North Protocol
A multimodal canvas.
An AI-native interface for film pre-production — text, image, and storyboard as one editable medium.
✺ — The problem
North Protocol's pre-production tool used to be three separate apps stitched together with copy-paste. Their users — directors and producers — wanted one canvas where they could write a scene, generate a frame, and revise both at the same time, without losing the through-line.
Sector
Creative tools
Year
2025
Duration
16 weeks
Team
1 Principal · 1 Designer · 2 Engineers
Stack
✺ — Approach
The same arc as every engagement — tuned to this problem.
Define · The unit of work
We spent a week shadowing two production teams. The atomic unit wasn't 'a prompt' or 'a frame' — it was 'a beat': a moment that connected text intent, visual reference, and motion direction. The whole interface had to compose around that.
Build · Streaming as a first-class state
Every panel renders partial state — text streams in word-by-word, frames generate progressively, edits are optimistic. The canvas never blocks; the user never waits without seeing progress.
Operate · A surface, not a chatbot
We retired the chat input. Every action lives where it belongs — on the beat, on the frame, on the camera path. The model is invoked by the interface, not addressed by the user.
✺ — Outcome
Three numbers we’d defend in public.
5.4×
more iterations per session vs. v1
82%
of sessions used multimodal edits, not text alone
$4.2M
Series A closed two weeks after launch
“Every other studio we talked to wanted to bolt a chat panel onto our app. Define AI asked what our directors actually do for a living, and then we redesigned the surface so the model fit underneath.”
Founder, North Protocol