AI Video
Swap
Upload a clip and reference photos, and swap in a different person, pet, object or outfit while the motion, camera, timing and original sound stay the same.
Swap a person, pet or object
Genjutsu Light Object Swap
Replace one subject in your clip with the one in your photos. The shot, the movement and the sound stay.
3–11 credits / secondOpen the swap studioPut yourself in the video
Genjutsu Light Motion Transfer
Your reference photos take over the whole performance: face, hair, body and outfit follow the clip's motion.
3–11 credits / secondOpen the swap studioFull-length swap
Genjutsu Object Swap
For clips longer than 15 seconds: the full Genjutsu model swaps across the whole video at 480p or 720p.
9–18 credits / secondOpen the swap studioAI video swap replaces a subject in an existing clip with someone or something from your reference photos. On imageat it runs on Higgsfield Genjutsu inside the video studio, keeping the clip's motion, camera moves, timing and original audio. Light modes use the first 15 seconds from 3 credits per second; full modes use the whole clip.
Swap the Subject, Keep the Shot
Start Now!People, pets, objects, outfits
Replace a person, an animal, a product or what someone is wearing using reference photos.
Motion and camera kept
The original movement, camera path and timing carry over to the swapped subject.
Original sound kept
The clip's audio stays in place, so dialogue and music still line up.
Light modes
Genjutsu Light Object Swap and Light Motion Transfer use the first 15 seconds at 3, 6 or 11 credits per second for 480p, 720p or 1080p.
Full modes
Genjutsu Object Swap and Motion Transfer process the whole clip at 9 credits per second at 480p or 18 at 720p.
Multiple references
Add up to five reference images where supported for a closer match from different angles.
How it works
Upload the clip
Choose a source clip of at least 4 seconds.
Add references
Upload photos of the person, pet, object or outfit to swap in, and pick a Light or full mode and resolution.
Generate the swap
Higgsfield Genjutsu renders the new clip with the original motion and sound.
AI Video Swap
specs
| Engine | Higgsfield Genjutsu, in the imageat video studio |
|---|---|
| Light modes | Light Object Swap, Light Motion Transfer (first 15 seconds) |
| Light pricing | 3 cr/s at 480p, 6 at 720p, 11 at 1080p |
| Full modes | Object Swap, Motion Transfer (whole clip) |
| Full pricing | 9 cr/s at 480p, 18 at 720p |
| Minimum clip | 4 seconds |
| References | Up to 5 images where supported |
| Kept from source | Motion, camera, timing, original sound |
Get better videos
Test with Light at 480p
Try the swap on the first 15 seconds at 3 credits per second before running the full clip.
Use clear references
Well-lit photos from a few angles help the swapped subject hold up as it moves.
Pick a steady clip
Clips where the subject is clearly visible and not heavily blocked give cleaner results.
Swap with consent
Only swap in yourself, people who agreed, or your own pets.
Paste one and generate
- Swap my dog into a viral dancing clip
- Put myself into a movie-style walking shot
- Replace the sneakers in my product video with the new colorway
- Change my outfit in a dance video to a suit
- Swap a coffee cup for my brand's mug in a lifestyle clip
Pay per video
The script, voice, pictures, music and MP4 render are all included. A failed video refunds its credits automatically.
AI Video SwapMake your first
video swap
Upload a clip and reference photos, and swap in a different person, pet, object or outfit while the motion, camera, timing and original sound stay the same.