Obsidian Vault RAG

Obsidian Vault RAG allows you to import your Obsidian markdown vault into a local LanceDB vector database and expose it via the Model Context Protocol (MCP) for AI agents like Claude.

Features

• Local RAG: No data leaves your machine; embeddings are stored locally in LanceDB.
• Smart Ingestion: Splits notes into chunks and generates embeddings.
• PDF Support: Converts research papers and PDFs to Markdown using Docling.
• Claude Integration: Connect your vault directly to Claude Desktop.

Installation

Clone the repository and install the package:

# Using pip
pip install -e .

# OR using uv (recommended for speed)
uv pip install .

Quick Start

1. Configure: Run the setup wizard to link your vault.

   obsidian config
   

2. Ingest: Index your notes into the database.

   obsidian lance
   

3. Serve: Start the MCP server (or connect Claude, see below).

   obsidian serve
   

Manual File Creation

To manually add files to your vault, you can use an Obsidian Template to automatically insert the required metadata.

Create a template file in your Obsidian templates folder with the following content:

---
id: {{date:YYYYMMDDHHmm}}
title: {{title}}
authors: []
type: note
status: active
created: {{date:YYYY-MM-DD}}
tags: []
source: "personal"
---

# {{title}}

Your content here...

Documentation

• Usage Guide: Detailed commands for ingestion, PDF conversion, and serving.
• Configuration: Settings, environment variables, and logging.

Connect to Claude Desktop

To chat with your notes in Claude Desktop, add the server to your configuration file:

• Mac: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

Add the following entry (replace /path/to/your/python with your actual Python executable path):

{
  "mcpServers": {
    "obsidian-vault": {
      "command": "/absolute/path/to/project/.venv/bin/python3",
      "args": [
        "-m",
        "obsidian.cli",
        "serve"
      ]
    }
  }
}

Tip: You can find your python path by running which python (Mac/Linux) or where python (Windows) inside your project environment.

License

MIT