Create Videos With AI: Start From What You Already Have

What you have, and the door it uses

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, 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. 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, 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

Upload 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 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

1. Script: paste your own on Upload Script, or write a one-line brief in AI Create and let it draft the episodes.
2. 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.
3. 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.
4. Storyboard: generate episode 1 first (free on your first drama) and read it before you render anything. One place per segment.
5. 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

Updated November 2, 2026