A modular Python pipeline for generating structured, validated, and reproducible book and audiobook content — from outline to audio output.
Creating books and audiobooks manually requires coordinating writing, formatting, narration, audio chunking, and platform constraints at every stage. Book Factory replaces that ad-hoc process with a repeatable, schema-validated pipeline that produces consistent outputs on every run.
Outline → Chapter Generation → JSON Validation →
Markdown Rendering → Audio Chunking → Book Assembly → Manifest Report
Book_Factory/
├── Book-factory-pipeline-deterministic/ # Active production version
│ └── Template-driven, schema-validated narrative engine
├── gemini/ # In development
│ └── LLM-based pipeline using Google ADK + Gemini
├── docs/
└── README.md
No external LLM dependency required.
| Feature | Description |
|---|---|
| Outline generation | Structured chapter-level planning |
| Chapter generation | Deterministic, template-driven content |
| JSON schema validation | Enforced output contracts at every stage |
| Paragraph formatting guards | Consistency across long-form content |
| Markdown rendering | Clean text output for ebook formatting |
| Book assembly | Full manuscript from modular chapters |
| Manifest + run reporting | Traceability and reproducibility per run |
- Deterministic and reproducible outputs
- Strict schema contracts between pipeline stages
- Separation of orchestration and narrative engine
- Clean artifact generation
- Extensible architecture for future LLM integration
A second pipeline using Google ADK and Gemini will be added to the same repository, demonstrating LLM-based generation while preserving the same architectural separation and schema contracts.
Python · JSON Schema · Markdown · ElevenLabs · Google ADK (planned) · Amazon KDP (target platform)
Author: Leonardo Flores
Status: Active Development
Contact: LinkedIn