Does Claude Desktop support MCP tools on both macOS and Windows?

Yes. The config file path differs by platform, but the setup steps are the same. macOS uses ~/Library/Application Support/Claude/claude_desktop_config.json and Windows uses %APPDATA%\Claude\claude_desktop_config.json.

Why do the tools not appear after I edit the config?

The most common cause is invalid JSON, such as a trailing comma. Another cause is using a relative path or ~ in the command field. Use the absolute path from which crawler-mcp. Quit Claude Desktop completely and reopen.

Can I use crawler-mcp in Claude Desktop without installing the CLI?

Yes. The MCP server is a separate binary. You only need to run the install-mcp.sh script, which places crawler-mcp in ~/.crawler/bin/. The main CLI is optional.

How to Crawl a Website with Claude Desktop

Claude Desktop can run MCP tools during a conversation. Once crawler-mcp is connected, you can ask Claude to crawl a site, fetch a page, or map out a site’s structure - all through natural language.

This guide shows how to set it up and start crawling.

Step 1: Install crawler-mcp

Run the install script:

curl -fsSL https://install.crawler.sh/install-mcp.sh | sh

This downloads the correct binary for your platform to ~/.crawler/bin/crawler-mcp.

For more detail, see the installation guide.

Step 2: Wire it into Claude Desktop

Edit your Claude Desktop config file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Add the server entry:

{
  "mcpServers": {
    "crawler-sh": {
      "command": "/Users/you/.crawler/bin/crawler-mcp",
      "args": [],
      "env": {}
    }
  }
}

Use the absolute path from which crawler-mcp. Quit Claude Desktop completely and reopen.

For full setup details, see the Claude Desktop docs.

Step 3: Crawl a website

Start a new conversation and ask:

Crawl https://example.com to depth 2 with crawler-sh and tell me what the site is about.

Claude will request permission to call crawl_site, then summarise the returned pages.

Step 4: Fetch a single page

For one-off lookups:

Use crawler-sh to fetch https://example.com/about and summarise the page.

Claude calls fetch_page and reasons over the Markdown response.