Fetch MCP Server

Official

Retrieve public web pages as readable Markdown with the reference Fetch MCP server, including pagination for longer responses.

Local serverstdioPythonMIT★ 63.0k

What is the Fetch MCP Server?

The reference Fetch MCP server retrieves a URL and returns web content to an assistant. Its default conversion produces Markdown, making an ordinary documentation page easier to inspect in a text-based workflow.

Read a specific source

Give the assistant the URL of the page you want it to check and the question the page should answer. Fetch supports a maximum response length and a start position for continuing through longer content. Its raw option can return the unconverted page when that is useful.

Understand the retrieval boundary

Fetch is a content retrieval tool, not a browser session that logs in, clicks controls or runs a full user journey. Pages that depend on a signed-in browser or client-side interaction may need another integration. Public-page retrieval does not require a provider API key. The reference README documents installation through uvx mcp-server-fetch.

Setup and tool documentation. Explore related MCP servers.

Recorded tools

Check the linked publisher documentation and your connected client's tool list. Names and availability can vary with the server version and account permissions.

ToolWhat it does
fetchRetrieve a URL as Markdown or raw content, with response length and start-position options.

How to install the Fetch MCP Server

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

Configuration

Python and uv for the documented uvx installation.

Example prompts to try

  • “Fetch this documentation URL and explain the configuration option it describes.”

Frequently asked questions

Yes. The fetch tool documents max_length and start_index arguments for limiting and continuing the response.