How to Install Fetch MCP Server in Cursor

Fetch any URL and convert web content to clean markdown for AI consumption

What Fetch MCP Server does

The simplest and most essential MCP server for web access. Fetch any public URL and receive the content as clean, AI-readable markdown — perfect for reading documentation, articles, and web pages. Features: - Fetch any HTTP/HTTPS URL - Automatic HTML-to-markdown conversion - Preserves links and code blocks - Handles JavaScript-rendered pages via optional Playwright backend - Respects robots.txt - Configurable user agent - Returns raw HTML if markdown conversion not needed The Fetch MCP is the go-to solution when your AI needs to read current documentation, follow a URL from search results, or process web content. Simpler than Puppeteer for read-only use cases. Installed via uvx (Python) — requires Python 3.10+ and uv package manager.

fetch

http

web

html

markdown

Installation steps

1Copy the config below
2Open or create .cursor/mcp.json
3Paste inside "mcpServers": {}
4Restart Cursor or reload the window

Configuration

File location: .cursor/mcp.json (project) or ~/.cursor/mcp.json (global)

.cursor/mcp.jsonjson

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

Full server details GitHub Open Config Editor

Install in other clients

🤖Claude Desktop 💻Claude Code 🏄Windsurf ✨OpenAI ♊Gemini

For maintainers

If you maintain Fetch MCP Server, add this badge to your README to show it's verified on CuratedMCP:

[![CuratedMCP Verified](https://curatedmcp.com/api/badge/fetch-mcp-server)](https://curatedmcp.com/marketplace/fetch-mcp-server)

Related servers

Brave Search MCP

Real-time web and local search using the Brave Search API — no tracking, no ads

Free

Firecrawl MCP

Advanced web scraping with JavaScript rendering, PDF support, and autonomous research agents

Free

Tavily MCP

AI-optimised web search built for agents — grounded, real-time answers with source citations

Free