Linkup

The Linkup MCP Server lets your AI assistants perform real-time web searches and fetch content from web pages. It connects to Linkup’s API to deliver natural language search results and page content, enabling up-to-date information access and extraction from trusted sources on demand.

How to use

Connect with an MCP client of your choice (Cursor, Visual Studio Code, Claude Code, or another MCP-compatible client). Configure either the remote HTTP MCP endpoint or a local stdio-based instance, then ask your AI agent to search the web or fetch webpage content.

Practical usage patterns include: asking for current news, weather, or stock data via real-time web search; fetching article text or documentation from specific web pages; and running deep research that analyzes multiple sources. Use full questions to get the most accurate results and, when necessary, enable JavaScript rendering for pages that require client-side content generation.

When you need quick facts, use standard search for direct answers. For more complex, source-wide research, use deep search to compare information across multiple sites.

How to install

Prerequisites: Node.js v18 or newer and npm. You will also need an API key from Linkup to access the MCP server.

Choose your installation path: remote HTTP MCP server, or local stdio MCP server. Use the commands shown in the setup options below to configure your MCP client.

Install and run using the HTTP remote endpoint

Configure your MCP client to connect to the remote Linkup MCP server. The remote endpoint uses the HTTP transport and requires your API key.

{
  "mcpServers": {
    "linkup": {
      "type": "http",
      "name": "linkup",
      "url": "https://mcp.linkup.so/mcp?apiKey=LINKUP_API_KEY",
      "args": []
    }
  }
}

Install and run using the local stdio transport

You can run a local MCP server instance via the stdio transport using a runtime command. The example uses npx to start the server with your API key.

{
  "mcpServers": {
    "linkup": {
      "type": "stdio",
      "name": "linkup",
      "command": "npx",
      "args": [
        "-y",
        "linkup-mcp-server",
        "apiKey=LINKUP_API_KEY"
      ],
      "env": []
    }
  }
}

Using the client configuration snippets

You can copy the provided configuration blocks into your MCP client configuration file. The HTTP variant connects to the remote endpoint, while the stdio variant starts a local process that exposes the MCP interface to your client.

Additional setup notes

If your client supports OAuth or a similar authentication flow, you can reference the remote MCP server URL directly. For local stdio runs, ensure the API key is supplied as shown in the command arguments.

Notes on usage and capabilities

The MCP endpoints provide two main capabilities: search the web in real time and fetch content from pages. You can specify search depth as standard for direct-answers queries or deep for more involved research that aggregates information from multiple sources.

Security considerations

Keep your API key secure. Do not share the keys embedded in configuration blocks publicly. Use environment management features of your MCP client or secret storage to protect credentials.

Troubleshooting

If you encounter connection issues, verify that the MCP server address is reachable and that your API key is valid. Check that the MCP client is configured to use the correct transport (http for remote, stdio for local). Review any client logs for transport errors or authentication failures.

Available tools

linkup-search

Real-time web search tool that queries the web to retrieve current information, facts, and news from trusted sources.

linkup-fetch

Fetch and extract content from any webpage, with optional JavaScript rendering for dynamic pages.