Genjutsu AI: Genjutsu-Style Viral Motion, Starring You
The Hotel Lobby clip's cuts, timing and moves play out with you in it, rapping as your own twin under one hanging mic. One photo of you or your pet is all it takes.
Updated : Page published with the Hotel Lobby motion template; the STORM II crowd version lives on the Storm Trend AI page.
1Add your photo
Front-facing, well-lit selfies work best. One face per photo.
2Choose quality
By generating you confirm you have permission to use the people in your photos. Terms
In one sentence
Genjutsu-style motion transfer is an AI video that keeps the motion, cuts and timing of a viral clip and swaps in a new character, as in the Hotel Lobby and STORM II remakes; videosi makes a 15-second Hotel Lobby version from one photo of you.
- Fast facts
- First seen
- Stage
- Just starting
- Started on
- Higgsfield
- Sound
- No single sound: Genjutsu is a tool, and each remake uses the sound of the clip it copies, such as Hotel Lobby (Quavo & Takeoff) or STORM II (GENER8ION and Yung Lean).
- Meaning
- "Genjutsu AI" refers to Higgsfield's Genjutsu video-to-video tool, which keeps a clip's motion, camera and timing and swaps in new characters and places, and to the viral remakes made that way, such as Hotel Lobby and STORM II.
Page updates
- October 26, 2026 — Page published with the Hotel Lobby motion template; the STORM II crowd version lives on the Storm Trend AI page.
How it is made
The Genjutsu AI prompt and model
Genjutsu AI is made with Seedance 2.0 / 2.5 (you pick the tier) by ByteDance. The prompt behind it has 5 parts, and the timing is what makes or breaks it.
- Model
- Seedance 2.0 / 2.5 (you pick the tier)
- Mode
- Photo + motion reference to video
- Length
- 15 seconds, 9:16, 1:1 or 16:9
- Input
- 1 photo
- Identity lock Your photo is the only source for both performers' face, hair, build and outfit, and a pet stays the same animal.
- Motion reference A 15-second Hotel Lobby remake supplies the cuts, framing and back-and-forth gestures, and nothing else.
- Scene A flat orange studio, a low stage block and one silver mic hanging between the two performers.
- Timeline Five beats over 15 seconds, placed where the original clip cuts.
- Rules Exactly two performers, mouths moving in rhythm, no merged bodies, no children, no celebrity faces, no text or logos.
Photorealistic live-action music-performance footage, 15 seconds, studio session look, soft even key light, no camera shake. There are exactly two performers…
Why it goes wrong when you do it yourself
- Without a motion reference, a general video model invents its own moves and the clip no longer reads as the viral original.
- A motion reference also carries faces and clothes, so unless the prompt separates motion from identity the result drifts toward the people in the reference clip.
- With two copies of one face on screen, models tend to drift one twin toward a different person unless the prompt locks both to the same photo.
- Each try costs credits and several minutes, and landing the cut and the drop on the right second can take more than one try.
Genjutsu AI examples
Two cats as a rap duo
Shiba and Rottweiler duo
Owner and dog duo
How to make your Genjutsu AI video
Upload one photo
A sharp, front-facing photo of yourself, waist up or more, or a clear full-body photo of your pet.
Pick the format
9:16 for TikTok, Reels and Shorts, 1:1 or 16:9. The Hotel Lobby motion, studio and prompt are built in.
Generate and download
Press Generate. The 15-second clip renders in a few minutes as a silent MP4 with no watermark.
Add the sound and post
The video is silent. Add the Hotel Lobby sound from your app's sound library when you post, and label it as AI-generated.
What is the Genjutsu AI trend?
Updated · videosi is an independent tool. It is not affiliated with TikTok, Instagram, YouTube, or with the owners of any trend, song, game or brand named on its pages.
"Genjutsu AI" usually means Genjutsu, a video-to-video tool from Higgsfield, and the remakes people make with it. The idea is motion transfer: keep a clip's movement, camera and timing, and rebuild who is in it and where from reference images. Higgsfield's own launch post is dated 31 August 2026 and describes two modes, Motion Transfer and Object Swap. The word itself is borrowed from Naruto, where a genjutsu is an illusion.
The best-known results copy a handful of clips. In Hotel Lobby remakes, two cats, two dogs, a couple or two friends rap under one hanging microphone in a flat orange studio, after Quavo and Takeoff's 2022 COLORS performance. In STORM II remakes, a person walks into a silent crowd and stands still while everyone around them breaks into the dance from the GENER8ION and Yung Lean music video. Both waves spread on TikTok from September 2026.
videosi is not Genjutsu and is not affiliated with Higgsfield. This page does the same kind of motion transfer with ByteDance's Seedance model on the most-copied clip, Hotel Lobby: you upload one photo and get a 15-second silent video in which you (or your pet) play both rappers as twins, following the original cuts, timing and gestures.
Want the STORM II crowd instead? The Storm Trend AI page puts you in the middle of the crowd from one photo. For the Hotel Lobby duo with two different people or pets, use the Hotel Lobby AI page.
Genjutsu AI: shot by shot
- 0–2 s
Wide on the stage
You and your twin stand on a low stage facing each other under the hanging mic, orange wall all around.
- 2–5 s
Two-shot, first bars
Cut to waist up. The right twin leans in and points at the lens while the left one nods.
- 5–9 s
Hands on the beat
Open-hand gestures and a bounce on every beat, the other twin swaying at the mic.
- 9–13 s
Turn and take over
The right twin turns away, hand behind the head, and spins back; the left one leans in to the mic.
- 13–15 s
Face to face
Both trade gestures under the mic, ending with a finger to the lips.
Tips for a better result
Show your outfit
The outfit in your photo is the outfit in the video, so a waist-up or full-body photo works better than a face crop.
Keep the face clear
No sunglasses, masks or hats, and no other people in the photo. Both twins copy this one face, so sharpness matters.
Dress for the orange
The studio is flat warm orange. A white, black or blue outfit stands out; an orange top blends into the wall.
Pets: full body
For a pet, use a photo that shows the whole animal, so the model knows its legs, tail and markings when it stands up to rap.
Good to know before you make a Genjutsu AI video
- videosi is not Genjutsu and is not affiliated with Higgsfield. You cannot upload your own source video here: the motion is the built-in Hotel Lobby clip.
- Both rappers come from your one photo, so you appear twice as twins. For two different people or pets, use the Hotel Lobby AI page; for the STORM II crowd, use the Storm Trend AI page.
- It is a performance, not an exact lip-sync of the song, and likeness is strongest in the waist-up shots.
- The video is silent and likeness is not guaranteed. A failed, timed-out or blocked render returns its credits automatically.
Genjutsu AI: questions
What is Higgsfield Genjutsu?
Genjutsu is a video-to-video AI tool from Higgsfield, launched with a post dated 31 August 2026, that keeps a clip's motion, camera and timing and rebuilds the cast or setting from reference images. videosi is a separate tool, not Genjutsu: it does Genjutsu-style motion transfer from one photo onto the 15-second Hotel Lobby motion.
Is there a Genjutsu AI alternative?
Yes. videosi is an independent Genjutsu AI alternative for the viral remakes: upload one photo and get a 15-second Hotel Lobby video of you following the original motion, and the STORM II crowd version is on videosi's Storm Trend AI page. It runs ByteDance's Seedance with the motion reference built in. videosi is not affiliated with Higgsfield.
How do I make a Genjutsu-style video from one photo?
On videosi, upload one photo of yourself or your pet, choose 9:16, 1:1 or 16:9 and press Generate. The 15-second clip keeps the Hotel Lobby cuts, timing and gestures and puts you in them as both rappers. You write no prompt, and the finished MP4 is saved in My creations. For the STORM II crowd, use the Storm Trend AI page.
Is Genjutsu AI free?
The Genjutsu-style videos on videosi are not free: there is no free generation or trial. A 15-second video on the Standard model at 720p is 13 credits; better models and higher resolutions cost more. Credit packs start at $6.99 for 10 credits and never expire, plans start at $29.90 a month, and a failed render returns its credits.
How much does Higgsfield Genjutsu cost compared with videosi?
Higgsfield lists its own credit prices for Genjutsu on its site, and they depend on length and resolution. On videosi, a Genjutsu-style 15-second video is 13 credits on the Standard model at 720p, and credit packs start at $6.99 for 10 credits. videosi prices come only from videosi's pricing page and are not linked to Higgsfield's.
What AI model and prompt does videosi use for Genjutsu-style motion transfer?
videosi uses ByteDance's Seedance model (Seedance 2.0 Mini by default; Fast, Pro and Seedance 2.5 Ultra are one click away) with a five-part prompt and a 15-second Hotel Lobby motion reference clip. We do not publish the full prompt. It locks both performers to your photo and takes only motion and camera from the reference.
What is the Hotel Lobby AI trend?
The Hotel Lobby AI trend, which spread on TikTok from September 2026, recreates Quavo and Takeoff's 2022 COLORS performance with new characters, often cats, dogs, couples or friends, rapping under a hanging mic in an orange studio. On videosi's Genjutsu AI page you play both rappers as twins in 15 seconds; the Hotel Lobby AI page does it with two different people or pets.
Trending this week
All trends this week
More effects like this
All viral effects
Go deeper on Genjutsu AI
Storm Trend AI ↗
The STORM II crowd version of Genjutsu-style motion transfer: you stand still in the middle while the crowd breaks into the dance, with a choice of blazers, nurses or firefighters.
Hotel Lobby AI ↗
The Hotel Lobby video with two different people, a couple, or two pets trading bars under the mic.
Make your Genjutsu AI video
Upload a photo and press Generate. 13 credits, refunded if it fails.
Get the next trend first
One email when new trends land on videosi, about once a week. Unsubscribe in one click.
See videosi first on Google
Add videosi as a preferred source and our trend pages show up first in your Top Stories and AI answers on Google.