How many photos does the shot need?

Two, one per figure, and the panel will not start without both — the shot is a duo rather than a single subject. Both are previewed on this page, and neither is uploaded until you sign in.

What is the stage, and can I run it without one?

The stage is the reference clip the motion comes from — the camera, the gestures, the backdrop and the timing are read from it rather than described to the model. A prompt can only ask for a locked-off camera and a pair of gestures, and the model then invents the specifics, so the result is a shot in that style rather than that shot; a clip shows it instead, while the two faces still come from your photos. There is no way to run without one. A run always carries a clip, which is what makes the output the shot this page promises rather than an approximation of it, and you can bring your own clip in place of any stage's.

Four, and each is a different backdrop with its own reference clip: Hotel Lobby, the trend's own orange staging; Luxury Lobby; Studio Booth; and Street Cypher. Hotel Lobby ships in two lengths — 7 and 30 seconds — because a clip's length is the shot's length; the other three ship in one.

Which models can shoot this template?

5 of the video models this deployment has keys for can shoot this template, and those are the 5 offered on this page. The shot takes two image references plus a clip for the motion, so a model whose request has nowhere to put a clip cannot shoot it. Each states its own length range, resolutions and reference limit before you start.

How long can the clip be?

Up to 30 seconds, which is the longest single run among the models that can shoot this. A single generation is the finished shot — the panel does not stitch clips together or edit video — so that is the ceiling for one run.

Whose photos can I put in it?

Your own, or anyone's you have permission to use. What the template does not change is the ordinary rule about likeness: using a real person's photo to make it look as though they performed this is a different thing from casting yourself and a friend, and it is the same line every AI video tool has to draw.

Does generating cost anything?

Generating spends credits, priced by the model and the settings you choose. You can hold credits through a monthly subscription or permanent credit packs that never expire, with exact cost estimates previewed before every run.

Two photos in, one fixed shot out

Hotel Lobby AI Video Generator

The staging is fixed and the pair is yours. Add two photos, pick the stage the motion should come from, and the panel takes the camera and the gestures from that stage's clip, keeps both faces, and films the shot the trend is built on.

Two-photo duo shot

0/2 photos added

The stage (the motion)

Stage clip

The camera, the gestures and the backdrop come from this stage's clip.

140creditsAdd both photos

5 models here will hold two photos and a reference clip in one run

MiniMax H3

MiniMax H3 Max

Seedance 2.5

Seedance 2.0

Wan 3.0

What the Hotel Lobby AI trend is

It is a staging, not a model. Two people stand shoulder to shoulder against a flat orange wall, one microphone hangs on a cable between them, the camera is locked off at chest height, and the two of them keep a deadpan expression while making the same small hand gestures in time. In late September 2026 creators began running that staging through video models with their own two photos standing in for the performers, and the pairs people cast into it — pets, cartoon characters, friends, couples, athletes, film characters — are what turned it into one of the season's most-copied clips.

What a template like this fixes is the staging, and there are two ways to hand a staging to a video model: describe it, or show it. Describing it works — a prompt can ask for a locked-off camera and a pair of gestures — but the model then invents the specifics, so what comes back is a shot in that style rather than that shot. Showing it is what a reference clip is for: the model reads the camera and the choreography from the clip and the identity from your two photos, which is why the panel below asks for both, and why it holds the model to ones that can take a clip and two images at once.

The pair is the part you bring. That is the whole trade: the wall, the light, the microphone and the movement come from the reference clip, and the two faces come from your two photos.

One thing worth being clear about, because it is the difference between joining a trend and misusing one: this page names the trend because that is what the trend is called. It has no connection to the record the staging came from and it ships no audio from it — and the shot it makes is a duo in that style rather than a copy of anybody's video. The clips shown here are re-hosted so that the stages can be run at all, and each one is captioned as what it is; none of them is presented as a run made on this page.

Two performers in orange outfits against a flat orange backdrop, with a microphone hanging between them. — The staging the trend is built on, as it was filmed. This page names the trend because that is what it is called; it is not affiliated with it, and the shot it makes is a duo in that style rather than a copy of this one.

The stages this page runs in

Four backdrops, each with its own reference clip. The first is the trend's staging in two lengths, shot on a clip we generated for it; the other three are demo clips published by rapduo.ai, whose stage names this page borrows.

Hotel LobbyThe original orange staging — two figures, one hanging microphone, a camera that never moves.

Luxury LobbyA chandelier-lit hotel lobby — marble underfoot, one microphone each, a slow walk-up.

Studio BoothA recording booth — neon rim light, a window onto the live room, one microphone between you.

Street CypherA night street — a graffiti wall, warm street light, a retro microphone each.

Three steps, all on this page

Nothing here sends you to a workbench to finish the job.

Add two photos

One photo per figure, previewed on this page. Nothing is uploaded until you are signed in — the panel keeps them locally until then.

Pick the stage

Each stage is a reference clip, and the clip is where the camera, the gestures and the backdrop come from — the difference between this template and a description of it. Hotel Lobby ships in two lengths, the others in one, and you can bring your own clip in place of any of them.

Generate, and watch it here

Keep or rewrite the prompt, choose the model and the resolution, then generate. The clip plays below the panel; reloading clears it, and the run stays in your video history.

What the panel gives you

Counted over the 5 models on this page that can shoot the template, not over the whole catalog.

2 photos and a clip

The panel takes 2 photos, one per figure, plus a reference clip for the motion. Generate stays closed until both photos are in; the clip is optional, and the panel says which of the two it is running when it is not there.

Faces and outfits carried from your photos

Up to 30 seconds

That is the longest single run the models here declare, and it is the ceiling: the panel does not stitch clips together or edit video, so one generation is the finished shot.

No cuts, no stitching, no editing

720p · 2K · 480p · 1080p · 4K

Between them the models here cover those tiers, and each one declares its own — the panel offers exactly those rather than one shared list.

The cost is shown before you press generate

Sound

2 of the models here put their own generated audio on every clip; the rest come back silent. Either way the trend's audio is yours to add where you post it.

Finished exports carry no watermark

Video Engine Capability Matrix

Different video models excel at different horizons — from 4K realism and native voice synthesis to extended 30-second continuous scenes. Compare real runtime capabilities directly.

Video Model	Max Resolution	Max Duration	Audio / Sound	Specialized Strength	Studio Direct
MiniMax H3minimax-h3	2K Quad HD	Up to 15s	Native Auto Audio	Multimodal references, native sound effects & fluid physics	Launch
MiniMax H3 Maxminimax-h3-max	720P HD	Up to 15s	Native Auto Audio	High-energy commercial actions & dynamic motion rendering	Launch
Seedance 2.5seedance-2-5	1080P Full HD	Up to 30s	Switchable Sound	Extended 30-second continuous scenes & character consistency	Launch
Seedance 2.0seedance-2-0	4K Ultra HD	Up to 15s	Standard	Frontier cinematic generative video synthesis	Launch
Wan 3.0wan-3-0	1080P Full HD	Up to 30s	Standard	Frontier cinematic generative video synthesis	Launch

Frequently Asked Questions

Yes. Video clips produced on active paid memberships include commercial licensing rights, allowing you to use them in client deliverables, paid ad creatives, social broadcasts, and commercial productions.

It is a staging that creators run through video models: two people shoulder to shoulder against a flat orange wall, one microphone hanging on a cable between them, a camera locked off at chest height, and the same small hand gestures from both of them. What changes from clip to clip is the pair — pets, cartoon characters, friends, couples, athletes, film characters — and that is what this page takes two photos for.

Have more specific requirements or questions?

PRICING DETAILS →LAUNCH STUDIO →

Ready to cast your pair?

Two photos, one shot, and the clip plays on this page a couple of minutes later. Every run is kept in your video history, where you can download it or send it on to the canvas.

Open the video workbench Animate a still picture instead