatar video platform: you pick a template and an avatar, type what it should say, choose a voice, then edit and export — with voice cloning, video translation and e-learning export on higher plans, and a free allowance of 30 credits a day. SceneMixer is built around stories: a script becomes an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with several characters acting and speaking in 15 languages. Choose Vidnoz for presenter, training and sales videos from templates; choose SceneMixer for scenes and episodes with characters who act.

## A presenter in a template, or a cast in scenes

Vidnoz describes its workflow as four steps — choose a template and avatar, create the voiceover, add touches, generate2. Its Free plan lists 30 credits a day (about 60 seconds), videos up to 3 minutes and 720p export1. SceneMixer does not make a presenter-to-camera video; it makes a scene in which characters talk to each other.

## At a glance

Vidnoz facts as stated on its own pages, checked November 2, 2026

| | Vidnoz | SceneMixer |

| Built around | AI avatars, video templates and voiceover: “Scale Video Production 10X Faster”12 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Presenters and templates | 1,600+ AI avatars and 3,400+ video templates on Free; 1,700+ avatars on paid plans; photo avatars1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Free plan | 30 credits a day (about 60 seconds), videos up to 3 minutes, 720p export1 | 50 free credits and 15 free reference images for new accounts |

| Plans | Starter $26.99/month ($19.99 billed yearly), 450 credits a month, 1080p, no watermark; Business $74.99 ($56.99 billed yearly), 900 credits a month, voice clone and video translation1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Video length | Up to 60 minutes a video on paid plans1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| Voices and languages | Voices from 890 on Free to 2,680+ on paid plans; its home page says 140+ languages12 | Lines are performed in the generated video itself, in any of 15 languages |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Ikrar Jakarta”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Vidnoz if…

- You need a presenter speaking to camera — avatars with lip sync, from templates12.

- You make training or e-learning video — its Starter plan lists SCORM export1.

- You translate videos — Business lists video translation and voice cloning1.

- You want a free daily allowance to try it — 30 credits a day on the Free plan1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- The same faces must hold from shot to shot. Reference sheets ride along with every segment, linked by id rather than by name. How consistency works.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- Your story needs scenes, not a presenter. Characters act, move and speak to each other in locations drawn for the story. AI video with talking characters.

## Where Vidnoz lands in SceneMixer

| In Vidnoz | Closest step in SceneMixer |

| Avatar | A character sheet, drawn once; the character acts in the shot rather than speaking to camera. Regenerate or replace one sheet |

| Video template | The project type in step 1 — Short Drama, TV Series or Film — and a script read into an episode outline. Upload a script and read it |

| Voice cloning | A voice per character chosen from a library. Pick character voices |

| Video translation | The dialogue language is set once in step 1 and the lines are performed in it. Choose the dialogue language |

## Moving a story over

- Bring the script — paste it or upload .txt, .docx or .md; a presenter script can become dialogue lines in a scene.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Vidnoz alternative?

For stories told in episodes by several characters, yes. For avatar presenters, templates and training video, Vidnoz is built around those.

### How much does Vidnoz cost?

On November 2, 2026, its pricing page listed Free (30 credits a day), Starter at $26.99 a month ($19.99 billed yearly) and Business at $74.99 ($56.99 billed yearly). Check vidnoz.com/pricing.html for current prices.

### Is Vidnoz free?

Its pricing page, checked November 2, 2026, listed a Free plan with 30 credits a day (about 60 seconds), videos up to 3 minutes and 720p export.

### Does SceneMixer make avatar presenter videos?

No. It makes scenes in which characters act and speak to each other; there is no presenter-to-camera mode. Native dialogue vs dubbing.

## Sources and corrections

- Vidnoz AI Pricing (vidnoz.com/pricing.html) — plans, credits, free allowance, avatars, templates, video length
- Vidnoz AI (vidnoz.com) — what the tool does, steps, voices and languages

Facts about Vidnoz come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Vidnoz's site before you buy. Vidnoz is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.


---

# Vidu Alternative for AI Drama Series | SceneMixer

URL: https://scenemixer.com/alternatives/vidu/
Vidu makes 3–16 second clips with multi-reference consistency; SceneMixer turns a script into an episodic drama. How they compare, and when Vidu fits.

Alternatives

# A Vidu Alternative for AI Drama Series

Updated November 2, 2026

Quick answer: Vidu is a video model and app: clips of 3 to 16 seconds, multi-reference consistency, smooth 2D animation, batch generation and an off-peak mode on credit plans. SceneMixer builds a whole drama from a script: an episode outline, a cast drawn once, a storyboard per episode and video segment by segment, with characters performing their lines on screen in 15 languages. Choose Vidu for quick, consistent clips; choose SceneMixer for a series.

## Clips with references, or a series with a cast

Vidu's plans are sized in credits and references, with fast channels, batches of four and off-peak quotas on higher tiers1. SceneMixer keeps references inside the project: every character, location and prop has one sheet, and every segment that needs it uses it.

## At a glance

Vidu facts as stated on its own pages, checked November 2, 2026

| | Vidu | SceneMixer |

| Built around | Video generation with multi-reference consistency, 2D animation, off-peak mode and batch generation1 | One series project: script → cast → episode outline → storyboard per episode → video per segment → one file per episode |

| Clip length | 3–16 second generations1 | Up to 30 seconds per segment on some tiers; an episode is as many segments as its storyboard needs |

| References | 10 references a month on Free; 300 on paid plans1 | The cast is generated once as reference sheets and attached to every segment a character appears in; paid plans can upload their own image instead |

| Price to start | Free plan; Standard listed at $10/month (shown at $8.25 with a first-purchase discount), 800 credits a month1 | Credit packs from $1.49; Pro plan $7.99/month with 700 credits. New accounts get 50 free credits and 15 free reference images. |

| Commercial use | “Generated content can be used commercially” on paid plans1 | You own the results, including commercial use, on free allowances too; uploads and outputs are not used to train models unless you agree. |

Models differ between platforms, so compare a test clip of your own scene, not only the price per second.

## What a fixed cast looks like

Reference sheets SceneMixer generated once for “Rust City Knockout”. Every segment that features one of these characters is rendered against their sheet, so the face does not have to be re-described shot by shot · watch the sample

## Stay with Vidu if…

- You want quick clips with reference consistency from one app1.

- You generate at off-peak times to stretch your plan1.

- You make 2D animation clips one at a time1.

## SceneMixer fits if…

- You are telling a story in episodes. One read of the script returns the cast, locations, props and an episode-by-episode outline with an end-of-episode hook; each episode gets its own storyboard. Upload a script and read it.

- Your cast should speak a particular language. Pick it in step 1 and the lines are performed in that language as the shot renders — no dubbing pass. Native dialogue vs dubbing.

- You are adapting one story for a market. The target region adapts names, casting and scenery for Western, East Asian, Southeast Asian or Latin American audiences. Relocate a story by region.

- You would rather pay as you go. Packs start at $1.49 and plan credits stay in your balance. Do credits expire?

- You want commercial rights from the first credit. Our terms give you ownership of the results, including commercial use, on free allowances too.

## Where Vidu lands in SceneMixer

| In Vidu | Closest step in SceneMixer |

| References | Character, scene and prop sheets in step 2. Regenerate or replace one sheet |

| A clip | A storyboard segment, joined into an episode |

## Moving a story over

- Bring the story — a script, a document, or one sentence.

- Set the project in step 1: Type (Short Drama, TV Series or Film), aspect ratio, dialogue language and target region.

- Reuse character art you own. On a paid plan, upload it in place of a generated sheet; otherwise generate sheets — the first 15 on an account are free.

- Draft cheaply, finish the keepers. Generate at the lowest tier, then regenerate only the segments that will be seen most at a higher resolution. Draft at 480P, finish at 1080P.

## Frequently asked questions

### Is SceneMixer a Vidu alternative?

For a drama in episodes with a fixed cast and spoken dialogue, yes. For quick reference-based clips, Vidu is built around that.

### How much does Vidu cost?

On November 2, 2026, its pricing page listed a free plan, Standard at $10 a month (shown at $8.25 with a first-purchase discount) with 800 credits, Premium at $43 and Ultimate at $230. Check vidu.com/pricing for current prices.

### Can SceneMixer make 2D animation?

Yes. Choose a 2D style in step 1 and every sheet and shot is drawn in it. Make an AI animated short film.

## Sources and corrections

- Vidu Pricing Plans for AI Video Generation (vidu.com/pricing) — plans, credits, clip length and features

Facts about Vidu come only from the pages listed above, as checked on November 2, 2026. Prices and plans change often, so check Vidu's site before you buy. Vidu is a trademark of its owner and is named here only to identify the product; SceneMixer is not affiliated with it. Spotted something out of date? Email support@scenemixer.com and we will correct it.