A video a day, with nothing to film

You paste in a piece of writing or a product link. The machine writes the shot list, makes the picture, reads the narration, mixes the sound, and hands back a vertical file ready to upload. No camera, no presenter, no editing software.

  1. Write

    Cuts the writing into scenes, one frame per line.

  2. Draw

    Makes a picture for each scene, or finds real footage where it exists.

  3. Read

    One voice, read in a single take, never chopped line by line.

  4. Assemble

    Picture, captions, sound effects and a music bed into one cut.

  5. Done

    1080×1920

    A 1080×1920 file, with the caption written for each place you post it.

These five stages are the machine's real stages. The cards here illustrate them; they are not a live queue.

Cheap because it reuses, not because it cuts corners

Every part of a video is stored on its own: one frame, one stretch of voice, one assembled segment. Change one part and only that part is paid for again. Assets are banked per channel rather than per video, so the second episode of a series costs materially less than the first.

180+videos rendered
This is the real library that has run through the pipeline, not a demo.

See what it makes

Real videos, built by the pipeline above.

Real video goes here

Real video goes here

Real video goes here

Real video goes here

Not open yet

Studio currently runs our own channels. Leave an email and we will tell you when it opens — no newsletter.

What you do

Used only to tell you when it opens. Never shared.