AI Book Agent MCP Server

# AI Book Agent An MCP (Model Context Protocol) server that provides ML textbook knowledge to Claude Desktop using fully local models - no external API keys required. ## Project Overview This project enables Claude Desktop to access authoritative ML textbook knowledge through a local MCP server. Built with privacy in mind, it uses local models for both embeddings and text generation, ensuring your queries and book content never leave your machine. ## Problem Statement - Organizations need accurate ML documentation based on established best practices - Authoritative textbooks contain valuable knowledge but are difficult to search through - External AI APIs raise privacy concerns for sensitive projects - Need for source-grounded responses with proper citations ## Solution An MCP server that: 1. **Processes** ML textbooks locally (EPUB format) 2. **Indexes** content using local sentence transformers 3. **Provides** semantic search through MCP tools 4. **Generates** responses using local Ollama models 5. **Integrates** seamlessly with Claude Desktop ## Core Features - **MCP Integration**: Works directly with Claude Desktop via Model Context Protocol - **Local Models**: Sentence transformers + Ollama (no API keys needed) - **Privacy-First**: All processing happens locally on your machine - **Source Attribution**: Always provides citations from source books - **Fast Setup**: Automated installation with UV package manager - **Extensible**: Easy to add new books and customize behavior ## Use Cases 1. **Technical Documentation**: Generate docs grounded in textbook best practices 2. **Knowledge Lookup**: Answer ML questions with authoritative sources 3. **Concept Explanation**: Get detailed explanations from multiple textbooks 4. **Best Practices**: Access established methodologies from expert authors ## Current Status - ✅ MCP server with 4 core tools - ✅ Local EPUB processing and vector indexing - ✅ Sentence transformers for embeddings - ✅ Ollama integration for text generation - ✅ ChromaDB vector storage - ✅ Automated setup with UV ## Technology Stack - **Language**: Python 3.11+ - **MCP Framework**: Official Python SDK - **Embeddings**: sentence-transformers (local) - **Generation**: Ollama + Qwen 2.5 7B (local) - **Vector Store**: ChromaDB - **Package Management**: UV ## Getting Started ### Quick Setup ```bash # Clone the repository git clone <repository-url> cd ai-book-agent # Run automated setup (installs everything) ./setup-uv.sh ``` ### Manual Setup ```bash # Install UV package manager curl -LsSf https://astral.sh/uv/install.sh | sh # Create virtual environment uv venv --python 3.11 source .venv/bin/activate # Install dependencies uv pip sync requirements.txt # Install Ollama curl -fsSL https://ollama.ai/install.sh | sh ollama pull qwen2.5:7b # Process books python scripts/index_books.py ``` ### Configure Claude Desktop Add to Claude Desktop config: ```json { "mcpServers": { "ai-book-agent": { "command": "/path/to/ai-book-agent/.venv/bin/python", "args": ["/path/to/ai-book-agent/server.py"] } } } ``` ## Project Structure ``` ai-book-agent/ ├── src/ # Core MCP server code ├── data/ # EPUB files and processed data ├── scripts/ # Setup and utility scripts ├── config.yaml # Single configuration file ├── server.py # MCP server entry point ├── setup-uv.sh # Automated setup script └── pyproject.toml # Modern Python project config ``` ## MCP Tools The server provides these tools to Claude: 1. **searchBooks**: Semantic search across all indexed books 2. **getChapterContent**: Retrieve specific book sections 3. **generateSection**: Create documentation from sources 4. **citeSources**: Generate proper citations ## Contributing See [USER_GUIDE_STANDALONE.md](USER_GUIDE_STANDALONE.md) for detailed instructions on: - Adding new books to the system - Extending the MCP server - Testing and validation - Best practices for development ## Requirements - **OS**: macOS, Linux, or Windows - **Python**: 3.11+ - **Memory**: 8GB RAM minimum - **Storage**: 10GB free space ## License [Your license here] ## Acknowledgments Built to leverage authoritative ML textbooks including: - "Designing Machine Learning Systems" by Chip Huyen - [Additional books as you add them]