More about this tool
Act out a scene yourself and let any character perform it. This is AI motion capture without a suit, markers or a special camera.
How it works
1. Upload a photo or video of the character you want to bring to life.
2. Upload a performance video of a person talking and moving, up to 30 seconds long.
3. Pick a shape and an expression level, then create. The character copies the face and body motion, with the voice from your performance clip.
Tips for better results
• Film the performance in good light, with the face clearly visible and the camera steady.
• Use a character image where the face and upper body are easy to see.
• Turn off Body Control in the advanced settings if you only want facial expressions and lip sync.
• Raise Expression Intensity for cartoon characters, lower it for a calmer, more natural look.
Good to know
What files can I use?
For the character: PNG, JPG or WebP images, or videos like MP4, MOV and WebM. For the performance: MP4, MOV, WebM and other common video formats.
How long can the performance video be?
Up to 30 seconds.
Where does the audio come from?
From your performance video, so you can record your own voice and have any character say your words.
Which shapes can I choose?
Vertical 9:16, portrait 3:4, square 1:1, landscape 4:3, widescreen 16:9 and ultra-wide 21:9.