Skip to main content
Glama

Obsidian MCP Server

License: MIT Python 3.11+ MCP Compatible Obsidian Integration Claude Desktop Custom Skills

An advanced MCP (Model Context Protocol) server that turns your Obsidian vault into a dynamic brain for your AI (Claude Desktop, Cursor, Claude Code, Cline, etc.). Much more than a file reader: it is an ecosystem of tools for knowledge management, workflow automation, and semantic analysis.


Features

Tool Ecosystem (30+)

The server exposes a wide variety of tools categorized by function:

  • Navigation: Intelligent listing, recursive reading, and advanced searching.

  • Creation and Editing: Automatic template usage, location suggestions, and non-destructive editing preserving metadata (frontmatter/YAML).

  • Analysis and Quality: Vault statistics, tag synchronization with the official registry, and integrity checks.

  • Graphs and Connections: Backlink analysis, orphan note detection, and local graph visualization.

  • Skills (Agents): Dynamic loading of AI personalities/roles from your vault ({vault}/.agent/skills/).

  • Semantic Search (RAG): Meaning-based searches, suggestions for non-obvious connections, and vector indexing.

  • YouTube: Extraction of transcripts to feed your knowledge base.

Built-in Intelligence

  • Vault-Agnostic Architecture: Independent of your folder structure; it uses intelligent auto-detection to find templates and resources.

  • Security: Strict protection of sensitive folders via .forbidden_paths and vault privacy configurations.

  • Customizable Skills: Define specific AI roles directly within your vault (.agent/skills/) for specialized tasks.


Quick Setup

Prerequisites

  • Python 3.11+

  • uv (Recommended)

Steps

  1. Clone:

    git clone https://github.com/Vasallo94/obsidian-mcp-server.git cd obsidian-mcp-server
  2. Install:

    make install # For semantic search capabilities (RAG): pip install "obsidian-mcp-server[rag]"
  3. Configure:

    cp .env.example .env # Edit .env with the absolute path to your Obsidian vault

Usage

The MCP server connects to compatible clients using the uv run obsidian-mcp-server command. Below are setup instructions for popular AI clients.

Claude Desktop Integration

Add the following to your claude_desktop_config.json:

{ "mcpServers": { "obsidian": { "command": "uv", "args": ["run", "obsidian-mcp-server"], "env": { "OBSIDIAN_VAULT_PATH": "/Absolute/Path/To/Your/Vault" } } } }

Cursor & Cline Integration

For other clients like Cursor or Cline, you can add it as a new MCP server in their settings panel by specifying:

  • Type: command

  • Command: uv run obsidian-mcp-server

  • Ensure the OBSIDIAN_VAULT_PATH environment variable is available to the instance running the command.


Technical Documentation

To dive deeper into how the server works and how to customize it, check our detailed guides located in the docs/ folder:

  1. Architecture: Modular structure and data flow of the project.

  2. Tool Reference: Complete list of available MCP tools and their parameters.

  3. Server Configuration: Guide on environment variables and technical configuration.

  4. Agent Setup: How to organize your vault (.agent/) with skills and contextual rules.

  5. Semantic Search (RAG): Deep dive into vector indexing and RAG mechanics.

  6. Future Roadmap: Planned improvements and next steps for the server.


Development & Quality

Command

Description

make test

Run the test suite (pytest)

make lint

Run static checks (Ruff + Mypy + Pylint)

make format

Automatically format code

make dev

Run the MCP inspector for live testing


License

This project is licensed under the MIT License.

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Vasallo94/obsidian-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server