My build in public journey: creating an Ebook generator

Short summary

A solo founder built Ebookr.ai, a multi-agent ebook generator using six specialized agents (Architect, Writer, Coder, QA, Infographic, Layout Designer) orchestrated via LangChain and LangGraph. The stack is Django, Celery, Postgres with pgvector, and Playwright for PDF rendering with a custom pagination engine. The author uses Claude Code as a coding partner rather than autocomplete, describing a workflow of intent + constraints, AI implementation, test gating, and human review.

• •Ebookr.ai uses a six-agent pipeline (Architect, Writer, Coder, QA, Infographic, Layout) orchestrated with LangChain/LangGraph
• •PDF rendering via headless Chromium + Playwright with a custom pagination engine for table/list splitting
• •Solo founder workflow: Claude Code as coding partner with test-gated review cycles

Generated with AI, which can make mistakes.