How to Create a Hotel Lobby AI Video (Steps, Prompts & FAQ)
Step-by-step guide to making a Hotel Lobby AI video: the photos to use, the exact scene and motion prompts, settings, fixes, and how to generate one in one click.
The Hotel Lobby trend is everywhere on TikTok and Reels: two people stand in a bright orange booth, one microphone hangs between them, and they trade verses like a rap duo in a single still, full-body shot. You don't need a studio, an orange backdrop or a camera crew to make one. With two ordinary photos and an AI video model, you can create a Hotel Lobby AI video in a few minutes.
This guide walks through exactly how we make Hotel Lobby AI videos at ViralEffects: the steps, the prompts we use, the settings that matter, and the mistakes to avoid. At the end, you'll see how to do the whole thing in one click.
What makes a Hotel Lobby video recognizable
Before writing a single prompt, it helps to know which details make viewers instantly recognize the format. Get these right and the video reads as "Hotel Lobby" even without the sound:
- A seamless orange booth. The back wall, side walls and floor are one saturated orange with no visible corners or seams.
- One hanging microphone. A single black studio mic on a thin cable, centered between the two people at mouth height.
- Two people, full body. Both performers are visible from head to shoes, standing side by side about an arm's length apart.
- A locked-off camera. No zoom, no pans, no cuts. One continuous wide shot.
- Trading verses. One person raps with big hand gestures while the other nods along, then they swap.
- Vertical 9:16. It's made for TikTok, Reels and Shorts.
The format takes its name from the song "Hotel Lobby" by Quavo and Takeoff, which is the sound most creators pair with it.
The two-stage workflow
The most reliable way to create a Hotel Lobby AI video is to split the job in two:
- Compose a still image of both people inside the orange booth with an image editing model.
- Animate that still into a 10-second performance with an image-to-video model.
Why not go straight from your photos to video? Image-to-video models animate the first frame they're given. If the first frame is a photo of you in your kitchen, the model has to invent the booth, the mic and the second person mid-shot, and the result drifts. Building the scene first gives the video model a clean, correct starting frame, so all it has to do is make the people perform.
At ViralEffects we use Nano Banana (image edit) for the still and Kling 2.6 Pro (image-to-video) for the animation.
Step 1: Pick the right two photos
Photo quality decides most of the result. Choose one photo per person:
- Full body, head to shoes. The format is a full-body shot, so the model needs to see legs, shoes and outfit.
- Facing the camera with the face clearly visible and well lit.
- Sharp and evenly lit. Avoid heavy shadows, backlight and motion blur.
- One person per photo. Group shots confuse the model about who should appear.
Avoid sunglasses, hats that hide the face, and tight face close-ups. Close-ups force the model to invent the body, and faces tend to drift toward a generic look. And only use photos of people who have agreed to be in the video.
Decide who stands where: put the photo of the left-hand person first.
Step 2: Compose the orange booth still
Place the two photos side by side in a single image (left person on the left half, right person on the right half) and send it to the image editing model with a prompt like this one, which is the exact prompt we use:
Create one photorealistic photo of the two people shown in this image. The person on the left half of the image stands on the left, and the person on the right half stands on the right. Keep each person's face, hair, skin tone, body and outfit exactly as they are. They stand side by side, about an arm's length apart, both fully visible from head to shoes and facing the camera, inside a small seamless studio booth: the back wall, side walls and floor are one solid saturated orange color with no seams, corners, props, logos or text. A single black studio microphone hangs on a thin cable from above, exactly between the two people at mouth height. Soft, even frontal lighting. Locked-off wide shot of a live music performance session. Vertical 9:16 frame.
Why each part matters:
- "Keep each person's face, hair, skin tone, body and outfit exactly as they are" protects identity. Without it, image models love to "improve" outfits and faces.
- "No seams, corners, props, logos or text" keeps the booth clean. Otherwise you'll get random posters, chairs or a visible room corner.
- "Exactly between the two people at mouth height" places the mic. Leave it vague and it ends up above someone's head or in someone's hand.
- "Vertical 9:16 frame" sets the shape early, so nothing gets cropped later.
Check the still before moving on. Both faces should look like the source photos, both people should be fully in frame, and there should be exactly one mic. If anything is off, regenerate now. It's much cheaper than fixing it in video.
Step 3: Animate the performance
Send the still to the image-to-video model with a motion prompt. Here's ours:
Static locked-off camera, one continuous shot, no cuts, no zoom. The two people perform a rap duet into the single hanging microphone between them, trading verses: the person on the left raps with energetic hand gestures while the person on the right nods and bounces to the beat, then the person on the right leans in and raps while the person on the left vibes along. Natural lip movement, head bobbing, confident live performance energy. Both stay in their places and fully visible. The seamless orange booth and the microphone stay unchanged.
And a negative prompt to block the usual failure modes:
camera movement, zoom, cuts, extra people, extra microphones, distorted hands, blurry faces, text, watermark
Settings we use:
| Setting | Value | Why |
|---|---|---|
| Duration | 10 seconds | Enough for both people to take a turn |
| Aspect ratio | 9:16 | Native for TikTok, Reels and Shorts |
| Audio | Off | You'll add the real trending sound when posting |
The key idea in the motion prompt is turn-taking. Describing "the left person raps while the right person vibes, then they swap" gives the video a clear structure. If you just write "two people rap," both tend to move their mouths at the same time and it looks like noise.
Step 4: Add the sound and post
The generated video is silent on purpose. Upload it to TikTok, Instagram Reels or YouTube Shorts and pick the trending Hotel Lobby sound from the app's music library. Using the platform's licensed sound keeps you clear of music rights problems, and videos that use a trending sound get grouped with the trend, which helps people find yours.
A few posting tips:
- Trim so the first verse lands right on a strong beat.
- Put a short hook in the caption ("we finally did the Hotel Lobby trend").
- Tag the friend who's in the video so it shows up on both profiles.
Prompt variations to try
Once the basic version works, small prompt changes give you a fresh take:
- Different energy: swap "energetic hand gestures" for "smooth, laid-back swagger" for a calmer vibe.
- Back-to-back finish: add "for the last two seconds both lean in and rap together into the mic."
- Matching outfits: in the still prompt, replace "outfit exactly as they are" with "both wearing matching black tracksuits" (this changes clothes but keeps faces).
- Different booth color: orange is the signature, but "solid saturated teal" or "hot pink" makes a recognizable remix.
Change one thing at a time so you know what made the difference.
Common problems and fixes
| Problem | Likely cause | Fix |
|---|---|---|
| Faces don't look like the photos | Low-res or close-up source photos | Use sharper, full-body photos facing the camera |
| Hands look warped | Fast gestures are hard for video models | Regenerate, or try "relaxed hand gestures" |
| Camera zooms or pans | Motion prompt too loose | Keep "static locked-off camera, no zoom" at the start and in the negative prompt |
| Two microphones appear | Mic not anchored in the still | Regenerate the still until there is exactly one mic |
| Someone gets cropped | Still wasn't full body | Check the still before animating |
| People swap sides | Photo order | Put the left-hand person's photo first |
How to create a Hotel Lobby AI video in one click on ViralEffects
Everything above, including composing the still, writing the prompts, tuning the settings and checking the frame, is already built into our Hotel Lobby AI video generator. You just bring two photos:
- Open the Hotel Lobby AI effect and sign in.
- Upload two photos, one full-body photo per person. The first photo stands on the left, the second on the right.
- Press Generate. We compose the orange booth scene and animate the 10-second vertical performance for you. It takes a few minutes, and you can leave the page: the finished video is saved in My videos.
- Download, add the trending Hotel Lobby sound in TikTok, Reels or Shorts, and post.
Each Hotel Lobby video costs 37 credits. Credits come in one-time packs that never expire; see pricing. No prompt writing, no model accounts, no editing software.
Create your Hotel Lobby AI video →
FAQ
What is the Hotel Lobby AI trend?
It's a viral short-video format where two people stand in a seamless orange booth with one microphone hanging between them and trade verses like a rap duo, filmed in one still, full-body shot. The AI version lets you put yourself and a friend in that booth from ordinary photos. The name comes from the song "Hotel Lobby" by Quavo and Takeoff.
Can I make a Hotel Lobby video with just one photo?
The format is a duet, so it needs two people. You can upload two different photos of yourself for a "me vs. me" version.
Does the AI video include the song?
No. The video is silent so you can add the official trending sound inside TikTok, Reels or Shorts. That keeps you clear of copyright problems and links your video to the trend.
Which AI models are used to create Hotel Lobby videos?
We use Nano Banana to compose the orange booth still from your two photos, and Kling 2.6 Pro image-to-video to animate the 10-second performance.
How long does it take?
A few minutes. You don't need to keep the page open; the finished video appears in My videos.
Why do the hands or faces sometimes look off?
Fast hand gestures and small or blurry faces are the hardest parts for current video models. Sharp, full-body, front-facing photos fix most issues. If a result still looks wrong, try a different pair of photos.
Can I use photos of other people?
Only with their permission. Don't create videos of people who haven't agreed to it.
What aspect ratio and length is the video?
Vertical 9:16, 10 seconds, ready for TikTok, Instagram Reels and YouTube Shorts.