BookForge

Overview

BookForge is an end-to-end non-fiction book creation pipeline designed for Amazon KDP publishing. The system uses a two-phase workflow: first, AI-driven market research and niche discovery to identify viable book topics, then chapter-by-chapter content generation with citations and KDP-formatted DOCX export.

The writing phase uses Claude for long-form content generation, producing well-structured chapters with proper citations. Perplexity handles real-time research, providing current data and sourced claims. The entire pipeline follows a structured service layer pattern, with each stage (research, outlining, writing, formatting) handled by a dedicated service.

Key Features

• Two-phase workflow: market research + content generation
• AI-driven niche discovery and competitive analysis
• Chapter-by-chapter generation with Claude for writing quality
• Real-time research integration via Perplexity API
• Proper citations and source attribution
• KDP-formatted DOCX export via python-docx
• Structured service layer pattern for each pipeline stage

Architecture

Phase 1: Research
Topic Input → Perplexity Research → Niche Analysis → Market Viability Score
                                          ↓
                                    Outline Generation

Phase 2: Writing
Outline → Chapter Generation (Claude) → Citation Insertion → DOCX Export
                    ↓                                                ↓
            Perplexity Fact-Check                              KDP Formatting
            (Real-time sources)                                (python-docx)

Phase 1 takes a topic and runs it through Perplexity for market research, analyzing competition, demand signals, and reader expectations to produce a viability score and structured outline. Phase 2 takes the outline and generates each chapter using Claude, with Perplexity providing real-time fact-checking and sourcing. The final output is a KDP-ready DOCX file with proper formatting, chapter structure, and citation pages.

Tech Stack