Films made with INKFLOW

Every character and every location in these excerpts is a trained rank-64 model. That is why a face survives a whole song, and why a room is still the same room three shots later.

Each clip carries a burned-in AI-generated disclosure, as it should.

Try it first

Every film here started as a book

And the first three chapters of yours are free. Write the opening, read it back, and only then decide whether you want it filmed.

Start writing free — 3 chapters
No credit card. 10,000 AI words included. Export DOCX, EPUB or PDF and keep whatever you write, whether you ever pay us or not.

Subtle — the night drive

A cab at night, rain on the tarmac, then inside. The same two faces hold through low light, reflections and a change of location — the conditions that usually break a generated face.

Watch Di Ride — full contact

Ice hockey at speed, with bodies colliding against the boards. Fast motion and heavy occlusion are where identity models usually come apart; here the jerseys, numbers and faces stay put.

Birds Eye View — scene nine

A single sustained scene from the middle of the film.

Two characters, one room

The hardest shot here. Two named characters share a frame in their own distinct wardrobes for eighteen unbroken seconds. An unpinned second character does not drift randomly — it copies the other one's outfit — so holding two people apart takes trained identity for both.

Use the same cast

A complete rank-64 character model is free, with its specification readable in the file header. Take it and check the claim yourself.

Free model on Hugging FaceSee the whole shelf