What does one clip cost?

It depends on the model and the length you pick, and the exact figure sits on the button before you press it. The cheapest ten second clip runs about $1.81 and the longest fifteen second one about $8.51. An optional opening frame adds about ten cents.

Do I need an account?

Not to build the prompt. Choosing a vehicle, a colour, a model and a length, and reading the finished prompt, all work signed out and cost nothing. You sign in only at the point of generating, because that is the point where money is spent.

Can I use a vehicle that is not in the list?

Yes. Type your own takes any vehicle you name, from a 1990 Yugo to a delivery van, and asks only which body type it is so the assembly order matches.

Why only ten or fifteen seconds?

That is the length this format is proven at, and every extra second costs money and gives the model another chance to fit a part the wrong way round. For longer clips the full video composer exposes everything the model allows.

Do the clips have sound?

Yes. Every clip is generated with audio, because the tapping, ratcheting and panel clicks are half of why this format works.

Will every clip look the same?

No. The workshop, the box, the lighting and the camera are drawn from a deck of 960 combinations on every generation, and the model runs with a different seed each time. There is a switch to hold the workshop still if you are building a series that should look like one channel.

Miniature assembly

The unboxing format, without the workflow

Two hands open a sealed model kit, assemble a miniature car or motorcycle bolt by bolt, and reveal it under a sweeping light. Pick a vehicle below and the whole prompt is written for you, free, before you spend anything. One click generates one finished vertical clip with sound.

Made with this page

Sound is half the format, so these do not autoplay. Press play on any of them.

Truck and trailer, 10s

Dirt bike, 15s

Supercar, 10s

Supercar, 15s

Supercar, another workshop, 10s

1Pick the vehicle

A glass engine cover over a fully detailed miniature V6, lit from inside at the reveal.

2Colour

or pick one

3Model and length

Vertical 9:16, sound on.

Ten and fifteen seconds are what this format is proven at. For a longer clip or any setting this page does not expose, the full composer has every option the model allows. To build a series where the same vehicle and workshop carry across every clip, Cinema Studio keeps them consistent by name.

4Opening frame (optional)

Generate the first frame and start the video from it. You keep it as a thumbnail.

5The prompt

Vertical 9:16, ultra realistic macro POV time-lapse, 15 seconds. An angled macro camera close to the mat, almost at surface height, over a brushed stainless steel workbench in a clean white technician's room, lit by low dramatic side light with deep shadows and bright highlights on metal. On a black rubber work mat sits a sealed premium model-kit box: deep navy card with a metallic emboss and a magnetic-close lid, with no readable text or logos anywhere. Two adult hands with medium skin tone, short clean nails and plain dark grey sleeves pushed to the forearm enter from the bottom of the frame.

0.0 to 2.3s: the hands break the seal, lift the lid off and set it aside. Tape peels, cardboard slides.

2.3 to 4.5s: they lift out a fitted foam tray and remove the tissue wrapping, revealing every miniature component seated in its own cut-out: chassis, engine block, gearbox, front and rear suspension arms, axles, brake discs, wheels with soft rubber tyres, seats, dashboard, steering wheel, doors, windows, body panels, rear wing, diffuser, front and rear light units, exhaust, and a small tray of screws and bolts. Foam clicks, tiny metal parts tick against each other.

4.5 to 12.0s: rapid assembly on the work mat, each step a short macro shot with smooth speed ramps between them. The chassis lies lengthways in the tall vertical frame, running from the bottom of the shot to the top, nose pointing away from the hands, with one wheel mount at each of its four corners and the engine bay behind the cabin, nearest the hands. The chassis is set down on the work mat, then the engine block is lowered into the rear bay, then the gearbox is mated to it and both are bolted down with a miniature ratchet, then the front suspension arms and springs are fitted, then the rear suspension follows, then the axles, then the brake discs and calipers go onto the hubs, then the wheels slide onto the axles and the wheel nuts are tightened, then the seats are dropped into the cabin, then the dashboard and steering wheel, then the doors and windows are clipped on, then the body panels are lowered onto the chassis and pressed home, then the rear wing is mounted and the diffuser is fitted underneath, then the light units and the exhaust are fitted last. Ratchet clicks, bolts biting, panels snapping into place.

12.0 to 13.5s: the hands wipe the finished Ferrari 296 GTB with a grey microfibre cloth and set it on a small rotating display stand. Cloth on paint, the soft whir of the turntable motor.

13.5 to 15.0s: the light drops to cinematic, a hard light sweeps across the glossy Rosso Corsa paint, the headlights click on, and the camera orbits slowly around the finished miniature for a final hero shot.

Consistency: the same two hands throughout, five fingers each, same skin tone, same plain dark grey sleeves, same workshop and same table. Every part keeps the shape and colour it had in the tray. The paint stays Rosso Corsa. The wheels stay the same size and style. No part appears before it is taken out of the box, and installed parts stay installed. Nothing moves unless a hand moves it. Every part is installed facing the right way: the body sits along the length of the chassis and never across it, all four wheels carry it at the corners, seats face forward toward the steering wheel with their backrests toward the rear, the steering wheel faces the driver, headlights face front and the exhaust faces rear. The finished Ferrari 296 GTB matches the parts shown at the start.

Style: ultra realistic macro miniature photography, cinematic, premium, shallow depth of field, realistic reflections, natural hand movement, professional product commercial lighting, sharp focus on whatever part is being worked on. Every action has its matching sound and no sound happens before or after its action. No narration.

Avoid: glitches, morphing, melting parts, colour shifts, duplicated components, missing wheels, extra fingers, extra or deformed hands, floating tools, parts fitted the wrong way round, disappearing screws, an unfinished vehicle, sudden environment changes, wrong vehicle proportions, a full sized real vehicle, cartoon or cheap plastic look, camera shake, text overlays, captions, watermarks.

Keep the same workshop across clips (for a series that should look like one channel)

Questions

What does one clip cost?: It depends on the model and the length you pick, and the exact figure sits on the button before you press it. The cheapest ten second clip runs about $1.81 and the longest fifteen second one about $8.51. An optional opening frame adds about ten cents.
Do I need an account?: Not to build the prompt. Choosing a vehicle, a colour, a model and a length, and reading the finished prompt, all work signed out and cost nothing. You sign in only at the point of generating, because that is the point where money is spent.
Can I use a vehicle that is not in the list?: Yes. Type your own takes any vehicle you name, from a 1990 Yugo to a delivery van, and asks only which body type it is so the assembly order matches.
Why only ten or fifteen seconds?: That is the length this format is proven at, and every extra second costs money and gives the model another chance to fit a part the wrong way round. For longer clips the full video composer exposes everything the model allows.
Do the clips have sound?: Yes. Every clip is generated with audio, because the tapping, ratcheting and panel clicks are half of why this format works.
Will every clip look the same?: No. The workshop, the box, the lighting and the camera are drawn from a deck of 960 combinations on every generation, and the model runs with a different seed each time. There is a switch to hold the workshop still if you are building a series that should look like one channel.