AI character swap for your own video
Film it once as yourself, then become someone else. Upload a clip of up to 60 seconds, choose a character from your library or from a few photos, and the same performance comes back with a different person in it.
On every plan from $9 a month·7-day free trial
Your performance, someone else on screen
Everything except the person stays exactly as you filmed it
This is not a generated scene that approximates your video. It is your video, with the person replaced.
- The performance: motion, gestures and timing exactly as filmed
- The camera: framing, movement and shot composition
- The scene: room, lighting and props
- The audio: your original track, re-attached bit-for-bit unchanged
- The speech: the mouth follows the original words, so it stays in sync
- The format: the aspect ratio you shot in, at 1080p on the short side
How it works
Upload your clip
MP4, MOV or WebM, 3 to 60 seconds, up to 100 MB. It should be your own footage, or footage you have the rights to use. Any aspect ratio works and the output keeps whichever one you started with.
Choose who they become
Pick a character you have already saved, one of the presets, or upload one to four photos of the character you want on screen. The same saved character can be reused across clips so a series stays consistent.
Render and download
The person is replaced and everything else is left alone. It takes a few minutes rather than seconds, because every frame is regenerated. Output is 1080p on the short side at 30 fps, in the aspect ratio you filmed.
What people use it for
One shoot, many presenters
Film an ad read once and ship it with several different people delivering it. The performance is identical, so the only variable under test is who is saying it.
UGC ad creativeStay off camera without losing the performance
Some people are good on camera and would rather not be recognisable on it. Film it yourself, publish as a character, keep your own delivery.
Faceless channelsA recurring character who is actually you
Save the character once and reuse it. Every episode has the same face and your timing underneath it, which is what makes a series feel like a series.
AI influencerLocalise the presenter, keep the take
Run the same footage with presenters suited to different audiences, without booking a second shoot for each one.
Spokesperson videoFootage you have the right to use
Before each render you confirm that you own the video or have permission to use it, that you have the right to use the likeness you are swapping in, and that you will not use it to impersonate or deceive anyone. That is a condition of use, not a formality: using this to misrepresent a real person breaches our terms.
Platforms including TikTok, Instagram and YouTube require you to label synthetic media, and the EU AI Act requires visible labelling where synthetic content is presented as real. Our AI disclosure covers what that means when you publish.
What it costs
18 credits per second of video, plus a flat 10 per render. A 15 second clip is 280 credits; a full 60 seconds is 1,090. It is included on every plan rather than gated to a higher tier, and the 7-day free trial carries 500 credits, which covers a clip of about 27 seconds.
Character swap FAQs
- What is AI character swap?
- You upload a video of a person and choose a character to replace them with. The motion, gestures, camera work, background and original audio all stay as filmed; only the person changes.
- How long can the video be?
- Between 3 and 60 seconds, up to 100 MB, as MP4, MOV or WebM. Clips longer than 15 seconds are processed in parts and joined back together.
- What does character swap cost?
- 18 credits per second of video plus a flat 10. A 15 second clip is 280 credits and a 60 second clip is 1,090. It is available on every plan, starting at $9 a month.
- Can I try it on the free trial?
- Yes. Every plan opens with a 7-day free trial carrying 500 credits, which covers a clip of roughly 27 seconds.
- Do I need a photo, or can I use a saved character?
- Either. Use a character already in your library, pick a preset, or upload one to four photos of the character you want. Saving the character means you can reuse the same face across a whole series.
- Does the audio change?
- No. Your original audio track is re-attached to the finished video exactly as recorded, bit for bit. Nothing is regenerated or re-voiced, and the mouth follows the original speech so it stays in sync.
- Whose footage am I allowed to upload?
- Your own, or footage you have permission to use, with the right to use the likeness you are swapping in. You confirm this before each render. Using it to impersonate or deceive anyone is not permitted under our terms.
- How long does a render take?
- A few minutes rather than seconds, because every frame is regenerated rather than filtered. An eight second clip took about six minutes in testing.
- What aspect ratio does it output?
- The same one you uploaded. Vertical 9:16, landscape 16:9 and square 1:1 all work, rendered at 1080 pixels on the short side at 30 fps.