Start from a clip you already have. Genjutsu AI keeps the original motion, camera, and timing, then rebuilds the cast, location, product, or a single object around it. Choose Motion Transfer for a full recast, or Object Swap when only one element should change.
Upload a source video and a reference image, then describe the change.
Video: MP4/MOV · Image: JPG/PNG · Up to 10 MB for images
Reference preview · Upload a video, image, and prompt to generate
GENJUTSU AI EXAMPLES
WORKFLOW
Pick Motion Transfer or Object Swap, add a source video and references, then describe the change and generate.
Use Motion Transfer when the motion, camera, cuts, and timing should stay, but the cast, location, or visual world should be rebuilt. Use Object Swap when the shot already works and you only want to replace a character, outfit, product, location, or object.
Upload one short source video, then add clear reference images for the new character, product, wardrobe, object, or location. Front-facing, well-lit references with visible face, costume, or product detail stay more consistent than text alone.
Write what should change and what must remain: motion, camera path, timing, composition, or the surrounding scene. Preview the result, then adjust one variable at a time if identity, lighting, or contact points drift.
A useful Genjutsu AI edit starts from a real performance, then changes only the parts you name.
Keep the original action, body movement, shot duration, camera motion, and edit timing. Rebuild the character, location, wardrobe, product, or scene from your references.
Replace one selected element — a person, outfit, product, prop, or location — while the background, other people, lighting direction, pose, and camera stay in place.
Show the new look with images instead of describing it only in text. Consistent references for face, costume, product shape, or environment reduce drift across the clip.
The source clip is the reference for framing, perspective, camera trajectory, cuts, and performance. Genjutsu AI changes the visual world without inventing a new camera move.
Restage a location, shift the look, replace a featured product, or swap a single object in frame. Name the change, and name what the original subject and motion must keep.
Upload source video and reference images you own or have permission to edit. Build an original variant of your shot rather than copying a protected performance or identity.
Answers about Motion Transfer, Object Swap, references, credits, and how Genjutsu AI differs from text-to-video.
Genjutsu AI is a video-to-video generator based on the Higgsfield Genjutsu workflow. You upload an existing clip and transform it while keeping selected parts of the original shot. Motion Transfer rebuilds the scene around the original motion and camera. Object Swap changes a selected element and preserves the rest.
Motion Transfer keeps the source motion, camera, cuts, and timing while rebuilding more of the visual scene — cast, location, wardrobe, product, or world. Object Swap changes one selected character, outfit, product, location, or object and tries to leave everything else intact. Choose the mode by how much of the original shot you want to preserve.
Choose Motion Transfer or Object Swap, upload a short source video, add reference images when the new look needs to be specific, and describe what should change and what must stay. Generate a preview, then check identity, object shape, contact, lighting, and continuity.
Use clear, well-lit images of the character, product, wardrobe, object, or location you want to introduce. Front-facing references with visible facial, costume, or product detail are more stable. If identities drift, shorten the source clip and simplify the cast.
Higgsfield documents Genjutsu for short source clips, commonly up to 30 seconds. Use a brief shot and follow the duration options shown in the generator. Longer clips and higher resolution use more credits.
Text-to-video invents a new shot from a prompt. Genjutsu AI starts from your footage. The original motion, camera path, and timing stay with the source clip, and the prompt plus references describe only the change.
No. This site bills Genjutsu AI with prepaid credits, per generation. Use the official Higgsfield app if you want the full Higgsfield suite on a subscription. Use this site when you want Motion Transfer and Object Swap without a monthly seat.
Each generation costs a flat amount by resolution: 720p uses 300 credits, 1080p uses 600 credits. The cost is shown in the generator before the job starts. There is no subscription.
Commercial use follows the Terms of Service. Only upload video and reference images you own or have permission to edit, and do not present a generated result as a real recording of someone else.
No subscription. Buy a one-time credit pack and spend credits on Genjutsu AI. Each job costs 300 credits at 720p or 600 credits at 1080p.
Starter
300 credits
1 × 720p Video
Creator
900 credits
About 3 × 720p Videos
Pro
1,500 credits
About 5 × 720p Videos
Business
3,000 credits
About 13 × 720p Videos
Upload a source video, add references for the character, scene, or product, and generate a Motion Transfer or Object Swap.