MODULAR-RAG-MCP-SERVER

Level ALocal

A modular RAG (Retrieval-Augmented Generation) system with MCP Server architecture. Using Skill to make AI follow each step of the spec and complete the code 100% by AI.

Overview

A modular RAG service framework that exposes tools to AI assistants like Copilot and Claude via the MCP protocol. It covers ingestion, hybrid search (BM25 plus dense vectors, RRF fusion, reranking), multimodal image captioning, Ragas-based evaluation, and a Streamlit dashboard, with pluggable components throughout. It also serves as a hands-on project for learning and interview prep.

⭐ 1.2kSearch & InfoPythonRepo ↗Updated 2026-03-10

Installation

MCP client config (paste into mcp.json)

{
  "mcpServers": {
    "jerry-ai-dev-modular-rag-mcp-server": {
      "command": "uv",
      "args": [
        "--directory",
        "<PATH_TO>/MODULAR-RAG-MCP-SERVER",
        "run",
        "python",
        "-c",
        "from src.mcp_server.server import main; raise SystemExit(main())"
      ]
    }
  }
}

Agent install prompt

Copy and paste into Cursor / Claude / other agents to install this MCP for you

Please read this MCP repository and install it into my current environment:
https://github.com/jerry-ai-dev/MODULAR-RAG-MCP-SERVER

Known working client config (can write into mcp.json):
{
  "mcpServers": {
    "jerry-ai-dev-modular-rag-mcp-server": {
      "command": "uv",
      "args": [
        "--directory",
        "<PATH_TO>/MODULAR-RAG-MCP-SERVER",
        "run",
        "python",
        "-c",
        "from src.mcp_server.server import main; raise SystemExit(main())"
      ]
    }
  }
}

Please: 1) check README for deps and env vars; 2) write config for my client; 3) if secrets are missing, tell me what to fill — do not invent values.

Related servers