Install in 1 click
Pick your AI client below. Each one connects to mcp.tripitaka-mcp.com/mcp — no API key, no signup, completely free.
11 tools · ~444K Pāli segments at parity with SuttaCentral · offered as Dhamma Dāna
mcp.tripitaka-mcp.com/mcpOr run it 100% offline
No hosted server, no network calls — the whole Pāli canon as a single SQLite file on your own machine. Needs Python 3.10+.
Serves 9 of the 11 tools — everything except search_semantic and search_hybrid, which need the hosted server's vector database (survey_corpus works locally in lexical-only mode; get_sutta's outline / around / pagination work fully offline too). Keyword search uses SQLite full-text (diacritic-insensitive) — it ranks results a little differently from the hosted trigram search, and the local database is a snapshot, so re-run tripitaka-mcp init to refresh it. Full hosted-vs-local comparison is in the README. Package on PyPI — works with any MCP client, and pairs with a local model (e.g. Ollama) for a fully offline AI stack.
Need help? Common issues + manual setup
Claude Desktop on Windows / nvm-using macOS — the copy-config above uses a generic npx. If Claude Desktop says "couldn't connect to server", it likely couldn't find npx on its PATH. See the full Claude Desktop walkthrough on the main page for the absolute-path + env-PATH variant.
VS Code didn't open? Make sure VS Code is installed and you've enabled MCP in settings: chat.mcp.enabled: true. VS Code Insiders users: replace vscode: with vscode-insiders: in the link target.
Behind a corporate firewall? The server speaks Streamable HTTP at https://mcp.tripitaka-mcp.com/mcp and legacy SSE at https://mcp.tripitaka-mcp.com/sse. Whitelist whichever your client uses.
Want a self-hosted copy? The full source is on GitHub with a Docker Compose setup. Free for non-commercial use; see NOTICE.md for data licences.
Discovery / browse on directories:
- Glama — score 4.4/5.0, full tool list
- Smithery — install commands for any client
- Cursor Directory · MCP.so