gnosis-mcp

Overview

Gnosis MCP is a powerful documentation search server that reduces token usage by 5-10× per lookup compared to traditional methods. It supports both keyword (BM25) and semantic search using local ONNX embeddings, ensuring your data never leaves your machine. The server indexes multiple document formats including markdown, git history, and crawled websites, with built-in evaluation tools to measure retrieval quality.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• search_docs

  Search through indexed documentation with keyword or hybrid search.

  Try:Search for information about authenticationTry:Find documentation related to API setupTry:How do I configure the database connection?
• get_doc

  Retrieve the full content of a specific document by path.

  Try:Get the full content of the README fileTry:Show me the complete documentation on embeddingsTry:Retrieve the content of src/auth.py
• get_related

  Find documents related to a specific document based on relationships.

  Try:Find documents related to the authentication moduleTry:Show me documents that reference the database configurationTry:What's connected to the payment processing documentation?
• ingest

  Ingest local documentation files into the search index.

  Try:Ingest all files in the docs/ directoryTry:Add new API documentation to the indexTry:Ingest the project's source code documentation
• ingest_git

  Ingest git commit history as searchable documentation.

  Try:Ingest git history for the current repositoryTry:Add commit messages from the last 6 monthsTry:Ingest git history for the authentication module
• crawl

  Crawl and ingest documentation from a website.

  Try:Crawl the Stripe documentation websiteTry:Ingest tutorials from the FastAPI documentationTry:Fetch documentation from docs.python.org
• stats

  View statistics about the indexed documents.

  Try:Show me statistics about the indexed documentationTry:How many documents are in the index?Try:What's the total token count of all indexed documents?
• prune

  Remove stale documents from the index that no longer exist on disk.

  Try:Prune deleted documents from the indexTry:Clean up the index after removing docsTry:Show me what would be pruned without actually deleting
• embed

  Generate embeddings for documents to enable semantic search.

  Try:Generate embeddings for all indexed documentsTry:Create embeddings for the API documentationTry:Update embeddings with the latest model
• savings

  Calculate token savings from using search instead of full document retrieval.

  Try:Show me token savings from the last 7 daysTry:How many tokens have I saved by using search?Try:Calculate token savings for the last month
• eval

  Evaluate the search performance with a test dataset.

  Try:Run evaluation on the current indexTry:Test search performance with golden queriesTry:Show me Hit@K and MRR metrics
• serve

  Start the MCP server to expose tools to AI clients.

  Try:Start the MCP server with default settingsTry:Serve the documentation index on port 8000Try:Run the server with HTTP transport

Comparable tools

Installation

pip install gnosis-mcp           # or: uv tool install gnosis-mcp
gnosis-mcp ingest ./docs/        # loads docs into SQLite (auto-created)
gnosis-mcp serve                 # starts MCP server

{
  "mcpServers": {
    "docs": {
      "command": "gnosis-mcp",
      "args": ["serve"]
    }
  }
}

FAQ

What document formats does Gnosis MCP support?

It supports .md, .txt, .ipynb, .toml, .csv, .json, and optionally .rst and .pdf files.

Can I use Gnosis MCP without an internet connection?

Yes, Gnosis MCP works entirely offline with SQLite by default. It uses local ONNX embeddings for semantic search with no API key required.

Compare gnosis-mcp with