AI video
Each clip is 4–8 seconds, generated from a detailed prompt. The teaser above stitches three together with auto-advance.
A three-clip assembled teaser, all three worlds realized as scenes, and an animatic — generated with the platform’s AI tools. Direction, not finished footage.
A documentary from a universe that doesn’t exist. Every character is invented. Every world is impossible. Except for the dog.
Black. The sound of a collar jingling. A breath.
Fade from black.
“When did you first see Lion?”
ADA VOSS, off camera. Title card: GOD IS DOG.
“Before I was born.”
Cut to MARA VELL. She does not laugh. Hold.
A small black French bulldog before a luminous door on an endless salt plain.
Clip 01. Slow push-in. Film grain. Last pink of evening.
In one world, a memory repairer finds a fragment dated tomorrow.
Clip 02. Amber archive. Dolly.
In a third, a door has been locked for a thousand years.
Clip 03. Lion walks toward the door.
“God has to go outside.”
Lion, preoccupied. Nobody gets an explanation.
GOD IS DOG. A documentary from a universe that doesn’t exist.
Title. Fade.
Each world realized as a scene — key art, close-up, and four beats. The full film, not one chapter.
In a city where memories can be taken apart and repaired, Mara finds a dog carrying a moment that belongs to no one.
Mara closes the last vessel of the day. Amber light. The sound of glass settling. She has restored a thousand memories. None of them were hers.
A vessel is returned without a card. Inside: someone laughing in a garden. Each witness names a different person. Lion recognizes the garden.
Mara lifts the fragment and reads the date. It is tomorrow. Nothing can be remembered before it happens.
For the first time in her career, Mara does not reach for the tools. She reaches for the dog. The question stops being whose. It becomes: what do I owe a memory that isn’t mine?
Generated frames in sequence. Faster and cheaper than video — used to test the cut before investing in motion.
“When did you first see Lion?” — Ada Voss, off camera.
Each clip is 4–8 seconds, generated from a detailed prompt. The teaser above stitches three together with auto-advance.
Storyboard and key art are generated images, used to lock composition and palette before motion.
Narration is synthesized text-to-speech. Final delivery would use a recorded human voice.