LiveFaceSwap.co
← Back to the blog

Published 2026-11-07 · Updated 2026-11-07 · LiveFaceSwap.co

The hotel lobby AI character swap trend, explained

The hotel lobby AI trend recreates a viral rap performance with new faces. How those clips are made, where live character swap fits, and the consent rules.

The "hotel lobby" AI trend is a wave of short videos that recreate Quavo and Takeoff's performance of "HOTEL LOBBY (Unc & Phew)" on A COLORS SHOW with different people, friends or fictional characters in their places. The versions covered in press and tutorials are usually made with AI video models from uploaded reference photos, not with a live camera face swap. That distinction matters for expectations, for the rights involved and for what our browser studio can and cannot do.

Where the trend came from

The original clip is a live performance on A COLORS SHOW: two rappers, a single hanging microphone and the show's burnt-orange backdrop. The trend takes that recognizable performance and puts new performers into it. Coverage in October 2026 described the videos as AI-generated recreations of the performance, and by then the format had spread to friends, celebrities and characters.

Creators settled on one section of the performance as the loop. A published tutorial for the trend uses roughly the 15 seconds that begin around 0:14.6 of the original, because that part holds the gestures people recognize: the forward point, the forearm across the face, the turn away and the shush.

How the clips are actually made

There are two production routes, and both are rendered rather than live.

Motion-reference recreation. You build a character sheet for each person from clear photos — front, three-quarter, profile and full-body views — then feed an AI video model a simplified motion reference of the performance and ask it to generate a photoreal version with your people. One public tutorial, for example, uses a line-drawing version of the motion clip with Seedance 2.5 in Kapwing and reports about 415 Kapwing credits for a 15-second generation. The original audio is added back in an editor.

Direct face replacement. You take the original clip, add a face image, and render a face swap over the footage frame by frame. This is an uploaded-video workflow: you wait for the file, review it, and re-render if something drifts.

Neither route uses a webcam. The clips are short, edited and usually paired with the original song, which is why they raise questions that a live effect does not.

Character swap, face swap and try-on

The trend is usually described as a "character swap", but three different things get mixed together:

  • Face swap changes identity: the face in the frame becomes the face from your reference image.
  • Restyle or character swap changes the look of the whole frame — the style, the setting, the character feel — rather than only the identity.
  • Try-on changes what you are wearing.

Our studio runs all three as live modes on LiveFaceSwap V1: Face Swap, Restyle and Try-On. What it does not do is rebuild a choreographed two-person performance with full body motion and a specific studio backdrop. That is AI video generation, not a live effect.

Social clips also use "character swap" for a different product: free uploaded-photo tools that replace one character in an image with another while keeping the scene, such as Live3D's AI Character Swap (file limits are published on the tool's page). Those are still-image edits. The live version of a character look is what Restyle and Try-On do on your camera.

What a live character swap can and cannot do

A live swap is worth understanding as a camera effect, not a video editor.

It can: give you a different face or character look on camera in real time, follow your expressions and head movement, let you talk to chat or a call as that character, and be captured with OBS or a screen share for a stream or recording. It is useful when the character is the point of the session and you want to react live.

It cannot: reproduce a choreographed dance, replace your whole body, move you into a burnt-orange studio, sync the original song, or produce a finished 15-second clip for you. The output is a live preview, not a file. We do not record the camera stream, so if you need a video you have to capture it yourself. There is also no 3D avatar output and no system virtual camera yet — the desktop app and virtual camera are coming soon, and today the browser output is captured from the window.

The consent problem the trend exposes

The trend looks harmless until you look at who is being recreated. Takeoff died in November 2022, so a viral format built on his performance keeps regenerating the likeness and movements of someone who cannot decide how that image is used. Some fans criticized those recreations. Coverage of the trend connects it to the wider questions of consent, copyright and the use of people's photographs by AI platforms.

Those questions apply to every tool in this space, including ours, so the rules are simple:

  • Use faces, outfits and images you own or have clear permission to use. A photo being public is not permission.
  • Do not use real people without consent — and not deceased performers, celebrities or public figures at all. Our acceptable use policy also bans impersonation, deception, fraud, harassment, sexual content, depictions of minors and attempts to bypass identity checks.
  • You must be 18 or older to use LiveFaceSwap.co.
  • Disclose AI-altered video when anyone could reasonably mistake it for real footage, on a stream, a call or a recording.
  • Check music and footage rights before you publish anything that uses the original performance or song, and follow the rules of whichever platform or AI service you use.

If you want to join the trend for fun, the safe version is to put yourself, a consenting friend or a fictional character in the clip — not a real person who has not agreed to it, and not someone who is no longer around to say no.

If you want the live version instead

The clips on social media are rendered after the fact. If what you actually want is to be a character while you talk to people live, the browser studio does that with Face Swap, Restyle and Try-On: 0.6 credits per live second, sessions up to 900 seconds with the unused reservation returned when you stop, 10 free credits for a new account, and one-time packs from $10 for 1,000 credits that never expire. Quarterly subscriptions start at $10/month for 650 credits; the pricing page lists every plan and pack.

Disclose the effect when it matters, respect the likeness rules above, and do not present the output as real footage. The studio produces a live preview; the editing and the upload are yours.

Sources

Where to go next