Duckdb RAG MCP Sample

🚀 DuckDB RAG MCP Sample

This is an example of embedding and vectorizing Markdown documents and using MCP for interpretation through the RAG method.

The Plamo-Embedding-1B model is used for vectorization.

🚀 Quick Start

✨ Features

• Extract and vectorize text from Markdown files.
• Perform vector search using DuckDB.
• Achieve persistent storage of vector data through Parquet files.
• Conduct vector search using MCP.

📦 Installation

Generate Vector Data

First, place the Markdown files you want to search in a specific directory. Then, run the following command to convert them into a Parquet file:

uv run main.py --directory ~/path/to/markdown/files --parquet vectors.parquet

Configure MCP

Build

Use the following command to generate a single executable file dist/server:

uv run pyinstaller --clean --strip --noconfirm --onefile server.py

Configure the MCP Client

Configure according to the client you need to use.

For example, for Claude Desktop, configure as follows: Specify the path of the previously generated file for VECTOR_PARQUET.

uv run mcp install server.py -v VECTOR_PARQUET=/path/to/vectors.parquet

Then make the following settings:

{
  "mcpServers": {
    "DuckDB-RAG-MCP-Sample": {
      "command": "/path/to/dist/server",
      "env": {
        "VECTOR_PARQUET": "/path/to/vectors.parquet"
      }
    }
  }
}

Start the Development Server

uv run mcp dev server.py

📄 License

The DuckDB RAG MCP Sample is provided under the Apache License, Version 2.0.