MCP -unified-docs-hub

Unified Docs Hub is an MCP server that builds a massive, searchable knowledge base from curated repositories and auto-discovered GitHub projects. It exposes a set of MCP tools you can use from your MCP-compatible assistant to search, index, browse, and retrieve documentation across 170+ curated repos and thousands of docs. Key tools include unified_search for full-text documentation queries, index_repositories to control indexing and discovery, list_repositories to browse indexed sources, get_repository_docs to pull docs for a specific repository, and get_statistics to inspect database metrics. This setup enables your AI assistant to answer questions with up-to-date documentation coverage, covering categories like Trading & Finance, AI/ML, Cloud/DevOps, Web Development, and more. Use the tools within your MCP client configuration to tailor the search and indexing to your needs.

Prerequisites:\n- Python 3.8 or higher (tested with Python 3.8+)

• Git installed
• Optional: GitHub Personal Access Token (recommended for higher rate limits and private repo access)
• MCP-compatible AI assistant installed (Claude Desktop, Continue.dev, etc.)\n\nInstallation steps:\n1) Clone the repository:\nbash\ngit clone https://github.com/yourusername/unified-docs-hub.git\ncd unified-docs-hub\n\n2) Create a Python virtual environment and install dependencies:\nbash\npython3 -m venv venv\nsource venv/bin/activate # On Windows: venv\Scripts\activate\npip install -r requirements.txt\n\n3) Set up your MCP client configuration (example for Claude Desktop):\njson\n{\n "mcpServers": {\n "unified-docs-hub": {\n "command": "/path/to/unified-docs-hub/venv/bin/python",\n "args": ["/path/to/unified-docs-hub/unified_docs_hub_server.py"],\n "env": {\n "GITHUB_TOKEN": "your-github-token-here"\n }\n }\n }\n}\n\n4) Start indexing (optional—the server can index automatically):\nbash\n# Index all curated repositories (example command for one-off indexing) python -c "import asyncio; from unified_docs_hub_server import index_repositories; asyncio.run(index_repositories('smart'))"\n

Tips and caveats:\n- Environment variable GITHUB_TOKEN boosts rate limits and access; set it in your MCP client configuration as shown.\n- The system uses SQLite with FTS5 for fast full-text search across documentation. Ensure your environment supports FTS5 (default in modern SQLite builds).\n- Automated indexing updates run on a schedule (daily for curated repos, weekly for discovery). You can also trigger manual updates through the provided MCP tools.\n- If you plan to index privately or limit which repos to include, update repositories.yaml accordingly.\n- For Claude Desktop users, ensure the config path matches your OS and that the Python path points to your venv.\n- If you encounter rate limit issues with GitHub, consider using a GitHub Token and/or adjusting the discovery parameters (min_stars, count).