Feedback
minimax face swap
MiniMax face swap recasts a finished clip without touching its timing, camera moves, cuts or audio. 5–30s, four people.
All Tools
Discover our comprehensive AI-powered animation toolkit
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.
Seedance 2.0
The Future of AI Video Is Here.
Veo3.1
Create Stunning Videos with Veo3.1
FLUX 3 Video Generator
Kling 3.0
Next-Gen AI Video Generator
Grok Video Generator
Create Videos from Text or Images with AI
MiniMax H3 video generator
MiniMax H3 AI Video Generator
Why use MiniMax face swap
A recast tool, not a generator: the head turns, lens movement, cuts and soundtrack stay, and only the faces change.
- A casting swap, not a new generationNothing about the motion or the edit is regenerated; the tool re-casts footage you already have, so the output slots back where the original left off.
- Left-to-right identity mapping with prompt overrideOne photo per newcomer, matched left to right in every shot and held across cuts; a prompt can remap the pairing.
- The source clip is the performanceHow heads turn, how a gesture lands, the on-set light and the lens drift all survive; faces and outfits come from your photos.
How MiniMax face swap works
Three steps from a finished take to a new cast, with the length and sound of the source left as they were.
Where MiniMax face swap holds up
Built for takes that are already right except for the casting: motion, edit rhythm and audio stay, and only the people change.
Cast up to four, left to right
Four photos max, one per person, each taking over a main role in every shot it appears in; a prompt can override the order.
Performance stays, people change
The acting survives: gestures land on the same beat, light falls the same way, the camera drifts as before. Faces and outfits are the new part.
Cuts and sound carried over
The cut pattern and the mix are inherited, which means the swap can sit in a locked edit without any re-timing or re-recording.
768p test pass, 1080p final
Preview at 768p to see if the likeness holds, then commit to 1080p; output runs as long as the source, so price tracks duration.
What makes a reference photo work
A sharp, evenly lit, front-facing portrait is ideal; a half-body frame carries hair and clothing better than a tight headshot. Skip sunglasses, hats and group shots.
Where the recast holds and where it slips
Plan around crowds, distant faces, fast action and occlusion after hard cuts; in long shots or profile turns the likeness can slide back toward the original performer.
MiniMax face swap — FAQ
Common questions about MiniMax face swap: source limits, casting order, reference photos and where the recast holds.
What does MiniMax face swap actually do?
It puts the people from your photos into footage you already have, keeping the source motion, camera work, cuts and audio — a casting change on a locked clip.
How long can my source video be?
5 to 30 seconds, with no shot longer than 15 seconds. Split longer takes first; the recast comes back at the same length you sent.
How many people can I recast, and in what order?
Four at most, one photo each. By default photos take the main roles left to right; a line like 'the woman on the right becomes person 1' flips that order.
What kind of reference photos work best?
Sharp, evenly lit, front-facing portraits, ideally half-body so hair and clothing carry. Skip sunglasses, hats, hands across the face, group shots and heavily filtered images.
Should I test at 768p or render at 1080p?
Do a quick 768p pass to see whether the casting reads correctly, then render at 1080p once it does; price follows clip duration.
When should I not use it?
Not when you need fresh motion, new camera moves or a different location — the source decides those. Crowds, distant faces and fast action after hard cuts are weak spots too.
Try MiniMax face swap on Fuser
Drop new faces into a finished clip and keep its timing, cuts and soundtrack. Price follows clip length, with MiniMax H3 and Hailuo 2.3 close by.