Scenes and the spoken script

How a brief becomes scenes

You never storyboard by hand. The orchestrator reads your brief, plans the scenes itself, renders each one, and stitches them into a single cut. A short ask, up to the model's native take, renders as one continuous shot. A longer ask gets broken into segments that are rendered and joined. Either way you get one finished clip, not a folder of loose shots to assemble yourself.

Because the planner owns the scene breakdown, the real lever in your hands is the brief. Naming distinct moments is how you get distinct scenes: a brief that spells out three beats tends to produce three beats. Think of the brief as the outline you give a director, not the camera you operate yourself.

The spoken-lines field is a script, not a hint

The brief describes what the clip shows. The optional spoken-lines field does something different: it holds the exact words you want spoken, line for line. Whatever you type here becomes the spine the scenes are cut to, and the per-scene timing is fit around it.

This field is never auto-filled, and that is deliberate. Leave it blank and the scene planner writes the dialogue itself from your brand voice and the brief. Fill it in and the clip says your words and nothing else. Use it when the wording has to be exact: a specific claim, a compliance line, a founder's precise phrasing. Leave it blank when you are happy for the planner to draft the talk track for you.

One honest note. The words you put here are timed against the duration you set. A long script squeezed into a short duration forces the planner to rush or trim lines, so match the two before you generate.