mcp-local-rag MCP Server

What does “mcp-local-rag” MCP Server do?

The mcp-local-rag MCP Server is a “primitive” Retrieval-Augmented Generation (RAG)-like web search Model Context Protocol (MCP) server that runs locally without requiring external APIs. Its main function is to connect AI assistants with the web as a data source, allowing large language models (LLMs) to execute web searches, fetch and embed search results, and extract relevant content—all within a privacy-respecting, local environment. The server orchestrates the process by submitting user queries to a search engine (DuckDuckGo), fetching multiple results, ranking them based on similarity using Google’s MediaPipe Text Embedder, and extracting relevant context from web pages. This enables developers and AI clients to access up-to-date web information, which can enhance workflows such as research, content creation, and question answering without relying on proprietary web APIs.

List of Prompts

No specific prompt templates are mentioned in the repository or the documentation.

List of Resources

No explicit MCP “resources” are described in the available repository content.

List of Tools

No detailed tool definitions are directly listed in the available files or documentation.

Use Cases of this MCP Server

• Real-time Web Search for LLMs: Enables AI assistants to access current web information, making outputs more accurate and up-to-date for research or news-based queries.
• Content Summarization: Allows LLMs to fetch web pages and extract relevant context, supporting summarization and fact-checking.
• Retrieval-Augmented Generation: Supports workflows where LLMs require external knowledge from the web to augment their responses, ideal for answering questions beyond their training data.
• Developer Productivity: Useful in coding assistants for searching documentation, Stack Overflow threads, or recent technical articles.
• Educational Assistance: Can aid in pulling in fresh learning resources or examples for students and educators.

How to set it up

Below are the general setup instructions for integrating the mcp-local-rag MCP Server with various MCP clients. Please adapt the configuration JSON as needed for your specific client.

Windsurf

1. Ensure you have uv (for the uvx method) or Docker installed.
2. Locate your MCP client configuration file (see here).
3. Add the following JSON to your mcpServers object:

   {
     "mcpServers": {
       "mcp-local-rag": {
         "command": "uvx",
         "args": [
           "--python=3.10",
           "--from",
           "git+https://github.com/nkapila6/mcp-local-rag",
           "mcp-local-rag"
         ]
       }
     }
   }
   

4. Save the configuration and restart Windsurf.
5. Verify the MCP server is running and accessible in your client.

Claude

1. Install uv or Docker as needed.
2. Open the MCP configuration for Claude Desktop.
3. Insert the following into your MCP server settings:

   {
     "mcpServers": {
       "mcp-local-rag": {
         "command": "uvx",
         "args": [
           "--python=3.10",
           "--from",
           "git+https://github.com/nkapila6/mcp-local-rag",
           "mcp-local-rag"
         ]
       }
     }
   }
   

4. Save and restart Claude Desktop.
5. Confirm that the “mcp-local-rag” server appears in your tools.

Cursor

1. Make sure either Docker or uv is installed.
2. Find and open the MCP server configuration file for Cursor.
3. Use the Docker configuration for improved isolation:

   {
     "mcpServers": {
       "mcp-local-rag": {
         "command": "docker",
         "args": [
           "run",
           "--rm",
           "-i",
           "--init",
           "-e",
           "DOCKER_CONTAINER=true",
           "ghcr.io/nkapila6/mcp-local-rag:latest"
         ]
       }
     }
   }
   

4. Save and restart Cursor.
5. Check the MCP server status within Cursor’s UI.

Cline

1. Install Docker or uv as required.
2. Access Cline’s MCP server configuration (refer to its documentation).
3. Add the appropriate JSON snippet (see above uvx or Docker example).
4. Save the configuration and restart Cline.
5. Ensure the server is listed among the available MCP integrations.

Securing API Keys

No external API keys are required for mcp-local-rag, but if you need to set environment variables (for Docker or other purposes), use the env object in your configuration:

{
  "mcpServers": {
    "mcp-local-rag": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "--init",
        "-e",
        "DOCKER_CONTAINER=true",
        "ghcr.io/nkapila6/mcp-local-rag:latest"
      ],
      "env": {
        "EXAMPLE_ENV_VAR": "value"
      },
      "inputs": {}
    }
  }
}

How to use this MCP inside flows

Using MCP in FlowHunt

To integrate MCP servers into your FlowHunt workflow, start by adding the MCP component to your flow and connecting it to your AI agent:

Click on the MCP component to open the configuration panel. In the system MCP configuration section, insert your MCP server details using this JSON format:

{
  "mcp-local-rag": {
    "transport": "streamable_http",
    "url": "https://yourmcpserver.example/pathtothemcp/url"
  }
}

Once configured, the AI agent is now able to use this MCP as a tool with access to all its functions and capabilities. Remember to change “mcp-local-rag” to whatever the actual name of your MCP server is and replace the URL with your own MCP server URL.

Overview

Overall, mcp-local-rag is a straightforward, privacy-respecting MCP server for web search, but lacks detail in its prompt/templates, resources, and tool specification documentation. It is easy to set up and use with major clients, but is best suited for simple web RAG use cases.

MCP Score