Hotel Lobby AI Template
Upload two reference photos, and the template places both subjects into an original two-person studio performance — orange stage, one microphone hanging at the centre. Updated October 1, 2026.
What the template actually fixes
The template holds the staging, the framing and the look: two performers on opposite sides of a single hanging microphone, against a warm orange backdrop, fixed camera. You supply the two people as reference photos. It does not take a name, a song or a video and copy it, and no original commercial recording is bundled.
Once the scene is set, the model composes a new clip around your references. That is where the variables live — photo quality, lighting and how clearly each subject is separated.
What you still control
- Duration — 5 or 10 seconds
- Resolution — 768P or 2K
- Framing — 9:16 vertical for short-form feeds, 16:9 landscape, or 1:1 square
- Creative direction — free text for staging and movement
Everything in the list above changes the credit charge, which is displayed before you submit.
Using it in three steps
- Upload one clear reference photo per subject. For people, use consenting adults whose photos you have permission to use.
- Choose your settings, add any creative direction, and review the displayed credit charge.
- Submit. Once the safety check and generation succeed, preview the result in My creations and download the video.
Submitted tasks continue running on the server if you close the page. Return with the same browser session or your saved recovery code.
Template or manual prompting
Use the template when you want the recognised format and you would rather not spend credits experimenting. Write your own scene direction instead when you want a different stage entirely, a different framing, or a product or a pet rather than two people.
A working direction keeps the scene and the positions in front, and keeps names out. For example:
Create an original two-person performance in a seamless orange studio with one microphone hanging at the center. Keep the first reference subject on the left and the second on the right. Give each performer natural, distinct movements and maintain a fixed camera.
In practice, a weaker result is usually a photo problem before it is a prompt problem.
Photos that work, photos that do not
| Works well | Reduces consistency |
|---|---|
| One person, well lit, head fully in frame | A group shot used as a single reference |
| Front-facing or slight three-quarter angle | Heavy side shadow across the face |
| Two separate files, one subject each | Two people inside one photo |
| JPG, PNG or WebP under 10 MB | Screenshots of screenshots |
| Same person twice, if you want one performer on both sides | A cropped forehead or chin |
What one run costs
| Output | Credits |
|---|---|
| 5 seconds, 768P | 50 |
| 10 seconds, 768P | 100 |
| 5 seconds, 2K | 85 |
| 10 seconds, 2K | 170 |
Planned packs are Starter — 300 credits for US$9 — and Creator — 1,000 credits for US$28. They are one-time purchases with no subscription and no automatic renewal. Public paid checkout is not open yet, so the exact price and credit quantity are disclosed at checkout. Rejected prompts are not charged, and a confirmed provider failure returns the credits. See Pricing and credits.
Two things to understand before your first run
- It generates a new scene; it does not swap a face into footage you already have. There is no workflow here for replacing a person in an existing video.
- No specific likeness or performance is guaranteed. Processing times vary with the model and the queue, and results vary with your references.
Reference photos are screened with OpenAI image moderation before generation. A blocked or unavailable check prevents the run. See the Acceptable Use Policy.
Frequently asked questions
Do I need the original Hotel Lobby video?
No. Nothing from it is uploaded and no frame of it is used. The generator builds a new scene from your two reference photos.
Can I use two photos of the same person?
Yes. Upload the same person as both references if you want one performer appearing on both sides of the mic.
Can I use pets or objects instead of people?
The template is built around two subjects. The uploader only restricts file type and size: JPG, PNG or WebP under 10 MB each.
How many credits does one run cost?
Between 50 and 170 credits, depending on duration and resolution. The exact charge is shown before you submit.
Is the Hotel Lobby AI template free?
There is no free generation allowance at the moment. Credit packs are planned but public paid checkout is not yet open. See Pricing and credits for the current status.