The output loses the source motion
Use a shorter, clearer source video and keep prompt changes limited.
VIDEO TO VIDEO AI
Quick answer
Video to video transforms an existing clip into a new AI video. Use a source video for timing and motion, then add a subject or style image and prompt for what should change.
Input requirements
Match the source type to a conservative first motion target. That improves the keeper rate before spending credits on longer clips.
| Input | Best for | Avoid |
|---|---|---|
| Source video | Motion, timing, camera rhythm, pose, or action reference | Long clips, rapid cuts, or heavy camera shake |
| Subject or style image | Guiding who or what should appear in the transformed clip | Conflicting style or identity references |
| Prompt | Clear change direction and preservation rules | Changing both subject, style, action, and camera at once |
Failure reasons
Use a shorter, clearer source video and keep prompt changes limited.
Provide a stronger subject or style image and avoid contradictory wording.
Ask for one transformation goal instead of rewriting the whole clip.
FAQ
It is best for remixes, style changes, subject changes, and motion-preserving transformations from existing footage.
Short clips between 3 and 30 seconds are easier to control and estimate.
This workflow is strongest with a subject or style image. Use text to video when you only have a prompt.
Related pages
Use character replacement when the main goal is swapping the visible subject.
Open replace characterREFERENCEUse a reference image when you want image guidance without a source video.
Open reference to videoPROMPTGenerate a new clip from text when no source footage is available.
Open text to video