Your AI doesn't know
the APIs you shipped last month.

BookLib detects post-training knowledge gaps in your dependencies and fixes them automatically — delivering context via MCP to 14 AI tools.

$npm install -g @booklib/core
booklib analyze on vercel/ai-chatbot

Real output on vercel/ai-chatbot — 274 post-training APIs across 158 files

What It Does

Gap Detection
10
Package ecosystems
Scans npm, PyPI, Maven, Cargo, and 6 more — finds every post-training dependency in your code
Runtime Injection
MCP
Pre/PostToolUse hooks
3-10 lines of context injected before each AI edit via a pre-computed context map
AI Tools
14
Auto-configured
Claude, Cursor, Copilot, Gemini, Codex, Windsurf, Roo Code, Goose, Zed, Continue, and more
Expert Skills
24
From canonical books
Effective Java, Clean Code, DDD, and 21 more — distilled into actionable principles

How it works

BookLib detects what your AI doesn't know, fixes it automatically, and delivers context via MCP and runtime hooks. Knowledge flows to your AI as it writes code.

  • ✓ Post-training gap detection — 10 ecosystems, cross-referenced with source code
  • ✓ Runtime injection — PreToolUse/PostToolUse hooks powered by context maps
  • ✓ Auto-resolution — Context7, GitHub releases, web docs fetched automatically
  • ✓ Hybrid search — BM25 + vector + cross-encoder reranking
  • ✓ Processing modes — fast (BM25), local (Ollama), or cloud AI
  • ✓ 14 AI tools — Claude, Cursor, Copilot, Gemini, Codex, Windsurf, and more

Explore Skills

Filter by profile or search for a specific book

All Skills

Community Skills

Open source skills from the community — GitHub repos, registries, and npm packages.

Reviewer Agents

Global Rules