What is the Hotel Lobby AI trend?

The reference is Quavo & Takeoff’s “HOTEL LOBBY” performance on A COLORS SHOW: two performers, a hanging microphone, and a distinctive orange studio. The AI version puts a different duo into that familiar setup. Despite the name, the scene is a performance space, not a hotel reception area.

The recent wave is more than an old music clip resurfacing. In an October 2026 AP interview with Quavo, the conversation explicitly turns to the performance’s AI-driven resurgence. A likely reason for the format’s appeal is its recognizable composition: viewers immediately understand the scene, while changing the two people supplies the joke. That is an interpretation of its appeal, not proof that one app or post started the trend.

There are two different workflows behind similar-looking results. A video face swap replaces faces in existing footage. A reference-guided video generator creates a new scene using portraits and a template or description. The second can change clothing, movement, and framing as well as faces. Decide which result you want before uploading; a text-only video generator may not preserve either person’s identity reliably.

About this guide

These steps describe the workflow to look for in a compatible tool; button names and features vary by provider. Our homepage currently offers a local photo preview. Video generation and downloading are not yet connected there.

What you’ll need before you start

Prepare two separate portraits, one for each person, and choose a service that explicitly supports two identities in the same video. A tool that can animate one face does not necessarily support a duo. If you are replacing faces in existing footage, you also need a compatible clip that you have permission to use.

Choose clear faces and even light

  • Use a front-facing portrait. Keep both eyes, the nose, and the full mouth visible. Leave some room around the forehead and chin instead of cropping tightly against the face.
  • Look for soft, even lighting. Face a window or another diffuse light source. Avoid a bright window behind the person or strong shadows covering half the face.
  • Start from the original photo. A small screenshot or repeatedly compressed image loses facial detail. Check sharpness at normal viewing size before uploading.
  • Keep each image unambiguous. Use one person per photo, without sunglasses, a hand over the mouth, heavy beauty filters, or a busy crowd behind them.

Avoid the common setup mistakes

Do not assume a larger file fixes a blurry face; upscaling cannot recover the original expression. Match the source angle to the template when possible, and avoid a wide-angle selfie taken extremely close to the nose. Save the portraits as “person-left” and “person-right” so you can check their placement later. Use your own photos or get permission from both people, and read how the chosen service handles uploaded images.

How to make your video, step by step

STEP 01Pick and prepare your two photos

Open both portraits side by side. Check that each face is sharp, unobstructed, and large enough to inspect. Crop away other faces while keeping the entire head visible. Use JPG or PNG if your chosen service accepts them; if a phone photo is HEIC, export a compatible copy rather than merely changing the filename extension.

Decide who belongs on each side of the scene. Keep a second portrait of each person available in case the first image performs poorly. A slightly different angle or more neutral expression is often a more useful next test than repeating the same input.

STEP 02Upload and assign each identity

Open the tool’s video face-swap or two-person reference workflow. Upload the portraits into separate identity slots. If the interface detects faces in a target clip, explicitly assign each portrait to the intended performer. Preview thumbnails can help you spot reversed identities before using generation credits.

For an existing-footage workflow, select the permitted target clip as well. For a template workflow, confirm that both uploaded portraits are used as identity references. Read any upload limits and data-retention information before continuing. A successful upload only confirms that the file was accepted; it does not guarantee a successful likeness.

STEP 03Choose the studio template and style

Start with the simplest orange studio template available: two people, one hanging microphone, and a steady camera. Keep stylization low for the first attempt so you can judge the faces. Choose the output shape before rendering. For a vertical social clip, check that both people stay within the crop instead of assuming a centered crop will fit a wide duo shot.

If the service supports prompts, describe the setting and placement clearly. For example: “Two people side by side in a warm orange recording studio, one suspended microphone between them, steady medium-wide framing, preserve each uploaded person’s identity.” Treat this as direction, not a guarantee of exact movement or lip sync. Our planned Hotel Lobby AI prompts guide will expand on ways to describe the scene.

STEP 04Generate, inspect, and download

Make the shortest available test before committing to a longer or higher-resolution render. Watch the whole result, then pause where a person turns, opens their mouth, or moves a hand across their face. Check that identities remain on the correct bodies, facial edges do not flicker, and the second person stays consistent when the first person moves.

If something fails, change one input at a time: replace a portrait, correct the face assignment, or reduce motion. Once you are satisfied, use the tool’s download option and play the saved file outside its preview player. Confirm the resolution, watermark, and audio are what you expected. A face swap alone does not add music or synchronize new speech. Use audio you are allowed to share and identify the result as AI-generated when posting.

Free vs. paid tools: what should you choose?

Start by checking capabilities, not the price badge. A free trial is useful only if it lets you test the actual two-person workflow. A paid subscription is not evidence that a tool supports the orange studio template, keeps both identities stable, or includes the music you want.

Compare these details before generating
CheckFree optionPaid option
Two identitiesCan you test both faces together?Is multi-face video included in the plan?
ExportCheck download access, watermark, and resolution.Confirm the advertised quality applies to downloaded video.
RetakesCheck the credit allowance and whether previews use it.Check credit cost per attempt and failed-render terms.
Data handlingRead retention and deletion options.Check the same terms; payment alone does not ensure privacy.

For a one-off experiment, test a free allowance before subscribing. If you need repeated exports, compare the total cost of several attempts rather than a headline price per clip. Check whether a trial renews automatically and whether unused credits expire. Our planned free Hotel Lobby AI video guide will cover this decision in more detail. No provider prices or free quotas are assumed in this tutorial.

Frequently asked questions

1. How many photos do I need?

Prepare two portraits: one of each person. Some services support additional references, but requirements vary. Avoid substituting one group photo unless the tool explicitly lets you detect and assign each face separately. Separate portraits make the identity mapping easier to check.

2. Can I make a Hotel Lobby AI video on my phone?

You can use a mobile browser if the chosen generator supports it. Confirm photo-format compatibility and keep enough storage for the downloaded file. If a completed video opens in a browser player, use its download or save option, then check your Files or Downloads app.

3. Why doesn’t my face line up correctly?

First check the assigned identity slot. Then inspect the source portrait for a tilted head, covered features, or extreme close-up distortion. If alignment fails only during turns or when the microphone crosses the face, the target motion may be difficult for that model. Try a clearer portrait or a calmer, shorter clip.

4. Why do the two faces switch places or blend together?

Recheck the left and right assignments and use distinct, separate portraits. If the problem appears only when the performers overlap, try a template with more space between them. Some models struggle to keep multiple identities consistent; another prompt alone may not solve that limitation.

5. Can I generate and download the video for free?

That depends on the service. A free preview may still require payment to download, remove a watermark, or export at higher resolution. Check the final export terms before spending time on a render. Our homepage’s local photo preview is free, but its video generator is not connected yet.

6. Does the video automatically include the song and lip sync?

Do not assume it does. Face replacement, animation, audio, and lip synchronization can be separate features. Check a provider’s actual export options. When adding audio separately, use a track or platform music option you are permitted to use and preview the timing before publishing.

Start with the faces. Build from there.

A clear portrait and a simple first test give you something useful to evaluate. Review both identities before adding stronger styling or longer motion.

Ready to explore the look? Visit the Hotel Lobby AI homepage to try the local two-photo preview and see the orange studio style concepts while video generation is being prepared.