AI image and video generator

AI video and image generator: what each output is for
| Output | What it is | Where you see it |
|---|---|---|
| Character sheet | A reference picture of one person, made once and attached to every shot they are in | Step 2, the character card |
| Location and prop sheets | The same for a place or an object that matters | Step 2, the location and prop cards |
| State pictures | An alternate look that lasts: a wedding dress, a bandaged arm, a room after the fire | On the card, under its states |
| Cover | The poster picture for an episode or a series | The cover panel |
| Video segment | The shots of one storyboard segment, rendered with voices and sound | The editor, segment by segment |
The order is the point: the images are drawn and approved first, so that every video segment starts from the same faces and places. Your scene image becomes the set.
The image models, and what a picture costs
Each picture is drawn by the image model you pick in the image settings. A few of them, with the price of one picture:
| Image model | One picture |
|---|---|
| Qwen-Image 3.0 | 7 credits |
| SeedDream 5.0 Lite | 8 credits |
| Qwen-Image 3.0 Pro | 18 credits |
| Nano Banana 2.1 | 19 credits |
| SeedDream 5.0 Pro | 25 credits |
| Nano Banana Pro | 46 credits |
The first 15 setup images on an account are free on the default Qwen-Image 3.0; the other models are charged from the first picture. A sheet can be redrawn as often as you like, each time at its model’s price, and the earlier drawings are kept. Which image tier for sheets compares them one by one.
The video models, and what a second costs
Video is rendered segment by segment on the video model chosen for the project, and any single segment can be rendered on another. A few of them:
| Video model | One second |
|---|---|
| Wan 3.0 · 480P | 10 credits a second |
| Wan 3.0 · 720P | 19 credits a second |
| MiniMax H3 | 25 credits a second |
| Seedance 2.5 Flash | 25 credits a second |
| Kling 4.0 | 38 credits a second |
| Seedance 2.5 Pro | 75 credits a second |
A segment runs up to 30 seconds on the tiers that allow long takes. Changing the model for one segment does not change the pictures it is made from, so every model starts from the same faces. Wan 3.0 or Seedance 2.5.
Why the images come first
A video model asked for the same person twice, from words alone, draws two different people. Given the same picture twice, it keeps the face, the hair and the clothes. That is why the images are made and checked before any video: a wrong face is a cheap picture to redraw, and an expensive clip to render again.
- Check the sheets at the size they will be seen, on a phone if the video is vertical.
- Redraw the sheet, not the clip, when a face is wrong in every shot.
- Add a state picture for a change that lasts, rather than describing it in every shot.
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: set Type to Short Drama; Scene Style Street Life; Costume Style Streetwear; Visual Style Digital HD; Target Region; dialogue language; 9:16 Portrait. Then read the script.
- Step 2: Draw the cast and places, and look at each sheet before you spend anything on video. 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.
| Images | Video | |
|---|---|---|
| What | Cast, places, props, states, cover | Each storyboard segment |
| Priced | Per picture, by image model | Per second, by video model |
| Free | The first setup images on an account, on the default model | One short preview |
| Made | First, and approved by you | After, from the approved images |
FAQ
Is there an AI image and video generator in one?
Yes. SceneMixer draws the images a story needs and then renders the video from them, in the same project.
Can I generate a single image without a video?
On SceneMixer, images are made inside a project: character, location and prop sheets, their states and covers. Each can be drawn without rendering any video.
Which image model is the default?
On SceneMixer the default is Qwen-Image 3.0 at 7 credits a picture; other models are picked in the image settings.
Are the first images free?
Yes, on the default model: the first 15 setup images on a SceneMixer account are free on Qwen-Image 3.0. Other image models are charged from the first picture.
Can I mix video models in one episode?
Yes. SceneMixer lets you render any segment on another tier, and every tier starts from the same cast pictures.
Can I use my own picture instead of a drawn one?
Yes, on a paid plan: SceneMixer’s Upload & Replace puts your picture on a character, location or prop card.
What does an image cost compared with a second of video?
On SceneMixer’s defaults a picture is 7 credits and a second of video 10.
Is it an AI photo and video generator too?
In both directions: SceneMixer draws the cast, places and props as pictures and renders the video from them, and for an ad it starts from your own photos.
Updated November 19, 2026