Introducing the Postman API MCP server collection

Ask the Experts and Postman Tips
,
1

We’ve just launched an official Postman API Model Context Protocol (MCP) server. It allows AI agents like Claude and VS Code to help you manage your Postman resources—things like workspaces, collections, specs, mock servers, and monitors. The server translates your instructions into executable API workflows behind the scenes.

MCP is an open protocol that standardizes how applications share context with LLMs, making it much easier to build agents and complex workflows on top of them. An MCP server exposes APIs and other contexts as tools that AI agents can access. Communication between the agent and your API happens through defined transport methods—standard input/output (stdio) or server-sent events (SSE)/HTTP:

  • stdio — When using stdio, the MCP server runs locally on the agent machine. You can set this up by either downloading and running the source code yourself or spinning up a local Docker image.
  • SSE/HTTP — Securely links the agent to the server over HTTP(s), without requiring any extra setup on the agent. A simple way to use the Postman API MCP server with your preferred IDE or AI agent.

To use the Postman API MCP server with your IDE or AI agent, you’ll just need to get it set up—we’ll guide you through it using a few popular tools.

VS Code

The MCP configuration for the current repository resides in the .vscode/mcp.json file. Create or edit the file and add the following:

{
    "servers": {
      "postman-api-http-server": {
        "type": "sse",
        "url": "https://mcp.postman.com/mcp",
        "headers": {
            "Authorization": "Bearer ${input:postman-api-key}"
          }
      }
    },
    "inputs": [
      {
        "id": "postman-api-key",
        "type": "promptString",
        "description": "Enter your Postman API key"
      }
    ]
  }

In VS Code, you can start the server by clicking “Start” next to the server name. You’ll be prompted to enter your Postman API key, which is needed to authenticate requests and act on your behalf.

After the server is up and running and you’ve entered your API key, you can ask Copilot or any other integrated LLM in VS Code to help manage your Postman resources.

Claude

In Claude, you’ll need to use a proxy to send the Authorization header to the MCP server. Edit your claude_desktop_config.json file:

{
  "mcpServers": {
    "postman-api": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.postman.com/mcp",
        "--header",
        "Authorization: Bearer PMAK-YOUR-POSTMAN-API-KEY-HERE"
      ]
    }
  }
}

Afterwards, restart the Claude desktop app, and you’ll be able to ask Claude about your Postman resources.

Cursor

At this time, Cursor doesn’t support sending headers to SSE servers, and the Postman API MCP server requires the API Key in an Authorization header. In this forum post there are a few alternatives, like using a gateway. For example:

{
  "mcpServers": {
    "supergatewayExample": {
      "command": "npx",
      "args": [
        "-y",
        "supergateway",
        "--sse",
        "https://mcp-server-ab71a6b2-cd55-49d0-adba-562bc85956e3.supermachine.app",
        "--header",
        "Authorization: Bearer YOUR-POSTMAN-API-KEY"
      ]
    }
  }
}

Summary

The Postman API MCP server simplifies managing Postman resources with AI agents like Cursor, Claude, and VS Code. By standardizing interactions with LLMs, it streamlines workflows and easily integrates with your tools. We’d love to hear how you plan to use MCP with your tools—let us know in the comments!

3 Likes