Verified against ChatGPT · 2026-08-10
Turn a one-line premise into a full plot outline where each beat actually causes the next one
Expands a premise into a structured outline built on causation rather than a sequence of things that merely happen one after another, using whichever story structure fits the material.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are expanding a story premise into a full plot outline. The test for every beat in this outline is causation, not sequence — each beat should happen because the previous one did, not just after it. PREMISE A disbarred lawyer takes a job ghostwriting appeals for death row inmates, and finds evidence that could exonerate a client she's not allowed to legally represent. GENRE Literary thriller. STRUCTURE PREFERENCE Three-act structure with a clear midpoint reversal. PROTAGONIST'S CENTRAL GOAL Get the exonerating evidence to someone with standing to act on it before the scheduled execution date. SCOPE About 28 chapters, novel-length. PHASE 1 — Spine Before breaking out individual beats, state the causal spine of the whole story in one paragraph: what the protagonist wants, what specifically stands in the way, and what changes about them by the end. Everything in Phase 2 has to trace back to this spine. PHASE 2 — Beat outline Break the story into beats following Three-act structure with a clear midpoint reversal.. For every beat, state not just what happens but why it happens given the beat before it — if a beat could be removed without breaking the causal chain to the next one, flag it as optional rather than including it as load-bearing. Pay particular attention to the midpoint and the low point: the midpoint should change what the protagonist is pursuing or how they're pursuing it, not just raise the stakes of the same pursuit, and the low point should be a direct consequence of a choice the protagonist made earlier, not bad luck imposed from outside. PHASE 3 — Causality audit Go back through your own beat list and check each transition: does beat N happen because of beat N-1, or merely after it? Flag any transition where you had to rely on coincidence or a character acting out of established character to make the sequence work, since those are the points most likely to feel unearned to a reader. OUTPUT FORMAT 1. The spine paragraph. 2. The full beat outline, each beat with its one-line causal justification. 3. The causality audit findings from Phase 3, including anything flagged as weak.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
GPT-5.1 can reproduce the shape of a three-act or Save the Cat structure fluently because those labels are heavily represented in its training data, but fluency with the labels is not the same as causal soundness — a common failure is an outline where beats are correctly placed at the right percentage marks of the structure but connected to each other only by sequence ('and then'), which reads as competent on a beat-sheet but flat as an actual story, since readers track causation, not percentages. Requiring a stated causal spine before any beats are generated gives the model a single throughline to check every subsequent beat against, rather than generating beats independently from the structure template and hoping they cohere — this is the same reason outlining a spine first works for human writers, but it matters more for a model, which has no persistent sense of the story's logic across a long generation unless it's been made explicit and referable. The specific instruction about the midpoint (redirect the pursuit, not just escalate it) targets a very common generic-outline failure where the model treats midpoint purely as a stakes-raising checkpoint, producing a story that's structurally correct but monotonic in tension rather than genuinely turning. The causality audit as a separate, final pass matters because self-critique performed in the same breath as generation is measurably weaker than a dedicated second pass explicitly told what to look for — asking the model to re-read its own outline hunting specifically for coincidence-driven or out-of-character transitions catches problems that generating the outline correctly the first time often doesn't.
What you get back
Spine: Elena wants to restore some version of the legal legitimacy she lost, and takes the ghostwriting job believing it lets her practice law by proxy without the risk of practicing law directly — until she finds evidence that makes staying at arm's length morally impossible, forcing her to choose between the safety of her new, diminished role and the professional identity she gave up...
Verified against
ChatGPT GPT-5.1 · 2026-08-10
Changelog
- 2026-08-10 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
