Gnosis

Provides a searchable knowledge base from Markdown docs for MCP-compatible AI agents.

python

GitHub Stars

python

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "nicholasglazer-gnosis-mcp": {
      "command": "gnosis-mcp",
      "args": [
        "serve"
      ],
      "env": {
        "GNOSIS_MCP_SCHEMA": "public",
        "GNOSIS_MCP_BACKEND": "auto",
        "GNOSIS_MCP_WRITABLE": "false",
        "GNOSIS_MCP_TRANSPORT": "stdio",
        "GNOSIS_MCP_CHUNKS_TABLE": "documentation_chunks",
        "GNOSIS_MCP_DATABASE_URL": "postgresql://user:pass@localhost:5432/mydb",
        "GNOSIS_MCP_EMBEDDING_DIM": "1536"
      }
    }
  }
}

Gnosis MCP lets you turn a folder of Markdown docs into a searchable knowledge base that any MCP-compatible AI agent can query. It keeps your docs current, reduces hallucinations, and supports both keyword and semantic search with optional local embeddings.

How to use

You run the MCP server locally and connect your AI agent or editor to it. Start by ingesting your Markdown documents, then run the server to begin searching and querying your docs. Your agent can call tools like search_docs to find relevant passages, get_doc to retrieve full documents, and get_related to discover linked content. If you enable embeddings, you can perform semantic searches in addition to keyword matching.

How to install

Prerequisites: Python 3.8+ and a working Python environment. You may also use a lightweight local execution method if you prefer quick trials.

# Create and activate a virtual environment (recommended)
python -m venv .venv
source .venv/bin/activate   # on Unix/macOS
.\.venv\Scripts\activate.bat  # on Windows

# Install the MCP server package
pip install gnosis-mcp

# Ingest your docs and start the server
gnosis-mcp ingest ./docs/
gnosis-mcp serve

Notes on usage and patterns

Ingest loads Markdown files into a SQLite database by default. You can add embeddings later if you want semantic search.
The server exposes 6 tools and 3 resources over MCP. Your agent will call these tools automatically when it needs information from your docs.
You can watch for file changes to re-ingest automatically during development by running the server with watch mode.

Available tools

search_docs

Search documents by keyword or semantic similarity when embeddings are enabled.

get_doc

Retrieve the full content of a document by its path.

get_related

Find documents related to a given document or topic.

upsert_doc

Create or replace a document in the knowledge base.

delete_doc

Remove a document and its associated chunks from the knowledge base.

update_metadata

Modify a document's metadata such as title, category, or tags.

Built by

VeilStrat

AI signals for GTM teams