Same characters, outfit and light in every clip. Paste a script or a one-line idea.
1Script or ideapaste it in the box above: one sentence works
2Cast sheeteach character and setting described once
3One prompt per shotwith the sheet repeated in every one
Write the script or idea in the box above (one sentence is enough).
Why characters change between clips
Paste a script or a one-line idea. You get a cast sheet, with each character, outfit and setting described once, and a ready prompt for every shot in Veo 3.1 or Gemini Omni Flash. The sheet is repeated word for word in each prompt, so the face, the clothes and the light stay the same from clip to clip.
Each clip is generated on its own. The model does not remember the previous shot, so if shot 2 only says "Ana", it invents a new Ana: different hair, different clothes, different light.
What ties the clips together is the text you repeat. The cast sheet holds those descriptions in one place, and every shot prompt copies them exactly. Edit the sheet once and every prompt updates.
If you also use reference images in Flow, keep them: the text and the images then describe the same person.
How it works
1
Paste the script
A full script, a paragraph or one sentence. Pick the model and the target length.
2
Check the cast sheet
Characters, outfits and settings, each described once. Edit anything: no credit, no new run.
3
Copy shot by shot
One prompt per shot, with camera, action, sound and the sheet repeated. Paste each one into Flow, the Gemini app or the API.
You can split one script a day without an account. A free account gives you 3 a month (shared with the other AI actions) and lets you save projects and edit them later. Pro is unlimited.
Which video models does it write for?
Veo 3.1 and Gemini Omni Flash. The shot lengths follow each model: Veo clips are 4, 6 or 8 seconds; Omni clips run 3 to 10 seconds.
Does it generate the video?
No. It writes the prompts. You paste each one into the tool where you generate: Flow, the Gemini app, the API or a site that resells these models.
What about an animated explainer with charts and text?
Pick "Animated" as the model. Instead of shots for a video model, you get scenes and a prompt to paste into Claude or another assistant, which writes the on-screen text and numbers. That mode needs a free account.
What happens to my script?
It is sent to the AI model that splits it into shots. We keep the result in a cache so the same script does not cost twice, and we never use it to train models.