Hotel Lobby AI Prompt

This is the exact Hotel Lobby AI prompt LobbyVideo sends to the video model for the Orange Booth stage, plus the Red Lobby and Pet Duo variants. It is written for reference-image video models, where you attach two photos (Image1 for the left person, Image2 for the right) and the prompt describes the scene. On LobbyVideo you never type it: you upload two photos and the prompt is applied for you.

The Orange Booth prompt

A 10-second music performance video. Seamless matte-orange wall and floor, one black microphone hanging from the ceiling between two performers. LEFT: the person in Image1. RIGHT: the person in Image2. Both shown full body, side by side, in one centered continuous shot with a fully static camera. LEFT stays on the left half and RIGHT stays on the right half for the whole clip; neither crosses the center. LEFT raps into the hanging mic with confident hand gestures while RIGHT nods to the beat and waits; halfway through they switch roles. Keep each person's face, hairstyle, skin tone, clothing and relative height exactly as in their reference image. No zoom, pan, cuts, flash transitions, furniture or moving background. Soft even front lighting. Original hip-hop beat, no existing song, no lyrics from real artists. No text, no logos, no watermark.

Why the prompt is written this way

The set comes first

“Seamless matte-orange wall and floor” and “one black microphone hanging from the ceiling” are the two things that make the clip read as the 2022 COLORS performance. Putting them first keeps the model from inventing a studio with furniture, posters or a second mic.

Fixed left and right roles

Labelling the people LEFT and RIGHT and tying them to Image1 and Image2 stops the model from blending the two faces. In our first test render, without the line “neither crosses the center”, the right performer walked into the middle in the last two seconds and pushed the other out of frame. Adding that line fixed it.

A static camera

“One centered continuous shot with a fully static camera” plus “No zoom, pan, cuts, flash transitions” matches the original single-angle look and avoids cuts that swap faces mid-clip.

Turn-taking

“LEFT raps … while RIGHT nods … halfway through they switch roles” gives the clip its back-and-forth. Without it, many models make both people rap at once.

Identity lock

“Keep each person’s face, hairstyle, skin tone, clothing and relative height exactly as in their reference image” is what keeps your friend looking like your friend.

No real music

“Original hip-hop beat, no existing song, no lyrics from real artists” avoids copying the actual track. Add the real Hotel Lobby song as the sound in TikTok or Instagram when you post.

Red Lobby variant

Same cast and rules, but in an actual hotel hallway:

A 10-second music performance video in a red-toned hotel lobby hallway with warm wall sconces and a red carpet, one black microphone hanging from the ceiling between two performers. LEFT: the person in Image1. RIGHT: the person in Image2. Both shown full body, side by side, in one centered continuous shot with a fully static camera. LEFT stays on the left half and RIGHT stays on the right half for the whole clip; neither crosses the center. LEFT raps into the hanging mic with confident hand gestures while RIGHT nods to the beat and waits; halfway through they switch roles. Keep each person's face, hairstyle, skin tone, clothing and relative height exactly as in their reference image. No zoom, pan, cuts, flash transitions, furniture or moving background. Soft even front lighting. Original hip-hop beat, no existing song, no lyrics from real artists. No text, no logos, no watermark.

Pet Duo variant

For cats and dogs the cast section changes: animals stand on their hind legs, bob instead of rap, and keep their breed, markings and collar.

A 10-second music performance video. Seamless matte-orange wall and floor, one black microphone hanging from the ceiling between two performers. LEFT: the animal in Image1. RIGHT: the animal in Image2. Both shown full body, side by side, in one centered continuous shot with a fully static camera. LEFT stays on the left half and RIGHT stays on the right half for the whole clip; neither crosses the center. Both stand upright on their hind legs; LEFT leans into the hanging mic and bobs to the beat while RIGHT sways and waits; halfway through they switch. Preserve each animal’s breed, face, fur color and markings, collar and accessories exactly as in its reference image. No zoom, pan, cuts, flash transitions, furniture or moving background. Soft even front lighting. Original hip-hop beat, no existing song, no lyrics from real artists. No text, no logos, no watermark.

More pet tips are on Hotel Lobby AI with cats and dogs.

Using the prompt in other tools

• Model. We use Wan 3.0 (through KIE AI) in reference-image mode, 480p or 720p, 10 or 15 seconds, 9:16 or 16:9. Other reference-to-video models accept similar prompts.
• Image references. If your tool names images differently (for example @image1), change Image1 and Image2 to match.
• Text-only tools. If a tool cannot take reference photos, the prompt still produces the booth, but with random people. The trend needs your own faces, so reference images matter.
• Length. Replace “10-second” with the duration you request so the turn-taking fits.

What not to put in a Hotel Lobby AI prompt

• Real names. Asking for “Quavo” or “Takeoff” produces look-alikes of real people, which most tools (including ours) do not allow.
• Lyrics. Copying real lyrics invites copyright trouble and the model rarely lip-syncs them well anyway.
• Extra props. Chairs, logos or a band in the background break the look.

Skip the prompt

Writing and testing prompts costs time and credits. On LobbyVideo the prompt above is built in and tested. Upload two photos, pick a stage, and a 10-second video costs 10 credits ($2.50–$4.95). Make a Hotel Lobby AI video.