Hotel Lobby AI logoHotel Lobby AI
Back to guides

How to Make a Hotel Lobby AI Video from Two Photos

A visual walkthrough of photo uploads, left and right roles, video settings, credits and downloads, with real screenshots and practical photo tips.

Hotel Lobby AI team
Workflow illustration: two separate photos become the left and right performers in one orange-stage video.
Workflow illustration. The people shown are symbols, not a generated video result.

To make a Hotel Lobby AI video, upload two clear photos into the left and right slots on Hotel Lobby AI, choose a length and model, and generate. No prompt is required. The result is a 5, 10 or 15-second, 720p landscape MP4 with the original audio. Free-credit videos have a moving watermark; a purchase unlocks watermark-free downloads for eligible videos, including ones already made.

This guide follows the actual website interface. The original screenshots were captured on September 30, 2026; model controls and default prices were updated on October 1, 2026. You can open each screenshot at full size. For current offers and the final generation cost, check the generator and pricing page.

Jump to: Prepare photos · Upload · Settings and cost · Generate and download · Troubleshooting · FAQ

Looking for the Migos AI orange-room trend? Open the Migos AI video generator for the same two-photo tool, a quick trend explanation, and current credit and watermark details.

If you found the trend through the performers' names, the Quavo and Takeoff AI guide identifies the original COLORS performance and explains how your two photos become the left and right people.

What kind of Hotel Lobby video will you make?

This tool uses the orange-booth Hotel Lobby performance as a fixed template: two performers, a hanging microphone and the original soundtrack. Your first photo supplies the left character and your second photo supplies the right character. The template supplies the scene and performance.

You do not need to upload a reference video, build an orange background or write a motion prompt. This also sets the boundaries: you cannot choose a different song, invent new choreography or switch the output to vertical inside this generator. If you need a custom performance, use a workflow that supports your own video reference instead.

  1. Prepare two clear, separate photos

Start with one person per image. Choose a front-facing photo with even light and an unobstructed face. A waist-up or full-body picture gives the generator information about the outfit as well as the face. Use the original image rather than a screenshot of a tiny social profile picture.

  • Use: a clearly visible face, simple background and the clothes you want the character to wear.
  • Replace: sunglasses, masks, heavy filters, motion blur or a hand covering the face.
  • Separate: a couple or group photo into two individual images before uploading. If the crops become too small, use better source photos.
  • Check the file: use JPG, PNG or WebP, no larger than 20 MB. The shorter side must be at least 300 pixels; avoid extremely narrow crops or panoramas.

Use photos you have permission to upload. If you want to play both roles, you can use your own photo in both slots. Pet pairings are also supported, but a human portrait and an animal photo can produce different results; neither is a promise of exact likeness.

The website's photo checklist recommends one clear front-facing person per image and advises against groups, covered faces and heavily filtered selfies.
The photo checklist on the live website, captured September 30, 2026. Good inputs help, but do not guarantee perfect identity preservation.

  1. Put each photo in the correct slot

Open the Hotel Lobby generator. Select Upload photo under Left person, then choose the image for the performer who opens the verse. Add the second image under Right person.

Check both previews before continuing. A face in the wrong slot becomes the wrong performer. Replace the uploaded photos if you want to change sides; you do not need to describe the positions in a prompt.

Actual upload interface with two separate slots: Left person, Opens the verse; Right person, Joins on the right.
The two upload slots before photos are selected. Left and right refer to their positions in the video.

The orange-booth video beside the controls is a template preview. It shows the reference performance for the selected duration. Your finished video appears in that area after generation succeeds; the preview itself is not a result made from your photos.

  1. Choose the length, model and credit cost

Select 5s, 10s or 15s, then use the model dropdown to choose Seedance 2.0 Mini, Seedance 2.0 Fast, Seedance 2.0 or Seedance 2.5. Each option shows the exact cost for the selected length. Every length is made in one generation. All four output 720p.

Model dropdown with real Seedance names and the exact credit cost for the selected duration.
Updated controls, October 1, 2026. Default 10-second Mini: 24 credits. Five-second Mini: 12 credits.

Default costs updated October 1, 2026 (the live quote takes precedence):

Model5 seconds10 seconds15 seconds
Seedance 2.0 Mini12 credits24 credits35 credits
Seedance 2.0 Fast44 credits87 credits130 credits
Seedance 2.073 credits145 credits217 credits
Seedance 2.5110 credits219 credits328 credits

For a first attempt, consider 5-second Seedance 2.0 Mini. In this guide's historical example, an unused 20-credit sign-up balance covered that combination at the October 1, 2026 default cost of 12 credits; the initial 10-second selection cost 24 credits. This is not a promise about today's sign-up offer. Check the current credit costs and offer eligibility, then read the quote beneath Generate and confirm your eligible balance covers it.

For plan choices and watermark details, see Is Hotel Lobby AI free?.

  1. Generate, review and download

Sign in if prompted, check the two photos and cost, then select Generate Hotel Lobby video once. Uploading photos and changing the settings prepare the video; pressing Generate starts the job and uses credits.

Generation can take a few minutes and may take longer when the video service is busy. Follow the progress display rather than submitting the same attempt again. If you leave the page, check My videos while signed in to the same account.

When the result is ready, play it with sound before downloading. Check the faces when they turn, which character is on each side, the timing of the audio and any visible distortions. Then use the download control to save the MP4.

OutputWhat to expect
DurationThe selected 5, 10 or 15 seconds
Frame720p, 16:9 landscape
AudioThe original Hotel Lobby audio for the selected length
Watermark

A small moving site watermark on free-account output; eligible clean downloads unlock after a purchase

For a vertical social post, place the landscape video inside a vertical canvas in your editor. Cropping directly to 9:16 can cut off one of the performers. Label the posted clip as AI-generated. Included audio is not a blanket music license; follow the destination platform's music rules.

What should you change if something looks wrong?

The face does not look like the photo

Try a sharper, front-facing source with both eyes visible, less filtering and no facial obstruction. Change one input at a time so you can tell what helped. You can try another Seedance model if your budget allows, but no model guarantees a perfect face match.

The two identities seem mixed up

Check that each upload contains only one subject and that the previews show the intended people. Replace a crowded group crop with a separate portrait. Recheck the left and right slots before spending credits on another attempt.

The upload is rejected

Check the image format, file size and dimensions against the requirements above. A supported file extension alone does not fix a corrupt image or a crop that is too small. Export a fresh JPG, PNG or WebP if necessary.

The video is still processing

Look for the existing task in My videos. A slow task is not necessarily a failed task. If a generation fails, its generation credits are returned automatically. A completed video that you simply dislike is not the same as a failed generation; a new attempt uses credits again.

There is a watermark after downloading

Free-account output includes the site's moving watermark. After a purchase, return to the same account and download the eligible video again. The file you saved earlier will not change on your device. Read the watermark and existing-video explanation.

Hotel Lobby tutorial FAQ

Do I need a Hotel Lobby prompt?

No. This generator already contains the template. The controls are your two photos, duration and model; it has no prompt field for changing the scene or song.

Can I make it completely free and without a watermark?

Free registration does not guarantee enough credits for a completed video. Check the current offer, eligibility and generation cost. When the site watermark is enabled, accounts without a paid order receive a watermarked version; a purchase unlocks eligible clean downloads. These are separate conditions, explained in the free-credit guide.

Can I make a longer or vertical version?

The longest Hotel Lobby video is 15 seconds, made in one generation on every model with the original soundtrack. The output is always 16:9 landscape; for vertical posting, see the canvas tip above.

Will it perfectly copy my face and lip movements?

No guarantee is made. Clear photos improve the input, but identity, expressions and lip sync can vary. Review your own result before sharing it.

What should I do next?

Choose two photos and open the generator. Start by checking the quote for 5 seconds + Seedance 2.0 Mini, and confirm your eligible balance covers it before generating. See the current credit costs and sign-up offer if you need to check the terms.