Fetch MCP

What is Fetch MCP Server?

Fetch MCP Server is a specialized implementation of the Model Context Protocol (MCP) designed to enable Large Language Models (LLMs) to retrieve and process web content efficiently. It fetches URLs from the internet, converts HTML content to Markdown for easier consumption, and allows models to read webpages in chunks using the start_index argument. This ensures that models can access the exact information they need without retrieving unnecessary data.

Key Features

• Web Content Extraction: Fetches URLs and extracts content as Markdown for simplified processing.
• Chunked Content Reading: Supports reading webpages in chunks using start_index to specify the starting point of content extraction.
• Flexible Configuration: Customizable user-agent headers, proxy support, and robots.txt compliance settings.
• Efficient Content Handling: Truncates responses to a specified max_length to avoid overwhelming models with excessive data.
• Raw Content Option: Retrieves raw HTML content without Markdown conversion when needed.
• Cross-Platform Compatibility: Works seamlessly with various MCP clients like Claude.app and supports installation via multiple methods (uv, pip, Docker).

How to Use Fetch MCP Server

Installation

1. Using uv (Recommended):

   • No specific installation is needed. Use uvx to directly run the server:

     uvx mcp-server-fetch
     

2. Using PIP:

   • Install the package via pip:

     pip install mcp-server-fetch
     

   • Run the server as a script:

     python -m mcp_server_fetch
     

3. Using Docker:

   • Run the server using Docker:

     docker run -i --rm mcp/fetch
     

Configuration for Claude.app

• Using uvx:

  "mcpServers": {
    "fetch": {
      "command": "uvx",
      "args": ["mcp-server-fetch"]
    }
  }