Higgsfield Genjutsu · Character swap

Put anyone in the Hotel Lobby video.

The Higgsfield setup and the copy-and-paste prompt to swap two characters into the viral rap performance.

You've seen the AI versions of Quavo and Takeoff's COLORS performance with different characters doing the exact same moves. Same camera, same background, same audio, new faces. This is how it's made, step by step, with the prompt that keeps each character glued to the right performer.

Same performance, same audio, two new characters
6
Steps from download
to finished clip
1
Copy-and-paste
prompt
2
Reference images
(one per performer)
24s
Our test clip
720p · 16:9

Clip length, resolution and credit cost are from our own generation. Higgsfield's pricing and options change, so check the cost shown before you hit Generate.

Free Guide

Get the Hotel Lobby prompt.

Drop your email and the whole guide unlocks right here, with a copy button on the prompt and a Markdown download. You'll also get one practical AI system most weeks.

  • The six-step Higgsfield Genjutsu setup, in the order the app expects it
  • The exact prompt that keeps Image 1 on the left performer and Image 2 on the right
  • What makes a good reference image, and what makes the faces drift
  • The review checklist before you download (faces, hands, positions, audio)
  • What to change when the characters switch places or the cost is too high
  • How to point the same setup at any other two-person performance
No spam. Unsubscribe anytime. You'll join "Stay Ahead" by AI Systems Lab.
Start here

Same moves. Same song. New people.

This guide makes one video: the Hotel Lobby performance with two characters of your choice swapped in. It uses Higgsfield's Genjutsu tool, one motion reference video and two reference images.

Your first move
Get the clip and the two images ready before you open Higgsfield.
  1. Tick through the four things to get ready. The reference images matter more than anything else.
  2. Open Higgsfield, go to Video, pick Genjutsu.
  3. Upload the clip and the two images, then paste the prompt and link the references.
01 · What you're making

A character swap, not a remake.

1

The motion reference

The original Quavo and Takeoff COLORS clip. Every gesture, head turn and mouth movement comes from here.

2

Two characters

One image per performer. Face, hair, body proportions and clothing come from these.

3

Genjutsu does the swap

Higgsfield replaces each performer with its assigned character and keeps everything else as it was.

The reason this trend works is that nothing is re-animated from scratch. The camera move, the framing, the background, the props, the lighting and the audio are all the original. Only the two people change. That is why the prompt spends so many words on what to preserve and only a few on what to replace.

The one thing that goes wrong most. The two characters swap sides partway through, or blend into each other. The prompt handles this by assigning each image to the performer who starts on that side of the frame, and telling the model to keep the replacement attached as they move.

02 · Setup

Get these ready.

Tick them off as you go. Your progress saves in this browser. 0 / 4

  • A Higgsfield account with credits. Genjutsu is a paid generation. The cost shows next to the Generate button before you commit.
  • The original video, downloaded. Trim it first if you only want a section. Shorter clip, lower cost, faster result.
  • Image 1: the character for the performer on the left. Clear face, outfit visible. Full body if the clip shows full bodies.
  • Image 2: the character for the performer on the right. Same rules. Keep the two images clearly different from each other.
The original clipAdam's copy of the COLORS performance, ready to trim and upload.
Download the original ↓
Good reference image
Front-on, well lit, one person

Face fully visible, no sunglasses, no hand over the mouth. The outfit you want them wearing in the video. Plain background helps.

Weak reference image
Group shot, cropped, filtered

Two people in frame, a heavy filter, a tiny face in a wide shot, or a crop that hides the body when the clip shows full bodies.

If the clip shows the performers head to toe, the model has to invent whatever your image doesn't show. Give it a full-body photo and it has nothing to guess.

03 · The Higgsfield setup

Load the video, then the two people.

The order matters. Motion reference first, then Image 1, then Image 2.

  1. Open Higgsfield and go to Video

    At higgsfield.ai, open Video in the top nav and select Genjutsu.

  2. Pick the swap mode

    Inside Genjutsu you'll see a toggle for Motion transfer and Objects swap. Choose Objects swap. Higgsfield moves its labels around; if your version shows Motion Control instead, that's the same upload flow.

  3. Upload your motion reference

    Upload the original video as the movement reference. This becomes @Video1 in the prompt.

  4. Add your two characters

    Upload Image 1 for the performer who starts on the left of the frame and Image 2 for the performer who starts on the right. These become @Image1 and @Image2.

  5. Turn the prompt on and paste

    Switch on the Prompt field, paste the prompt from the next section, and link each @ reference to the file you uploaded.

Check the order before you spend credits. If you accidentally upload the right-hand character first, the whole clip comes back mirrored. Swap the images, don't rewrite the prompt.

04 · The prompt

Paste this, link the references.

One prompt. It tells Genjutsu who replaces whom, what to take from each image, and everything to leave alone.

The prompt · Hotel Lobby character swap · paste into Genjutsu
Use @Video1 as the source performance throughout the clip. Replace the performer who starts on the left side of the frame with the character from @Image1. Replace the performer who starts on the right with the character from @Image2. Keep each replacement attached to its assigned performer as they move. Use each image for that character's face, hairstyle, body proportions, and clothing. Keep both characters distinct and consistent throughout. Follow each original performer's gestures, head movements, body movement, facial expression, and mouth movements, matching the timing and rhythm of @Video1. Preserve the source camera movement, framing, background, props, and lighting. Keep the characters at the original performers' positions and scale, with natural body movement, believable contact, and shadows that match the scene. Keep the original two-person composition. Preserve the original audio.

How it's built: paragraph one assigns and anchors. Paragraph two says what to take from each image and what to copy from the video. Paragraph three is the preserve list. If you edit it, edit the first paragraph only.

Why each paragraph is there.

  • "Starts on the left" and "starts on the right." Performers cross and lean into each other. Anchoring to the starting position, then saying "keep each replacement attached", stops the swap from flipping mid-clip.
  • "Face, hairstyle, body proportions, and clothing." Naming the four things to pull from the image stops the model from keeping the original performer's outfit or build.
  • "Matching the timing and rhythm." This is a rap performance. Mouth movement that drifts off the beat is the first thing people notice.
  • "Preserve the original audio." Without it, some runs re-synth or mute the track. The output should play back with the original song untouched.
05 · Generate and check

Check the cost, then check the faces.

Look at the settings and the credit cost next to Generate, then click it. Our 24 second test at 720p, 16:9 showed 330 credits on a 55% off promotion (480 at the standard rate). Yours will differ with length, resolution and whatever Higgsfield is running that week.

24s
Our test generation

24.0 seconds, 720p, 16:9. One reference video, two reference images, the prompt above, no other settings changed.

Cost scales with clip length. If you're testing a new pair of images, trim to a few seconds first and only run the full clip once the faces hold.

Before you download.

  • Faces. Both characters look like their image the whole way through, not just in the opening frames.
  • Hands. Watch the mic hand and any gestures near the face. Extra fingers or a melted hand means regenerate.
  • Movement. Gestures, head turns and mouth movement land on the original beats.
  • Positions. The characters haven't switched places. Left is still Image 1, right is still Image 2, even after they cross.
  • Audio. The original track plays back unchanged.

If it passes, download it. If one thing fails, fix that one thing (usually the reference image) and run it again rather than rewriting the prompt.

06 · Make it yours

The same setup works on any two-person performance.

The Hotel Lobby clip is the trend, but the method doesn't care about the song. Anything with two performers who are clearly separated in the frame will work the same way.

  • You and a mate. Two phone photos, front on, in the outfits you actually wear. This is the version people share.
  • Your team. Two people from the business, same rules. Useful for a launch post or an internal joke that turns into a recruiting post.
  • Characters or mascots. Illustrated or 3D characters work if the image shows a clear face and body. Expect more variation between runs.
  • A different clip. Pick one where the two performers start on opposite sides and stay mostly upright. Heavy crossing, crowd shots and fast cuts all make the anchoring harder.

Keep it to two. The prompt is written for a two-person composition. Adding a third character means rewriting the assignment paragraph and uploading a third image, and consistency drops with each extra person.

Troubleshooting

If something goes wrong.

ProblemWhat to try
The characters switched sidesCheck the upload order. Image 1 must be the performer who starts on the left. Swap the images and run again; leave the prompt alone.
A face drifts or changes partway throughReplace that reference image with a clearer, front-on, well-lit photo of one person. Trim the clip shorter for the retest.
Wrong outfit or body shapeThe image is probably cropped. Use a full-body photo when the clip shows full bodies, and make sure the outfit is fully visible.
Hands or mic look wrongRegenerate. Hands near the face are the hardest part of any clip. A shorter section that avoids the worst gesture is a fair trade.
The audio is missing or differentMake sure the prompt still ends with "Preserve the original audio." If it does, regenerate; if it persists, re-attach the original track in your editor.
The credit cost is too highTrim the clip before uploading. Test a few seconds first, then run the section you actually want to post.
Keep it

Take the guide with you.

Hotel Lobby character swap guideMarkdown file. Every step, the prompt and the checklists. Keep it next to your reference images.

Where to go.

Higgsfield's labels, modes and pricing change. If your screen looks different from this page, follow the app over the page.

Want the full library of systems like this?

This is one build. AI Systems Lab is where the rest live: practical AI systems you can run in your own business.