DocuMCP

--- sidebar_position: 1 --- # DocuMCP Introduction Welcome to the documentation for DocuMCP, a comprehensive system for intelligent code documentation generation and contextual assistance. DocuMCP consists of two complementary MCP servers designed to work together for advanced documentation workflows. ## What is DocuMCP? DocuMCP is a TypeScript implementation of the Model Context Protocol (MCP) that serves as the foundation for seamless integration between AI systems and your codebase. It functions as a **standardized communication layer** between language models (like Claude) and your code repositories. The system includes: 1. **DocuMCP Server** - The core documentation generation server with RAG capabilities 2. **DocuMCP Manager** - A supervisor server for coordinating multiple documentation agents Using Retrieval-Augmented Generation (RAG) with a local vector database, DocuMCP enhances code understanding by providing: - Automatic and contextual documentation of codebases - Semantic search and retrieval of code snippets - Integration with local coding assistants - Multi-agent coordination for large-scale documentation projects ## Project Architecture DocuMCP follows a modular monorepo architecture: ``` DocuMCP/ ├── mcp/ # Core DocuMCP server │ ├── src/ # Source code │ │ ├── services/ # Vector DB & embedding services │ │ ├── tools/ # MCP tools for documentation │ │ ├── helper/ # Utility functions │ │ └── schemas/ # Data schemas │ └── ... # Configuration files ├── manager/ # DocuMCP Manager server │ ├── src/ # Source code │ │ ├── services/ # Agent management services │ │ ├── tools/ # Supervisor & coordination tools │ │ └── ... # Shared components with mcp/ │ └── ... # Configuration files ├── qdrant/ # Vector database (Qdrant) │ └── docker-compose.yml ├── chromadb/ # Alternative vector database │ └── docker-compose.yml └── docusaurus/ # Documentation website ``` ## Getting Started To get started with DocuMCP, see our [Getting Started](./getting-started.md) guide. ## Core Concepts The project is built around several key concepts: - [Model Context Protocol](./fundamentals/mcp.md) - The foundation for AI integration - [Retrieval-Augmented Generation](./fundamentals/rag.md) - Enhancing responses with codebase context - [Local Vector Database](./fundamentals/vector-db.md) - Efficient semantic search