Turn a prompt — or one extracted from a video you like — into a 4–30 second AI video. Pick a Seedance model, set the format, and guide it with reference images.
Why it works
Analyze a reference video, keep the scene prompts you like, and generate new videos from them in one place.
Every scene in a video analysis has a Generate video button that brings its prompt straight here.
Lock the first frame, or add reference images and mention them as @Image1, @Image2 in your prompt.
Credits are charged per output second, shown before you generate. Failed generations are refunded automatically.
Seedance generates dialogue, effects and ambience in sync with the picture. Turn it off for silent clips.
Analyze any reference video to get scene-by-scene prompts you can generate from.
Tip: start from a real video's scene prompt. Extract a prompt from a video →
Prompts, images and generated videos are checked automatically. NSFW, sexually suggestive, deepfake, violent or hateful content is not allowed. Acceptable Use Policy
The video starts exactly from this image.
Create synchronized sound with the video.