Video to video (edit a video)
Upload a video and change it with AI: outfit, background, lighting, style or action, while everything else stays the same.
Video to video edits an existing clip. You describe the change, and the model keeps the rest: the people, the motion and the timing.
How to use it
- Go to Create → Video to video, drop a video, or press Edit on any video in your Library.
- Describe what to change and what stays the same: "change her dress to a red silk slip; keep the room, the motion and the lighting".
- Pick the aspect ratio ("Same as source" by default) and quality.
- Tick the rights checkbox and press Generate.
Don't start your prompt with "Edit the video:"; the app adds that for you.
Good edits
- Outfit: "she now wears a black lace robe"
- Background: "move the scene to a hotel room with a sea view at night"
- Lighting: "warm candlelight instead of daylight"
- Style: "make it look like grainy 16mm film"
- Action: small changes work better than a whole new action
The Timeline uses the source length in this mode, so you can describe changes part by part.
Limits
- Wan 3.0 / Prime: the source can be at most 15 seconds. Trim it, or use Seedance.
- Seedance 2.5: no length limit on the source.
- The output has the same length as the source.
Cost
Video to video is billed for the output plus the input: 2 × the source length × the price per second. A 6 s source with Wan 3.0 at 480p costs (6 + 6) × 25 = 300 credits.