Hotel Lobby AI video starring you two

In September and October 2026, TikTok and X filled up with "Hotel Lobby" AI videos. Two people stand side by side in a bright orange booth and trade rap lines into one microphone hanging between them.

The look comes from the COLORS show performance of "HOTEL LOBBY" by Quavo and Takeoff, published on June 17, 2022: an orange backdrop and one microphone hanging between them. In September 2026 people started putting their own faces in that frame with AI video tools, and Quavo reposted the original on X on September 23, 2026 (XXL).

The full story, with sources and caption ideas, is in our guide to the Hotel Lobby trend.

Kiravu is not affiliated with the artists or with COLORS, and it does not use the original song or lyrics. Kiravu's orange booth stage recreates the look with your faces and an original verse written for the two of you.

The trend has a few fixed parts. Here is how the orange booth stage handles each one:

• The color. A saturated orange backdrop with soft, even light and no visible edges.
• The mic. The original hangs one mic between them. Here each performer gets a vintage mic on a floor stand at chest height, so both faces stay in view.
• The framing. Waist-up, shoulder to shoulder, one locked-off shot with no cuts and no zoom.
• The trade. Person 1 takes the first verse, person 2 answers, and both hit the last words together.

Trap (hard 808s, around 140 BPM) is the default beat because it matches the energy of the clips going around. Boom bap is slower and dustier, so each line lands on its own. Melodic turns the booth into a sung duet.

The booth is bright and plain, so your faces carry the whole video. A few things help the render:

• Use one clear, front-facing selfie per person, taken in good light.
• Keep your shoulders in the shot. The video frames you from the waist up.
• One face per photo. Crop out friends standing behind you.
• Skip sunglasses and heavy beauty filters so your eyes and mouth are easy to read.

JPG, PNG and WEBP files up to 10 MB all work. Photos are used only to render your video and are deleted from storage after 7 days.

To shoot the trend yourself you'd need an orange backdrop, a hanging mic, even studio light, a beat, lyrics and two people who can rap in time. Then a few dozen takes.

Here you send two photos, tweak the words if you want, and the render handles the performance, the beat and the lip-sync. It runs in the browser, so there is no app to install and nothing to edit afterwards.

A standard song (12 seconds at 720p on Standard) costs 10 credits; 480p is 5. Faster or sharper models, 1080p and longer clips cost more. If a render fails, the credits come back automatically.