So I love the idea of Dramatica theory as an anchor to the story, but what produces a lot of pain is essentially the process of adhering to its tenets and enumerations from an initial thought. Not every writer is essentially aware of the entire process at the start, and that feels kind of normal.
I’m building a system to use an LLM to basically just have a natural discourse with the user to build the storyform itself, and then step by step (chapter by chapter) literally evaluate for deviations or additions, surface to the user and decide on an action plan.
Making it directional like that also means that in theory, at the end, you should be able to compare the manuscript to the storyform. I have a system now that uses Gemini for a three party review:
Unknowing reader - What would someone without knowledge of the source material feel / think about it after reading it
Storyform evaluation - After reading the story, did we adhere or deviate from our storyfrom components?
Consistency of tone (character and author) - Were we mostly consistent in representation of characters and general prose
The storyform evaluation is mostly useful to me as a pre-editing handoff step. Don’t want to chuck a bunch of money for an editor to be like “I feel like your main character started down this path and that never resolved”
What you’re describing is actually already a core capability of the Dramatica Narrative Platform, particularly as it’s evolved over the last several months.
Storyform evaluation — comparing the manuscript itself against underlying structure to surface deviations, unresolved throughlines, or unintended arguments — is built directly into the platform workflow. It’s explicitly designed for iterative discovery and post-draft diagnostics, not for requiring full awareness up front.
Over the past six months, the platform has also seen significant upgrades to its structural algorithms, alongside new theory work focused on how narrative intent is expressed in text rather than merely predefined. That’s pushed Dramatica even further toward the exact use case you’re outlining: directional guidance, deviation detection, and pre-editorial sanity checks.
You might also be interested in the NCP (Narrative Context Protocol), an open-source standard we’ve been developing for conveying narrative intent between systems. It’s designed specifically to support interoperability around story structure, intent, and evaluation — including LLM-driven workflows.
Not sure how much of the recent platform and theory updates you’ve had a chance to see, but there’s a lot of overlap here with what’s already shipping.