The trend

Hotel Lobby AI: the trend, explained

Two faces, one orange booth, one verse each. Here is where the Hotel Lobby AI video came from, why it keeps coming back, and which duo ideas land best.

Make my video How to make one

Free 5-second preview · No watermark · About 3 minutes · Phone or desktop

Origin

Where the trend comes from

One 2022 performance, one very specific camera angle, and a beat that turned out to be endlessly reusable.

The clip everything copies is Quavo and Takeoff performing “Hotel Lobby” on A COLORS SHOW in 2022. The setup is almost comically simple: two performers, one warm orange backdrop, a handheld camera that stays close, and a verse each. No dancers, no cuts, no crowd.

That simplicity is exactly why it works as a template. Because the camera barely moves and the two performers stay in their own halves of the frame, a face can be swapped on each side without the shot falling apart. The composition does the hard work for you.

The clip went viral a second time in September 2026, when people started re-rendering it with new faces. It moved in waves: TV characters first, then two cats, then the entire internet — coworkers, couples, grandparents, and people putting themselves in both seats at once.

Timeline

How it spread

Four phases, in the order they happened.

When	What happened	What it proved
2022	Quavo and Takeoff perform “Hotel Lobby” on A COLORS SHOW	The shot is stable enough to swap faces into
September 2026	The clip resurfaces; early re-renders use TV and film characters	The format is legible with any two faces
Late September 2026	Two cats become the breakout version	You do not need humans, or even two people who exist together
Now	Friends, couples, coworkers, grandparents, and the same person twice	It is a repeatable, personalised format — not a one-off joke

Ideas

Duo ideas that actually land

The trend rewards a pairing people recognise instantly. These are the ones that get shared most.

Best friends

The original energy. Two friends trading verses about the trouble they get into together. Works best when the two photos are years apart — the contrast is the joke.

Same person, twice

Two photos of yourself in different outfits, and you take both sides of the booth. The model keeps the face identical on each side and gives each one its own movement.

Pets and cats

The version that broke the trend. One clear photo of each animal, taken from the front. Cats that are sitting still and looking at the camera convert best.

Coworkers

Two coworkers who carry the whole office and finally say it out loud. The funniest ones name the specific thing everyone already complains about.

Birthday shout-outs

Send it to the person on the right on their birthday. Type the shout-out into the topic field and the verse is written around their name.

Friendly roasts

A playful roast of each other's worst habits. Keep it affectionate — a roast that reads as mean does not get reposted.

Craft

What makes it read as the trend, not a face swap

Four things the convincing versions all have in common.

Both people move. The giveaway of a cheap edit is a frozen second face. In the versions that look real, both performers get the full choreography and trade verses independently.
The camera stays put. The reference shot is built into the template, so you never have to find or trim the original clip. The framing that makes the swap work is already there.
The booth stays orange. The colour of the backdrop is half the recognition. You can change the setting — studio, rooftop, neon club, birthday, Halloween — and the trend still reads, because the staging logic is the same.
The audio is the real thing. A version with a mismatched beat is instantly recognisable as fake. Every render here ships with the performance audio already attached.

The one rule that fixes most results

Use two separate photos, one person each. The generator binds photo one to the performer on the viewer's left and photo two to the right. A single photo of both people forces it to guess who is who, and it guesses wrong often.

Each photo should be front-facing, with the face clearly visible, no sunglasses and no heavy filters. Include as much of the body as you can — the more it knows about the shoulders, the better the movement.

Numbers

The short version

Everything above, as figures.

Question	Answer
Photos needed	2 — one person each, taken separately
Time to a finished video	About 3 minutes
What the free preview covers	10 credits = one 5-second 480p video
A full 15-second Standard video	60 credits
Cheapest plan	$1.99/month for 1 full videos
Templates	9 settings, including the original booth
If a render fails	Credits are refunded automatically

FAQ

Hotel Lobby AI questions

What people ask about the trend itself.

Is Hotel Lobby AI the same as the original video?

No. The original is Quavo and Takeoff's 2022 performance of “Hotel Lobby” on A COLORS SHOW. This is an independent tool that recreates that format with your own photos.

Do I need the original clip to make one?

No. The reference performance and the camera framing are built into the template. You upload two photos and nothing else — there is no clip to find, download or trim.

Can I use two photos that were taken years apart?

Yes, and it often looks better. The two photos do not need to be from the same day, the same place or the same camera. They just need one clearly visible face each.

Does it work with photos of two different people?

Yes — friends, couples, siblings, coworkers, grandparents. It also works with the same person twice, or with pets. The only requirement is one face per photo.

Why do so many versions use cats?

Because the format needs two clear faces and nothing else. Cats are easy to photograph front-on, and the contrast between an animal and a rap performance is funny on its own.