Globalping

Globalping MCP Server lets AI models interact with Globalping’s global network measurement capabilities using natural language. You can run measurements like ping, traceroute, DNS, MTR, and HTTP from thousands of probe locations, while authenticating securely to get higher rate limits and priority access.

How to use

You connect an MCP-compatible AI assistant to the Globalping MCP Server and then ask it to run measurements in natural language. For example, you can request tests from multiple locations, compare results across targets, or look up measurement details. The server translates your requests into concrete network tests and returns the results in a structured format that your AI assistant can interpret and present.

How to install

Prerequisites: you should have Node.js and npm installed on your machine, or use an environment that already supports MCP clients. You will also need access to one of the MCP-compatible assistants you plan to connect.

Step 1: Ensure you have an MCP client ready (for example Claude Desktop, Gemini CLI, Cursor, Windsurf, or another MCP-compatible tool). Step 2: Configure the Globalping MCP Server endpoints in your client. Step 3: Authenticate using OAuth or an API token to increase rate limits.

Authentication

The Globalping MCP Server supports two authentication methods. OAuth Authentication is automatically handled for secure access. API Token Authentication can be used for manual token configuration via the Authorization header and provides higher rate limits when you have an account.

Connecting AI Assistants

You can connect any MCP-compatible assistant to the Globalping MCP Server. This includes Claude Desktop, Anthropic Assistants, Cursor, Windsurf, or custom MCP client implementations. Set the server URL and authentication method in your tool’s MCP configuration so it can send natural-language requests that map to Globalping measurements.

Usage Examples

Ask for measurements described in natural language, for example: Ping a domain from multiple locations, run a traceroute from a specific region, or compare DNS resolution across different resolvers. The MCP server handles the interpretation and returns results you can present in your assistant.

Location and targeting guidance

Locations can be targeted using a range of descriptors such as continents, countries, city names, networks, ASNs, and cloud regions. You can combine targets to focus measurements from precise sets of locations.

Configuration and endpoints

The remote MCP server endpoints you will connect to are available at two transports:

Streamable HTTP transport: https://mcp.globalping.dev/mcp

SSE transport: https://mcp.globalping.dev/sse

Example MCP configuration for the Globalping server

{
  "mcpServers": {
    "globalping": {
      "type": "http",
      "name": "globalping",
      "url": "https://mcp.globalping.dev/mcp",
      "args": []
    }
  }
}

Notes on authentication tokens

If you provide an API token, the server uses it automatically for authentication and applies higher rate limits. You generate tokens in your Globalping account under the Tokens section.

Available tools

ping

Run a ping test from target(s) to measure latency and packet loss.

traceroute

Trace the path to a target to visualize routing and hop-by-hop latency.

dns

Perform a DNS lookup to resolve a domain and inspect DNS records from various resolvers.

mtr

Execute an MTR (My Traceroute) test to observe route changes and performance.

http

Send an HTTP request to a URL and measure response details.

locations

List all available Globalping probe locations.

limits

Show your current rate limits for the Globalping API.

getMeasurement

Retrieve details for a previously run measurement by ID.

compareLocations

Guide on performing comparisons of measurements across different locations.

help

Show a help message with available tools and usage notes.