Scenes
Distinct story sequences extracted from scene headings and structural context. Each scene has a location, time of day, and a list of the characters who appear.
When the Script Node finishes reading your script, it produces a structured breakdown: Scenes, Shots, Beats, and a full roster of the Actors, Props, and Sets your story needs. This page explains what each piece means, how to read the results tabs, and how to edit before anything else gets generated.
Scenes
Distinct story sequences extracted from scene headings and structural context. Each scene has a location, time of day, and a list of the characters who appear.
Shots
Subdivisions of each scene, defined by camera framing or coverage. Shots give PrePrompt the cinematic layer — “how is the camera looking at this.”
Beats
The atomic units — one beat per rendered frame. Each beat is tagged with its Scene.Shot.Beat address (e.g., 1.1.3).
Actor / Prop / Set nodes
Every named character, object, and location gets its own node on the Flow canvas — auto-spawned, ready to generate.
Once analysis completes, the Script Node switches to a tabbed results view. Five tabs, one per category, plus a stats bar across the top showing your totals — scenes, shots, beats, Actors, Sets, Props.
| Tab | What you see |
|---|---|
| Scenes | Each scene as an expandable section — location, time, shots, beats, dialogue, and tags |
| Actors | Cards for each character — description, scene appearances, wardrobe notes |
| Sets | Cards for each location — description, time of day, which scenes use it |
| Props | Cards for each object — description, which scenes it appears in, whether it’s set-integrated |
| QC Log | Any issues PrePrompt flagged during analysis (ambiguous references, missing details) |
The Scenes tab is usually where you start — scrolling through expanded scenes lets you see the whole story at a glance.
Open a scene in the Scenes tab and you’ll see:
INT. KITCHEN — MORNING, or whatever you wrote.Nothing in the results view has been rendered yet. This is structure only — the plan PrePrompt will use when you move to the Storyboard stage.
The QC Log is PrePrompt’s analysis checklist. If anything in your script was ambiguous or under-specified, it shows up here.
Ambiguous references
A character named “the detective” who’s otherwise unnamed, or a prop called “the artifact” without a description. PrePrompt still creates nodes for these, but flags them for your review.
Missing details
Scenes without clear time-of-day, characters without physical descriptions, or sets without explicit locations. PrePrompt makes reasonable inferences and shows its work.
Orphan entities
Anything referenced in dialogue or action but never introduced — a name that appears once in passing, for example. You decide whether to add the entity or edit the reference.
You don’t have to fix every QC issue. The log is advisory — PrePrompt will still generate with what it has, but cleaning up flagged items produces better downstream results.
You can adjust anything in the analysis before the pipeline moves forward.
Edit scene structure. Expand any scene and rename it, change its location, or adjust its time-of-day. You can also reorder shots within a scene.
Add, merge, or remove beats. Open a shot and add a beat between two existing ones; downstream beat numbers renumber automatically. Merge two beats into one if they should share a frame.
Rename or merge Actors. If PrePrompt created two Actor nodes for the same person (e.g., “SARAH” and “Sarah Walker”), merge them from the Actors tab. Renames propagate through every scene.
Edit Prop and Set descriptions. Sharpen the description on any asset card. The text here becomes the seed for hero-look generation.
Resolve QC issues. Click a flagged item in the QC Log to jump to the affected entity and fix it in context.
When the analysis looks right, tap Create Asset Nodes. PrePrompt spawns a node on the Flow canvas for each Actor, Prop, and Set — plus a DRB (Director Review Board) node downstream of all of them.
You’re now out of the Script Node and into the Asset stage. Open any new node to start generating.
If you want a production-planning view of everything your script requires — grouped by scene, with status, priority, notes, and assignments — the Shot List Node reads from the Timeline and your script structure. It’s not the same as the Shots extracted during analysis; it’s a downstream planning layer for crews on set. See the Shot List docs for detail.
Does analysis generate images? No. Analysis is structural only — scenes, shots, beats, and asset rosters. Image generation starts after you tap Create Asset Nodes and open individual asset nodes.
Can I re-run analysis on an edited script? Yes. Paste the updated script and submit again. PrePrompt re-analyzes from scratch, but existing Actor/Prop/Set generation work is preserved when names match.
What if PrePrompt missed a character? Add an Actor manually from the Actors tab, or spawn a new Actor node on the canvas after the initial asset creation. Name it carefully so downstream references can find it.
Can I change scene order after analysis? Yes. Reorder scenes from the Scenes tab; beat IDs renumber automatically.
What’s the difference between a “Shot” from analysis and the “Shot List” node? The Shots extracted during analysis are the cinematographic subdivisions inside a scene — one Shot contains multiple Beats. The Shot List Node is a downstream production-planning tool that takes shots from your Timeline and adds status, priority, and notes for crew use.
Will the analysis view stay editable after I create Asset Nodes? Yes. You can reopen the Script Node any time and adjust structure. Changes that affect assets (renaming, deleting, re-scoping) can require downstream work to rebuild.
How does PrePrompt decide what’s a Prop vs. a Set? Broadly: Sets are locations or environments, Props are physical objects. If something is a physical object that’s built into an environment (a throne in a throne room), PrePrompt marks it as both Prop and Set-integrated, which wires it into the Set for visual consistency.