Three steps. About a minute.
Upload
Drop in a photo or a clip. JPG, PNG or WebP up to 25 megapixels; MP4, M4V, MOV or WebM for video.
Pick the faces
Every face is detected for you. Use one source face for everyone, or give each person their own.
Generate
The cost is shown first. Run it, then download from your library at full resolution.
Every face found. You choose who changes.
Group photographs are where most tools stop being useful. Every face in the frame is detected and listed, and you decide what happens to each one — swap everybody from a single source, pick out one person, or give each face its own.
- Up to 25 megapixels in, 25 megapixels out
- Any number of faces in one photograph
- A different source face for each person, if you want one
- Face enhance included, at no extra cost
The whole clip, and only the seconds you use.
Every face is followed from the first frame to the last, through movement, cuts and changes of angle, and the original audio comes through with it. What is less usual sits underneath: an interrupted render carries on from where it stopped instead of starting again, and you are charged for the seconds you actually used.
- 480p through 1440p, clips up to 10 minutes
- Every face in the frame, held across cuts and angles
- An interrupted render continues rather than restarting
- Charged by the second, never rounded up to a minute
Full resolution out
What you upload is what you get back. No silent downscaling.
Face enhance included
The optional detail pass costs nothing extra on any plan.
Credits do not expire
Buy them when it suits you and use them when you need them.
Failed renders refunded
If a job fails through no fault of yours, the credits come back.
Pay for swaps, and see the price first.
One credit system across both tools. Video is billed by the second rather than rounded up to the minute, and the exact cost appears before you run anything.
Straight answers.
How many faces can I swap in one photo?
As many as are in the frame. Every face is detected automatically, and you can give each person a different source face, swap only some of them, or swap them all from one source.
What are the limits for video?
480p through 1440p, and clips up to ten minutes on a paid plan. The face is tracked frame by frame, and the original audio is kept in sync.
How do credits work?
An image costs one credit. Video costs 30 credits a minute, billed by the second, so a 40-second clip costs 20. The exact figure is shown before you run the job, and credits do not expire.
What happens to my files?
The file you upload is deleted as soon as the job has finished with it. Results stay in your library for two days on the free tier and fourteen on a paid plan, and anything you pin is kept for ninety.
Where is any of this stored?
In the European Union. The service runs on a server in Germany and media is stored in EU-resident storage.
Is my media used to train models?
No. The models are pre-trained and fixed; nothing you upload is used to change them.
Your first 20 credits are on us.
Enough for twenty images, or forty seconds of video. No card, and nothing to cancel.