If your feed has been full of two people squeezed around one hanging microphone against a flat orange wall, you have seen the Hotel Lobby AI trend. This guide covers where the clip comes from, why it spread in September 2026, the different ways people make their own version, and how to get a clean result from two ordinary photos.
The same orange booth is also searched as AI Rap Duo. The category page is AI Rap Duo. Generate on Hotel Lobby. That effect URL stays /effects/hotel-lobby.
What is the Hotel Lobby AI trend?
Despite the name, there is no hotel in it. "Hotel Lobby (Unc & Phew)" is a rap song released on May 20, 2022. Its best-known performance was filmed for A COLORS SHOW, the YouTube music series known for single-color sets, and published in June 2022: a bright orange backdrop, two performers sharing one microphone that hangs from above, trading verses back and forth.
According to Complex, the AI version took off in early September 2026 after a few clips of two cats "performing" the song went viral. Meme.com's write-up traces the earliest known edit to a TikTok posted on September 16 that put two TV characters in the booth. After that the format filled up with pairings: famous couples, film duos, footballers, superheroes, coworkers, pets.
On September 23, 2026, one of the original performers reposted the untouched performance and called it "The Original", which pushed the trend even further.
Sources: Complex, September 27, 2026 (the A COLORS SHOW performance, the release date, the cat clips and the September 23 repost) and the track listing on Spotify (release date May 20, 2022).
Why the format works
A few things make this clip unusually easy to remix:
- One flat color. The orange set has no props, no depth and no background detail, so a model has very little to get wrong.
- One shared mic. The single microphone between two faces reads instantly. Viewers recognize the setup in under a second, whoever is standing there.
- Two performers. The back-and-forth (one person raps, the other reacts, then they switch) gives the joke a built-in structure. Pick any two people who would never share a stage and you have a punchline.
- Short and vertical-friendly. A tight waist-up framing crops cleanly to 9:16 for TikTok, Reels and Shorts.
The part people argue about
It is worth knowing the context before you post. One of the two original performers has since died, and the trend reproduces his performance over and over. Some fans have criticized that, along with broader concerns about consent, copyright and what AI apps do with uploaded photos. Tom's Guide covers both sides.
Most viral versions are face swaps: they keep the original footage, the movements and the song, and paste new faces on top. That approach depends on someone else's video and audio. A cleaner route is to generate a new clip that borrows the look (orange set, one mic, two performers) without copying the footage or the recording. That is how the Hotel Lobby effect works. Face Swap takes one to three character images and a reference video, 150 credits for 15 seconds at 768P.
How to make a Hotel Lobby AI video
There are three common methods.
1. A phone template
CapCut and similar apps have community templates where you drop two photos into slots. It is quick, but you get whatever the template author built, the length is fixed, and many templates reuse the original performance.
2. Face-swap or motion-transfer apps
Tools like Starrd and Summrs, mentioned in the Complex piece, map two faces onto the original clip. The movement matches the real performance closely, because it is the real performance. The trade-off is the reliance on footage and audio you do not own.
3. Generate a new clip from two photos
A reference-to-video model takes your two photos as identity references and builds a new scene from a text description of the set. On HiVideoStudio this is the Hotel Lobby AI video generator:
- Upload photo 1 (the person on the left) and photo 2 (the person on the right).
- Optionally add a short line about outfits or energy, for example "both in matching tracksuits, very hyped".
- Generate. A 15-second vertical video costs 90 credits. It is one pass, not two clips joined together. The sample below is a separate 4-second file, not the video you generate.
The sample below is a 4-second loop from one of our own renders. It uses an AI-generated instrumental beat, not the Hotel Lobby track. If you want the actual song, add it from the TikTok or Instagram sound library when you post, where music is licensed for use on the platform.
A 4-second loop from one of our own renders: two people from two photos, an orange set and one hanging mic. No footage or audio from the original performance is used.
Getting a good result
We ran our test renders with CC0 stock photos, and a few things made a clear difference:
- One person per photo. Group shots confuse the model about who goes where.
- Similar framing. Two waist-up photos give a balanced duo. Mixing a close-up selfie with a full-body shot tends to make one person look oddly scaled.
- Clear, front-facing faces. Sunglasses, hats pulled low and heavy shadows reduce how closely the result matches the photo.
- Keep the prompt short. The set and the mic are already described in the effect. Extra direction works best for clothing and mood.
Expect some variation between renders. Faces and the orange set stay consistent; small things like the exact clothes, the hands and where the camera ends up can change. In our second test render the camera drifted back slightly over four seconds and a thin strip of ceiling appeared at the top. That is typical of current video models, and running a second generation usually fixes framing you do not like.
Mouth movement is not synced to lyrics, because there are no lyrics in the generated audio. If you add the real song in an editor, treat it as a performance vibe rather than a lip-sync.
Ideas that tend to land
- Two coworkers, captioned with the meeting that should have been an email.
- Your dog and your cat, which is how the trend started in the first place.
- You and your younger self, using an old photo.
- A couple, posted for an anniversary.
Using photos of other people
Only upload photos of people who have agreed to it, and do not upload photos of minors without a guardian's consent. Do not present the clip as real footage. These rules are in our terms, and they are also just good practice when the whole point of a format is putting someone in a scene they never filmed.
Quick answers
Is the Hotel Lobby trend filmed in a hotel? No. The orange set is a music-show studio. "Hotel Lobby" is the song title.
Where does the original come from? A June 2022 performance of the song on A COLORS SHOW, filmed against an orange set.
When did the AI trend start? Early-to-mid September 2026 on TikTok, with a repost by one of the original performers on September 23 giving it another push.
How many photos do I need? Two, one per performer.
Is this effect affiliated with the artists or A COLORS SHOW? No. HiVideoStudio is independent and not affiliated with the performers, their label or COLORS.
Ready to try it? Open the Hotel Lobby effect, upload two photos and see what you get.