N/02 · STORY / PRE-VIZ Pre-launch

StoryViz

Storyboarding for film & TV that grows with the production — draw a rough sketch, and walk it up a fidelity ladder to photoreal while the whole board stays in one consistent style.

IN rough sketch + shot description OUT consistent boards, at any fidelity

The problem

Boards evolve. Early in a production you need fast, throwaway sketches; by the time you're pitching sequences to a director or client, you need something that sells the shot. Traditionally that means redrawing — or hiring out — every pass.

And the moment quality ramps up, consistency breaks: styles drift between cells, characters stop looking like themselves, and the board reads as a patchwork instead of a film.

The fidelity ladder

Every shot starts as a drawing on a layered 16:9 canvas. "Send to AI" regenerates the flattened canvas at whichever rung the production has reached — the same board, climbing in quality as the project matures.

✏️

Rough sketch

anyone can draw it
✒️

Illustrator

clean pro linework
🎨

Colour sketch

light, palette, mood
🖌️

Cartoon

full stylised render
🎬

Photoreal

sells the final shot

What makes it different

USP 01

Unitize style

One click regenerates the entire project against a single anchor shot's aesthetic — every cell, one consistent look. The patchwork problem, solved at the board level.

USP 02

@Characters

Design a character once in final style; StoryViz back-generates a full style ladder for them. Mention @Name in any shot description and the right reference is injected into the generation — the character stays themselves in every cell.

USP 03

Draw over anything

Any generated version can become the locked background layer — sketch corrections on top, regenerate, and it becomes the next version. Human and AI take turns.

USP 04

Versioned like a pipeline

Every generation is stored as a flattened, numbered version — flick back through the shot's history, pick any version as the card image, or branch from it. It's shot versioning, because that's how film works.

Under the hood

A custom HTML5 layered canvas engine (locked background layer, reorderable draw layers) feeds flattened frames to image-to-image diffusion via fal.ai, steered by the shot description, the chosen style rung, and any @character references. Projects, sequences, shots and versions live in Postgres (Neon + Drizzle), imagery in Cloudinary, payments through Stripe — a full production-grade product, not a prototype.

i2i diffusion fal.ai Custom canvas engine Next.js / TypeScript Neon / Drizzle Cloudinary Stripe Clerk