P—10 / SHOWROOM ROOM SWAP

Furniture retail video prompt for MiniMax H3

She taps the desk. The room changes. She never moves.

A furniture retail spot with one mechanic: a woman sits still with her eyes closed, and every time her fingers tap the desk the entire room around her becomes a different room. Four rooms in seven seconds, one unbroken subject. It is the cleanest transition device in the library, and it works because the thing that stays still is a person.

Steps
2 steps
Mode
Image to Video
Duration
15s
Ratio
9:16
Resolution
2K
Audio
Music bed + room tone
Cost
$1.95

Reference render — not generated on this site. Source: pollo.ai

Output reference

A woman in a striped knit sweater standing in a sunlit doorwayImage
Step 01 — the still
A woman sits still at a desk while the furniture showroom around her changesVideo
Step 02 — the clip

JPG · PNG · WebP · HEIC · 30 MB

2160 / 7000

Images 0/9 · Clips 0/3 · Audio 0/3 · 12 files max

No card, no account: watch and download 12 real MiniMax H3 clips, and run one prompt of your own on Agnes Video V2.0 — 16:9, watermarked, once. After that, sign in free for 3 clips a day on the free engine. MiniMax H3 on your own prompt starts at $24.9 a month.

1 FREE CLIP · WATERMARKED

02 — Inside

Two prompts, one workflow

Each step is the result beside the text that produced it. Copy either prompt into the console above.

Reference output, step 01

A woman in a striped knit sweater standing in a sunlit doorwayImage
The still
Step 01the stillAny image model
1,282 chars

The subject

The one thing that must not change across four room swaps. Framing, wardrobe and light all have to survive being dropped into four different sets.

Photoreal full-length lifestyle portrait, vertical 9:16, natural daylight.

SUBJECT:
A young woman in her twenties standing relaxed and centred, weight on one hip, looking calmly toward the camera. Long wavy blonde hair falling past her shoulders. Soft natural makeup. She wears a loose olive-and-cream horizontally striped knit sweater slipping off one shoulder, white frayed-hem denim shorts, and simple black flip-flops. A small silver-grey leather handbag hangs from one forearm.

FRAMING:
Full body, head to feet, subject occupying the middle third of the frame with clear space above and below. Camera at chest height, straight on, no tilt. Neutral standing pose with arms loose — nothing gestural, nothing that would look wrong in a different room.

SET AND LIGHT:
A simple sunlit doorway in a pale plaster wall, soft dappled shade falling across the wall behind her. Warm late-afternoon daylight from camera left, gentle contrast, no hard shadow across the face. Background deliberately plain and shallow so the subject can be relit.

RENDER:
Photoreal, natural skin texture, no beauty retouching, sharp on the face, mild depth-of-field falloff behind.

RESTRICTIONS:
No text, no logo, no brand marks on clothing, no second person, no strong colour cast, no studio backdrop.

Paste it into whichever image model you already use. We do not run one here yet — when we do, this button generates instead.

Reference output, step 02

A woman sits still at a desk while the furniture showroom around her changesVideo
The clip
Step 02the clipMiniMax H3 · Image to video
2,160 chars

The room swaps

Three timed blocks. The middle block contains four rooms and one repeated gesture that triggers each of them.

Create a video ad from [image1].

9:16, 24fps, 15 seconds. Sophisticated documentary style, warm ambient lighting, modern Scandinavian furniture retail. No dialogue, music bed and room tone only.

0-3s:
Close-up shot. In a softly lit furniture showroom, the woman from the reference walks in, her expression calm and curious. She pauses to take in the space, surrounded by modern Scandinavian furniture in pale wood and neutral fabric. Other customers browse in the background, but she remains the focal point. The warm ambient lighting and the organised, welcoming layout emphasise the showroom's clean aesthetic.

3-5s:
Close-up shot. She pulls out a chair and sits at a modern light-wood desk. The room feels cosy and inviting. As the music starts, she closes her eyes and taps her fingers lightly on the desk, fully immersed in the calm around her.

5-12s:
Sophisticated documentary montage. The woman remains the absolute visual centre, always facing the camera, eyes closed, tapping. Each time she taps the desk the room around her transitions into a different space, while she stays exactly where she is:
Children's room — she sits in a playful, vibrant space, a colourful rug underfoot, bright walls, lively furniture and scattered toys. The camera holds on her calm expression as the energetic environment contrasts with her stillness.
Minimalist room — a serene living room in clean lines and neutral tones. The simplicity enhances the quiet, and she looks deeply relaxed.
Gaming room — a modern setup with an ergonomic chair and vibrant LED accents. The atmosphere is energetic; her stillness stands out against it.
Greenery room — a plant-filled space with natural light through large windows. Wooden furniture blends with the plants. The outside world moves past in subtle motion blur, emphasising her focus.

12-15s:
The final shot returns her to the first showroom. She opens her eyes and smiles. The camera pulls back to show her standing in the centre of the store, pushing a cart loaded with flat-packed furniture, heading toward the exit.

Negative Prompt: Any text, subtitles, watermarks, brand logos, harsh light effects, black background.

4 levers

Make it yours

What is safe to change. Most libraries publish only this list, which is why so many copied prompts come back worse than the original.

  1. 01

    The trigger gesture

    A finger tap on a desk. Small, repeatable, and audible — which matters, because H3 generates the sound and the tap becomes the cut point. Swap it for a blink, a breath, a page turn. Do not swap it for anything that requires her to move her body.

  2. 02

    The four rooms

    Children's, minimalist, gaming, greenery. They are chosen to be maximally unlike each other — loud, empty, dark, green — so that four changes read as four changes. Four tasteful neutral rooms would look like one room with the furniture moved.

  3. 03

    The unchanging element

    She faces the camera, eyes closed, in the same chair, in every room. That constant is the whole illusion. If she opens her eyes or turns her head mid-montage the transitions stop reading as room swaps and start reading as cuts.

  4. 04

    The return

    The last block puts her back in the first room. Without the return the montage has no frame around it, and the ad ends inside a fantasy rather than in a store you can walk into.

Three ways to break it

  • Naming a real retailer

    The source of this prompt named a specific furniture chain throughout and supplied its logo as a second reference image. We removed both. The model has a very strong prior for that brand — it will produce recognisable signage, the corporate colour pair, and a logo end-card, and you will have made an unauthorised advertisement. Describe the aesthetic instead: pale wood, neutral fabric, flat-pack, organised.

  • Letting her move during the montage

    The subject is the anchor. Any change in her pose gives the model permission to re-solve her appearance, and by the fourth room she is a different woman in the same sweater.

  • Making the rooms similar

    Four variations on tasteful neutral furniture produce a montage in which nothing appears to happen. The rooms have to differ in colour temperature and light level, not just in furniture.

What it does

What the showroom room swap template does

Transitions are the hardest thing to get out of a video model, because a transition is by definition a moment when the model has permission to change everything, and models take that permission enthusiastically. This template solves it by removing the permission: it names, explicitly and repeatedly, the one thing that is not allowed to change.

"The woman remains the absolute visual centre, always facing the camera, eyes closed, tapping." That sentence appears at the head of the montage block, before any of the four rooms are described, and it is the reason the montage works. The model reads it as a constraint that applies across all four sub-scenes, and then reads four room descriptions as things that happen around a fixed point.

Compare that to the obvious way to write the same idea — four scene descriptions each beginning "she sits in a children's room…", "she sits in a minimalist room…" — and you can see why it fails. Each of those is a complete generation brief containing a person, and the model produces four people who resemble each other.

The trigger gesture is the second piece of craft here, and it is specifically chosen to be audible. A finger tap on wood makes a sound; H3 generates the sound and the picture together, which means the tap gives the model a shared clock for both tracks. The room changes on the tap because the tap is a discrete event in a soundtrack the same network is writing. A gesture with no sound — a blink, a slow exhale — leaves the model to guess when the change happens, and the transitions drift out of rhythm with the music bed.

The four rooms are chosen for contrast rather than for range. Bright and loud, empty and pale, dark with coloured light, green and daylit. Two of them are warm and two are cool; two are cluttered and two are bare. That spread is what makes four transitions in seven seconds legible — each cut changes the colour temperature of the whole frame, which the eye registers instantly, before it has identified a single piece of furniture.

The return in the final block is what turns a device into an advertisement. Three seconds of her back in the real store, eyes open, pushing a trolley, converts the montage from a daydream into a claim about what the shop contains. Retail spots that end inside the fantasy test badly for exactly this reason: the viewer enjoyed it and has no idea what they are supposed to do next.

A note on what we changed. The source prompt named a specific furniture retailer eleven times and included that retailer's logo as a second reference image. Both are gone. This is not only a legal fix — the version without the brand is a better prompt, because the model stops budgeting capacity for reproducing signage it renders unreliably and spends it on the room swaps instead. If you are making this for a real client, put their actual store photography in step one rather than asking a model to imagine their brand.

Everything here runs in image to video.

6 questions

Showroom room swap — common questions

  • 01

    How do I keep the same person across four scene changes?

    State the constant once, before the scenes, and phrase it as something that does not change: same position, same framing, eyes closed. Then describe only what happens around them.

  • 02

    Why does the trigger gesture need to make a sound?

    Because H3 writes the audio and the picture in one pass, so an audible event gives both tracks a shared cut point. Silent triggers drift out of time with the music.

  • 03

    Can I name a real store or brand?

    You should not. Models reproduce well-known retail brands recognisably enough to be a problem and badly enough to be unusable. Describe the aesthetic, or supply real photography as your reference image.

  • 04

    How many room swaps fit in seven seconds?

    Four, at roughly 1.75 seconds each. Five is possible and starts to feel like a strobe; three gives each room a beat.

  • 05

    Why do the rooms have to be so different from each other?

    Because the eye reads colour temperature and light level before it reads furniture. Four tastefully similar rooms produce a montage in which nothing visibly happens.

  • 06

    Does the ad have to return to the opening location?

    For a retail spot, yes. The return is what converts the device into a claim about a place the viewer can actually go.

Copy it, change three things, run it.

Every character is on this page. 15s at 2K costs $1.95.

Sign in free · 3 clips a day