A project → Weave.
A scene list tells you what happens. It cannot tell you whether the B-plot has gone quiet for twenty
minutes, whether the romance ever resolves, or whether two threads ever touch — and those are the
failures that make a finished film feel wrong in a way nobody can point at.
### The grid is the feature
Threads down the side, scenes across the top, a mark where they meet. A thread going quiet is
literally a run of empty cells, shaded red inside that thread's own span. It is the only view where
that is visible rather than deducible.
Mark height carries how much of the scene the thread owns. Wide marks are climaxes and resolutions.
### What it checks
- A thread with no setup beat, so nothing tells the audience it exists.
- A thread that never resolves — unless you mark it meant to stay open, which is a real thing a
series does deliberately.
- A resolution beat before its climax, which is almost always a mislabelled beat.
- A thread that averages 5% of a scene across ten beats: being mentioned rather than advanced.
- A thread that never shares a scene with another one — running alongside the film rather than
through it.
- Scenes where three threads claim more than the whole scene between them.
### Nothing is rewritten
Every finding is a location and a description. The fix is yours. A tool that quietly restructures
your story is one you stop trusting the first time it is wrong, and it will be wrong — "goes quiet in
act two" is sometimes exactly the intent.
### Threads survive a re-import
Beats are keyed to scene order, not scene id, the same as the clue tracker. Pasting in a new
draft creates new scene rows; a structure keyed on ids would evaporate at exactly the moment you most
need it.
### Start me off
On an empty project it creates the A-plot, the B-plot and an arc for each of your first two
characters — empty. Which scenes carry which thread is a judgement, and guessing it from scene
headings would be a guess presented as structure.