AI product film

The whole room, in one call.

A meeting camera, taken apart and put back together. Every frame was generated with AI.

Look closer

Built to see everyone.

A lens on top takes in the table from every side, so nobody ends up sitting outside the shot.

Listening

Voices, from any seat.

  • Microphones set around the body pick up whoever is talking.
  • The camera follows the conversation, not the loudest chair.
  • Remote attendees hear the room as if they were in it.

Inside

Sound that fills the table.

A speaker at the base and a board at the core, laid out here as one floating diagram.

Back together

No studio. No camera. No renderer.

192 frames from text prompts, played back by your scroll.

Loading frames…

How it was made

A product ad without a shoot.

The question behind this page: can AI produce the kind of exploded-view product film that normally needs a studio, a physical product and a 3D artist? The frames were generated in Google Flow from text prompts, exported as an image sequence, and played back on a canvas that ties each frame to how far you have scrolled.

GenerateText prompts in Google Flow produced the assembled shot and the exploded view.
Sequence192 frames at 1280×720, taken straight from the generated video and cleaned up.
PlayYour scroll picks the frame, forwards to take it apart and backwards to rebuild it.

The inside of the device is imagined by the model. It is a plausible arrangement of parts, not a real teardown, so nothing here describes how the actual product is built.