How do I give Claude Desktop web access?

Give Claude Desktop web access by installing an MCP server that exposes a fetch tool, then adding it to your claude_desktop_config.json. The fastest path is AgentFetch: add {"mcpServers": {"agentfetch": {"command": "npx", "args": ["-y", "@agentfetch/mcp"]}}} to your config (located at ~/Library/Application Support/Claude/claude_desktop_config.json on macOS or %APPDATA%\Claude\claude_desktop_config.json on Windows), restart Claude Desktop, and the model will see fetch_url and extract_json tools in its toolbox. You can then ask Claude things like "fetch the top three TechCrunch headlines and summarize them" and it will call the tool autonomously. Alternatives include Anthropic's reference mcp-server-fetch (basic, no retry/extraction logic) and mcp-firecrawl (paid, wraps Firecrawl's API). Claude Desktop does not have native web browsing; the MCP server is the canonical extension point. After installation, check the hammer icon in the bottom-right of the chat input — it should show the new tools as available. If they don't appear, check the MCP logs at ~/Library/Logs/Claude/mcp.log.