globalping

Globalping MCP Server exposes Globalping’s network measurement capabilities (ping, traceroute, DNS, MTR, HTTP) through the Model Context Protocol so AI assistants can request measurements using natural language. You can connect any MCP-compatible AI tool or client (e.g., Claude Desktop, Gemini/Claude API, Cursor, Windsurf, or other MCP implementations) to this server and issue high-level requests like: ping a domain from multiple locations, run traceroutes from specific regions, or compare measurements across targets. The server supports both OAuth and API token authentication to enable higher rate limits when you have a Globalping account. When using an API token, include it in the Authorization header to switch from OAuth to token-based authentication seamlessly. To connect, configure the MCP client with the server’s streamable HTTP endpoint (https://mcp.globalping.dev/mcp) or SSE endpoint (https://mcp.globalping.dev/sse) and provide credentials if required by your setup.

Prerequisites:

• Node.js (recommended LTS version) and npm installed on your machine.
• Internet access to fetch MCP client bridge tools.

Installation steps:

1. Ensure Node.js and npm are installed. Verify: node -v npm -v

2. Install or use npx to run the MCP bridge that connects to the Globalping MCP server. The instructions below use npx to avoid a global install:

   • For Streamable HTTP transport: npx mcp-remote https://mcp.globalping.dev/mcp

   • For SSE transport (where supported): npx mcp-remote https://mcp.globalping.dev/sse

3. If you need API token authentication, prepare an authorization header and adapt the command accordingly in your MCP client configuration:

   • For clients that support header injection, set: Authorization: Bearer YOUR_GLOBALPING_API_TOKEN

4. Optional: If you’re integrating with a specific tool (Gemini, Claude, Cursor, etc.), follow that tool’s MCP integration guide and point it to the provided Globalping endpoints as shown in the README.

Tips and tips:

• If you plan to use higher rate limits, obtain an API token from dash.globalping.io and configure the Authorization header in your MCP client setup.
• When using GPUs/LLMs that require latency awareness, consider requesting measurements from multiple locations to get a broad overview.
• The server supports a variety of tools (ping, traceroute, dns, mtr, http, locations, limits, getMeasurement, compareLocations, help). Use 'help' to discover available operations from your client.
• If you experience connectivity issues, verify that you can access https://mcp.globalping.dev/mcp and that any firewalls or proxies permit WebSocket/SSE traffic as required by your client.
• For advanced automation, store credentials in environment variables or secure vaults and reference them in your MCP client configuration rather than hardcoding tokens.