Technical Documentation
Everything you need to understand how Writers Factory works under the hood. This is the system architecture that powers AI-assisted novel writing.
The Two-Stream Approach
The pedagogical foundation of Writers Factory—separating research from story to enable creative discovery.
The Intersection
How skeleton meets research—the Connection Point workflow
The Two-Track Architecture
The separation of Researcher (Kitchen) and Architect (Meal)
Writing Stages
The four stages of the Writers Factory workflow.
Research Preparation
What you do before the course starts—building your research in NotebookLM.
NotebookLM Deepening
Enrich your story with research ingredients
Deepening Prompts →
Copy-paste prompts for prose extraction
Genre Calibration
Teaching the system your genre's rules
Concept Generation
Turning ingredients into 5 novel concepts
Research Notebook Templates
Seven specialized notebooks for building your research library in NotebookLM.
The Arena
Competition, sports, and conflict settings
The Speculation
Science fiction and fantasy worldbuilding
Beliefs & Worldviews
Philosophy, religion, and ideology themes
Literary Roots
Literary traditions and influences
The Voice
Prose styles and narrative voices
Visual Language
Cinematic and visual aesthetics
The Rabbit Hole
Deep research obsessions
System Overview
Understand how Writers Factory works under the hood.
Context Engineering
The new prompt engineering—it's all about context
GraphRAG
Knowledge graphs for narrative consistency
Data Architecture
Three-system architecture: Research + KB + Story
LLM Models
Multi-model orchestration and routing
Technical Reference
Deep dives for Architect Track students who want to understand the implementation.
Scene Scaffolding
How the scaffold prompt is assembled from multiple sources
GraphRAG Architecture
Knowledge graphs for narrative consistency
Research Pipeline
Three-system architecture and data flow
Agent Instructions
How to communicate with AI agents effectively
LLM Orchestration
Multi-model routing and capabilities
Anti-Patterns
Common mistakes and how to avoid them
Source code access is available to Architect Track participants.
ISP Course Results
Findings from the January 2026 ISP course — what we learned generating 73,000 words across 4 manuscripts.
Scene Scaffolding
How the scaffold prompt is assembled from multiple sources
Context Engineering
Why context matters more than the model you choose