Updated 2026-10-07
How to Make a Hotel Lobby AI Video
The Hotel Lobby AI format draws on the orange-booth duet from Quavo and Takeoff’s COLORS performance. This site creates an original AI performance with the two people you choose. Here is the whole process, from picking photos to posting on TikTok. Preparing the photos takes a few minutes; rendering can take longer when the provider’s queue is busy.
Step 1 — Choose two photos
Pick one photo per performer. Front-facing, well lit, both eyes visible. If you want the outfit in the video, use a waist-up or full-body shot. The photos do not need to be taken together: friends in different cities, a grandparent from an old album, a coworker’s LinkedIn headshot (with permission) all work. Decide who is on the left (Photo 1) and who is on the right (Photo 2) — the preset asks the left performer to lead first, although timing can vary.
Avoid: group photos, sunglasses, heavy beauty filters, profiles, and dark bar lighting. These are the reasons faces blend or swap sides.
Step 2 — Upload and generate
Open the Hotel Lobby AI generator, add Photo 1 and Photo 2, pass the quick human check and press Generate. Sign in with Google when asked — that is how we give every account one free video. The booth (matte orange walls, one hanging mic, locked-off wide shot) and the alternating performance are preset, so there is nothing to type.
Allow several minutes for rendering; busy queues can take longer. You can keep the page open or come back later; the finished clip is saved on your videos page.
Step 3 — Download and post
Download the MP4. In TikTok, Reels or Shorts, upload it as a new post and keep the generated audio or select music from the platform’s sound library according to its usage rules. Trim to the moment both performers are in the frame, add a caption like “we did the hotel lobby ai trend”, and post. Most viral versions use one of three hooks: an unexpected pairing (grandma and the dog), a “before/after” cut from the photos to the booth, or a duet with the original clip.
Tips that raise the hit rate
- Match the energy of the booth. Photos where the person is already smiling or mid-gesture read better than passport-style shots.
- Keep the two photos similar in scale. One tight face crop plus one full-body shot gives uneven results.
- Try the reverse order. If the left performer looks off, swap the photos — the first slot carries slightly more motion.
- Same person twice works. Upload two different photos of yourself for the “me vs. me” version.
- Pets: eyes open, ears up. A clear, frontal pet photo is one of the most shared variants.
Other ways to do it
If you prefer to do it manually, the Hotel Lobby AI prompts page has six prompts for Kling, Veo and Seedance and explains why prompting alone tends to blend the two faces. If you just want the fastest possible version for a joke, a CapCut template pastes your face onto the original performers in under a minute. For the background of the trend — the 2022 performance, the cat video, the celebrity versions — see the Hotel Lobby AI trend.
FAQ
Do I need the original song?
No. This site generates a new performance and audio. You can optionally use music from your publishing platform’s sound library according to its rules.
Can I make it on my phone?
Yes. The generator is a web page; upload from your camera roll, allow several minutes for rendering, download the MP4 to your phone.
Why did my render fail?
Poor references or a rendering-provider error can cause failures. A photo with several faces, very low light or a side view can also produce weak results. Retry with a front-facing, well-lit photo — a failed render does not use your free video.