MCP.so
登录
服务器

Wandering RAG

@wandermyz

A CLI tool for personal RAG that retrieves data from Notion, Obsidian, Apple Notes, etc., stored in Qdrant and exposed as an MCP server.

概览

What is Wandering RAG?

A CLI tool for personal RAG that retrieves data from Notion, Obsidian, Apple Notes, and other sources, stores it in Qdrant, and exposes it as an MCP server (e.g., for use with Claude Desktop).

How to use Wandering RAG?

Install with uv pip install -e ., run a Qdrant server using the provided Docker Compose file, copy .env.example to .env, and specify your Markdown folders or Obsidian vaults. Use the CLI subcommands: wandering-rag md index for Markdown, wandering-rag notion index (work in progress), and wandering-rag mcp run-server to start the MCP server. Configure Claude Desktop by adding the server to its MCP configuration JSON.

Key features of Wandering RAG

  • CLI‑based personal RAG system
  • Indexes Markdown from folders or Obsidian vaults
  • Notion indexing (work in progress)
  • Stores indexed data in Qdrant vector database
  • Exposes a Model Context Protocol (MCP) server

Use cases of Wandering RAG

  • Query personal notes (e.g., “When did I adopt my cat?”) through Claude Desktop
  • Combine data from Obsidian, Markdown, and Notion in a single retrieval pipeline
  • Run a local, self‑hosted RAG system without external cloud dependencies

FAQ from Wandering RAG

What dependencies are required?

You need a running Qdrant server (a Docker Compose file is provided) and a .env file with your configuration (e.g., Markdown folder paths).

How do I configure Wandering RAG for Claude Desktop?

Add an entry to the mcpServers object in Claude Desktop’s config, pointing the command to your checkout’s wandering-rag script with arguments ["mcp", "run-server"] and setting the PATH environment variable to include ~/.local/bin.

Is Notion indexing fully supported?

Notion commands are listed as “WIP” (work in progress); only notion index is available, and its maturity is not stated.

Where is the indexed data stored?

Data is stored in Qdrant, a vector database that runs locally via Docker.

What file formats can be indexed?

The README explicitly shows support for Markdown files (via md index) and mentions Obsidian vaults (which are Markdown‑based). No other formats are confirmed as working.

标签

来自「记忆与知识」的更多内容