Valery Perelygin did not need a chat with a model. He needed a shop floor. We built the full-cycle Neuro-Publisher AI platform: from lore to volume.
What shipped: a graph lore-book, fractal plot, an AI screenwriter with dynamic context, style fine-tune, SVP psychotypes, a trend watcher with product placement, and a chronicler agent.
Modules
A graph lore-book holds entities. Fractal plot assembles arcs. The screenwriter works with dynamic context. Fine-tune holds the author’s voice. SVP psychotypes drive behaviour. A trend watcher suggests placement. A chronicler agent keeps the universe from drifting.
How the client earns
Pace: from one volume in two weeks to one volume a week. Royalties and product placement. Contract €15,000, 20 weeks / 4 stages; €15,000 returned via 25% of SaaS profit.
The twist
Only entities of the current location go into the LLM. Tokens drop 8×, and plot holes do not appear from “the whole universe in the prompt”.