Editing Existing Videos with AI

You can edit an existing video using natural language — describe what you want changed and the AI will modify the video while keeping the rest intact.

How it works

1. Go to Generate and select Video as the asset type
2. Under the source tabs, select the video edit option
3. Provide your source video — either:
   • Select from the media library (previously generated videos)
   • Paste an external video URL
   • Upload from your device
4. Write a prompt describing the edit (e.g. "Add snow falling in the scene" or "Change the woman's outfit to red")
5. Tap Generate

The AI understands the video content and applies your requested changes while preserving the rest of the scene.

Limitations

• Maximum input duration: 8.7 seconds
• Format: MP4 only
• Output resolution: Matches input, capped at 720p
• Output duration: Matches input (you cannot change duration when editing)
• Aspect ratio: Matches input (cannot be changed when editing)

Tips for better edits

• Be specific — "Add a red hat to the person" works better than "change something"
• One edit at a time — the AI handles focused edits better than multiple simultaneous changes
• Use short source videos — edits work best with clips under 5 seconds
• Try concurrent edits — generate multiple different edits of the same source video to compare results

Image to video

You can also animate a still image into a video:

1. Generate an image first (or use one from your media library)
2. On the job detail page, tap Create Video
3. Describe how you want the image to animate
4. Set your desired duration (up to 15 seconds)
5. Tap Generate