AI storyboard canvas — multi-shot continuity

What the canvas is for

The single-clip generator emits one video at a time. The storyboard splits a script into shot nodes — each with its own prompt, duration, aspect ratio and model — then writes the result back onto the node so you can redo one shot without rerunning the whole piece.

Ratio and duration still come from the model’s admin-configured lists. Changing model snaps illegal values and persists them in project metadata.

How it pairs with /sora

Prototype prompts on /sora. Once the look is right, move winning shots onto the canvas, attach character/scene boards, and batch-render. GP still follows the same live catalog as /credits.

From empty canvas to a shot list

1. Create or open a project and drop shot nodes on the canvas.
2. Generate or upload character and scene references; attach them to shots that must stay consistent.
3. Set model, ratio and seconds per shot and check the GP estimate.
4. Generate one shot or a batch, preview, then export or rewrite prompts.

FAQ

How is storyboard generation billed?

Each shot is billed from that model’s GP rate × seconds (or per generation), same catalog as /sora. A failed shot refunds on its own; successful shots stay paid.

Can every shot use a different aspect ratio?

Yes. Ratios live in project metadata. Switching model clamps the value to that model’s allowed list so the API never receives an illegal ratio.