A dancer mid-routine in a rain-slick neon alley at night.

Higgsfield AI Genjutsu

Genjutsu is Higgsfield's video-to-video tool: it keeps the motion of a source clip and swaps who performs it, where it happens, or one prop. This page is the free route to the same result.

Three Genjutsu-Style Swaps You Can Copy

Each prompt names what stays — the motion, the camera, the light — before it names what changes. That order is what keeps a swap looking like one take instead of two clips stitched together.

  • A dancer mid-routine in a rain-slick neon alley at night.

    Same dance, new world

    Keep the choreography and the camera path from my clip and move the same performance into a rain-slick neon alley at night: wet asphalt, sign reflections, sodium and magenta spill on the dancer's jacket. The body motion and timing stay frame for frame; only the world changes. Cinematic grade, shallow background blur. No text or signs in focus.

  • A single performer dancing a set routine in a bright loft studio.

    Same routine, new lead

    Keep the routine in my uploaded clip exactly as performed, camera and timing untouched, and replace the dancer with the person in my photo (cropped knit top, wide-leg cargo trousers, bare arms). Same studio, same window light from the left, same floor shadow. The face comes from the photo; the movement comes from the clip. No text or graphics.

  • A hand reaching for a tall ceramic vase on a wooden table in a sunlit room.

    One prop swap, mid-move

    Keep the shot in my clip as it is and swap only one prop: the object the person reaches for becomes a tall ceramic vase instead of a mug, same position, same size in frame, same hand contact. Everything else stays — the room, the light through the window, the fabric of the knit, the camera. Photoreal, natural contact shadows. No text.

Where a Motion Swap Breaks

Three things give a swap away: the new performer moves differently from the source, the light changes with the person, and a swapped prop stops touching the hand that holds it.

A single performer dancing a set routine in a bright loft studio.

The motion is the expensive part

A swap works because the performance already exists: the timing, the weight, the little off-beat in the shoulder all come from the source clip, so the model only has to draw a different body doing the same thing. Say that out loud in the prompt — keep the routine, the camera and the timing — before you describe the new lead. If you are starting without a clip, the Photo to Video AI Free builds the motion from a still instead.

A dancer mid-routine in a rain-slick neon alley at night.

A new world is a light change first

Moving a performance somewhere else fails when only the background changes: the dancer keeps studio light while the street is neon, and the shot reads as a cut-out. Name the light the new place throws — sodium and magenta on the jacket — and the same AI Video Editor relights the body with the room.

A hand reaching for a tall ceramic vase on a wooden table in a sunlit room.

Prop swaps fail on contact shadows

Replacing a mug with a vase looks fake the moment the fingers do not wrap the object or the shadow misses the table. Pin the position, the size in frame and the hand contact, then check the contact shadow. When the prop is the whole idea, the AI Image Generator is quicker for testing the look before you animate it.

How to Run a Genjutsu-Style Swap Here

A dancer mid-routine in a rain-slick neon alley at night.

Where a Motion Swap Earns Its Keep

One still covers most of what people want from this trend. The wording of the prompt changes with who is in the picture, not with the room.

  • A single performer dancing a set routine in a bright loft studio.

    One shoot, several leads

    Reuse a finished routine for a second creator without reshooting it: the choreography is already approved, only the performer changes. The AI Video Editor takes the source and the description.

  • A dancer mid-routine in a rain-slick neon alley at night.

    Testing a look before you shoot

    Try three worlds against the same routine before you book a location. Wet neon, warm kitchen, empty gallery — the motion is free after the first pass.

  • A hand reaching for a tall ceramic vase on a wooden table in a sunlit room.

    Fixing one detail, not the take

    The clip is right except for the object in the hand or the jacket colour. A single prop swap saves the take; the AI Image Generator settles the look first if you want to be sure.

Higgsfield Genjutsu FAQ

What is Higgsfield Genjutsu?

It is Higgsfield's video-to-video tool: you bring a source clip, and it re-renders the shot with a different performer, a different environment, or a different object while keeping the motion of the original. People use it to reuse choreography, restyle a scene, or fix one detail in a take.

How is this page different from using Genjutsu?

Genjutsu runs the swap on Higgsfield's own models and credits. Here you describe the swap in words and run it through the AI Video Editor, or build the shot from your photos if you have no source clip. Same outcome, different route, and you can try it without buying a subscription.

Is it free to try here?

You can run a first pass without signing up, so testing one swap costs nothing beyond the render itself. We do not promise unlimited free use, and account-gated capabilities stay gated.

Do I need a source clip?

No. With a clip, the existing motion carries the shot. Without one, generate the performance from a prompt and a photo — the Photo to Video AI Free is the shorter path.

Will the new performer keep my face?

Upload a clear, front-facing photo and name it as the only face reference, then state the rule twice: face shape, hair and age come from the photo, the body follows the original performance. A small face inside a group shot loses too much detail to hold.

Why does the swap look like two clips stitched together?

Usually because the light did not move with the person. The body keeps studio light while the new room is neon, or the shadow stays where the old floor was. Name the light of the new place and the shadow it throws and the seam mostly disappears.

Can I use a swapped clip commercially?

Treat the result as generated footage and check the rules of where you post it, plus anyone whose likeness appears. Do not present a generated performer as a real endorsement, and do not recreate a specific artist's likeness or a real product's packaging.

What happens to the clip and photos I upload?

They are used as the reference for the render you ask for. Keep to material you have the right to use, and avoid uploading footage of people who have not agreed to appear in a generated version.

Tools that restyle the same clip

Try the swap before you pay for credits

Run one version here, compare it with your source, and keep the take that holds the motion. The clip you already shot is the expensive part.