document-search-mcp

# Document Search MCP A Model Context Protocol (MCP) server that provides intelligent document search across multiple sources, starting with Google Drive integration. ## Overview This MCP server enables AI assistants like Claude Desktop to search and retrieve documents from connected sources. It implements the official MCP protocol and provides a clean, extensible architecture for adding new document connectors. ## Features - 🔍 **Multi-source document search** - Search across Google Drive documents, sheets, and presentations - 🔐 **OAuth 2.0 authentication** - Secure authentication with environment-based credentials - 📄 **Full content retrieval** - Get complete document content for analysis - 🔌 **Extensible plugin system** - Ready framework for custom enhancements - 🏗️ **Modular architecture** - Clean separation of connectors, models, and search orchestration ## Quick Start ### Prerequisites - Python 3.11+ - Google OAuth 2.0 credentials (for Google Drive integration) ### Installation ```bash # Install uv (recommended Python package manager) curl -LsSf https://astral.sh/uv/install.sh | sh # Clone and setup git clone <repository-url> cd document-search-mcp uv venv source .venv/bin/activate # Windows: .venv\Scripts\activate uv pip install -e ".[dev]" ``` ### Configuration 1. Set up Google OAuth credentials: ```bash export GOOGLE_CLIENT_ID="your-client-id" export GOOGLE_CLIENT_SECRET="your-client-secret" ``` 2. Configure Claude Desktop by adding to `~/.claude/mcp_servers.json`: ```json { "mcpServers": { "document-search": { "command": "document-search-mcp" } } } ``` ### Usage ```bash # Start the MCP server document-search-mcp # Or with debug logging document-search-mcp --log-level DEBUG ``` ## MCP Tools The server provides these MCP tools: - **`search_documents`** - Search across connected document sources - **`get_document_content`** - Retrieve full content from documents - **`list_sources`** - Show configured document sour…