Every story in Ozu is organized into three levels: Scenes, Shots, and Beats. This vocabulary is borrowed directly from the film industry, with one AI-specific adaptation: a beat is one rendered frame.
Understanding these three levels is the single most useful piece of vocabulary in Ozu. Every other feature — storyboards, timeline, exports, asset reuse — refers back to them.
The three levels
Scene
A distinct story sequence. Typically one location, one continuous stretch of action. Example: Scene 1 — INT. KITCHEN — MORNING.
Shot
A subdivision of a scene, usually defined by camera angle or framing. Example: Shot 1 — Wide of the kitchen, Shot 2 — Close-up on the coffee cup.
Beat
The atomic unit. One beat equals one rendered frame in your storyboard. Example: Beat 1 — She pours the coffee, Beat 2 — She looks up.
How they’re labeled
Ozu tags every beat with a three-number notation: Scene.Shot.Beat.
You’ll see this notation everywhere in the app — on the Flow canvas, inside the Storyboard Node, on Timeline clips, in the Asset Library. It’s the canonical address for any moment in your story.
A worked example
Here’s a short scene broken down:
Ozu would label these seven beats as:
Seven beats = seven frames = seven images to generate.
Why three levels?
Scenes and beats alone would be enough to tell the AI what to render. Shots are the middle layer where cinematic intent lives.
- Scenes answer: where and when are we?
- Shots answer: how is the camera looking at this?
- Beats answer: what exactly is happening in this instant?
Giving Ozu a shot’s framing intent (“medium two-shot, Dutch angle, golden hour”) produces dramatically better generation than beats alone. Shots are also the layer you’ll most often edit — re-framing a single beat, or regenerating an entire shot.
Thinking in shots comes with practice. If you’re new to screenwriting, start by letting Ozu auto-extract shots from your script. Then adjust individual shots in the Script Node’s analysis view — your instincts will sharpen fast.
Scene, Shot, Beat in the UI
You’ll encounter this hierarchy in several places:
- Script Node analysis — after you paste a script, Ozu auto-extracts scenes, shots, and beats from it. You can review and adjust before anything is generated.
- Storyboard Node grid — frames are laid out by scene/shot/beat, grouped visually.
- Asset Library — storyboard frames are organized under a scene/shot tree, with beats nested beneath.
- Timeline tracks — clips on the primary video track inherit scene/shot/beat tags, so you can navigate or re-order by structural position.
Frames and Animatics
A frame is the rendered image of a single beat. An animatic is the timeline assembly of all your frames into a playable video.Read more → Pipeline and Nodes
The Scene/Shot/Beat hierarchy is produced by the Script Node and consumed by every node downstream. This is how the pipeline stays coherent.Read more → FAQ
Can I have a scene with only one shot?
Yes. Short scenes often have a single shot with multiple beats.
Can I have a shot with only one beat?
Yes, though most shots have 2–5 beats. Think of a beat as a moment worth its own frame; if the whole shot is a single moment, one beat is correct.
What if I want to insert a new beat between two existing ones?
Open the Script Node’s analysis view and add a beat. Downstream beat numbers renumber automatically, but existing frames tied to specific beat IDs are preserved — so you won’t lose work.
Does beat count affect how many credits I use?
Yes. Each beat generates one frame; each frame consumes credits. A 60-beat scene costs roughly 6× a 10-beat scene (modulated by model choice, resolution, and variant count). See Credits for details.
Is this the same as “scenes” in video editing software?
Close but not identical. In most editors a “scene” is just a chunk of timeline. In Ozu, a Scene is a structural unit that propagates through the whole pipeline — script, assets, storyboard, timeline, and exports all know your scene boundaries. 
