Gnosis

v0.13.3Knowledge & Memorystable

Zero-config knowledge base for AI coding agents. Loads your markdown docs into a searchable database and exposes them as MCP tools — search, read, and manage documentation without leaving your editor. Works instantly with SQLite (no setup), upgrades

aiclaudedeveloper-toolsdocumentationknowledge-base
Share:
23
Stars
0
Downloads
0
Weekly
0/5

What is Gnosis?

Gnosis is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to zero-config knowledge base for ai coding agents. loads your markdown docs into a searchable database and exposes them as mcp tools — search, read, and manage documentation without leaving your editor....

Zero-config knowledge base for AI coding agents. Loads your markdown docs into a searchable database and exposes them as MCP tools — search, read, and manage documentation without leaving your editor. Works instantly with SQLite (no setup), upgrades

This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • Zero-config knowledge base for AI coding agents. Loads your

Use Cases

Knowledge base creation
Markdown documentation indexing
Semantic search for AI
nicholasglazer

Maintainer

LicenseMIT License
Languagepython
Versionv0.13.3
UpdatedMay 21, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

PIP

pip install gnosis-mcp

Manual Installation

pip install gnosis-mcp

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Gnosis

Gnosis MCP is a zero-configuration knowledge base server that ingests your local Markdown documentation, code notes, and other text files into a searchable SQLite (or PostgreSQL) database and exposes them to AI coding assistants as MCP tools. It supports hybrid keyword and semantic search using BM25 plus local ONNX embeddings, enabling AI agents to retrieve precise, token-efficient excerpts from large documentation sets instead of reading entire files. Developers use it to keep project wikis, architecture notes, and API references instantly searchable from Claude, Cursor, or any MCP-compatible client.

Prerequisites

  • Python 3.9+ with pip or uv installed
  • Documentation files in Markdown, text, JSON, TOML, or CSV format to ingest
  • Optional: PostgreSQL with pgvector extension for multi-user or large-scale deployments
  • An MCP-compatible client such as Claude Desktop, Cursor, Windsurf, or VS Code with Copilot
1

Install gnosis-mcp

Install the base package with pip or uv. Add optional extras for semantic search, PostgreSQL, or web crawling support.

# Basic install (SQLite, keyword search)
pip install gnosis-mcp

# With semantic/hybrid search
pip install "gnosis-mcp[embeddings]"

# With PostgreSQL support
pip install "gnosis-mcp[postgres]"

# Via uv
uv tool install gnosis-mcp
2

Ingest your documentation

Run the ingest command pointing at your docs directory. Gnosis walks the directory, chunks documents, and stores them in the local SQLite database. Use --embed to also generate semantic embeddings.

gnosis-mcp ingest ./docs/

# With embeddings for hybrid search
gnosis-mcp ingest ./docs/ --embed

# Prune deleted files and re-ingest
gnosis-mcp ingest ./docs/ --embed --prune
3

Test search from the command line

Verify the index is populated by running a test search before connecting your MCP client.

gnosis-mcp search "authentication flow"
gnosis-mcp stats
4

Configure your MCP client

Add the gnosis server to your Claude Desktop or editor MCP configuration. The server runs gnosis-mcp serve which starts the MCP stdio transport.

{
  "mcpServers": {
    "gnosis": {
      "command": "gnosis-mcp",
      "args": ["serve"]
    }
  }
}
5

Optionally enable write access and configure the environment

To allow Claude to create or update documents in the knowledge base, set GNOSIS_MCP_WRITABLE=true. Configure the database URL or embedding provider as needed.

{
  "mcpServers": {
    "gnosis": {
      "command": "gnosis-mcp",
      "args": ["serve"],
      "env": {
        "GNOSIS_MCP_WRITABLE": "true",
        "GNOSIS_MCP_EMBED_PROVIDER": "local",
        "GNOSIS_MCP_DATABASE_URL": "sqlite:///gnosis.db"
      }
    }
  }
}

Gnosis Examples

Client configuration

Claude Desktop config for gnosis-mcp with local embeddings and write access enabled.

{
  "mcpServers": {
    "gnosis": {
      "command": "gnosis-mcp",
      "args": ["serve"],
      "env": {
        "GNOSIS_MCP_WRITABLE": "true",
        "GNOSIS_MCP_EMBED_PROVIDER": "local",
        "GNOSIS_MCP_RERANK_ENABLED": "true",
        "GNOSIS_MCP_COLLAPSE_BY_DOC": "true"
      }
    }
  }
}

Prompts to try

Example prompts for searching and managing documentation with gnosis-mcp.

- "Search my docs for anything related to database connection pooling"
- "Get the full contents of the architecture/overview.md document"
- "Find documents related to the authentication module and show linked pages"
- "What does our onboarding guide say about environment variable setup?"
- "Add a new document summarizing what we just discussed about the API rate limits"

Troubleshooting Gnosis

gnosis-mcp command not found after pip install

Ensure your Python scripts directory is on your PATH. With uv tool install, run uv tool update-shell to add the tool bin directory to your PATH. On macOS you may need to add ~/.local/bin to your shell profile.

Search returns no results after ingest

Run gnosis-mcp stats to confirm documents were indexed. If the count is 0, check that the directory path passed to ingest is correct and contains supported file types (.md, .txt, .json, .toml, .csv). Re-run gnosis-mcp ingest ./docs/ with verbose output.

Embedding generation is very slow on first run

The first ingest with --embed downloads the ONNX embedding model (~100MB). This is a one-time download. Subsequent ingests use the cached model and are significantly faster.

Frequently Asked Questions about Gnosis

What is Gnosis?

Gnosis is a Model Context Protocol (MCP) server that zero-config knowledge base for ai coding agents. loads your markdown docs into a searchable database and exposes them as mcp tools — search, read, and manage documentation without leaving your editor. works instantly with sqlite (no setup), upgrades It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Gnosis?

Install via pip with: pip install gnosis-mcp. Then configure your AI client to connect to this MCP server.

Which AI clients work with Gnosis?

Gnosis works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Gnosis free to use?

Yes, Gnosis is open source and available under the MIT License license. You can use it freely in both personal and commercial projects.

Browse More Knowledge & Memory MCP Servers

Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "gnosis": { "command": "pip", "args": ["install", "gnosis-mcp"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Gnosis?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides