MCP servers in Postman Flows

View as Markdown

In Postman Flows, you can create and deploy a flow that functions as a Model Context Protocol (MCP) server with MCP tools, prompts, and resources. When deployed to a public URL, the flow can process incoming requests and send the results as responses.

This document describes the concepts behind MCP servers in Flows. For the steps to create one, see Create an MCP server with Postman Flows.

A deployed flow has a public URL in the Postman cloud. Users can send requests to the flow’s URL. The URL can also be triggered by external systems like webhooks, third-party apps, and other APIs. The flow processes incoming data using the blocks on its canvas, then sends the result as a response.

To function as an MCP server, a deployed flow needs an MCP block that defines the server’s MCP tools, prompts, and resources. Data items required by MCP tools are called arguments. The body of the MCP block must follow the MCP tool definition structure.

Test your MCP server in Postman

You can use your MCP server from any client that supports the Model Context Protocol, such as Claude or VS Code GitHub Copilot. But you can also quickly test your MCP server from Postman, using an MCP request.

When you create an MCP request in Postman and connect to the MCP server, the request shows the available tools. You can click a tool to see its required arguments. You can then enter the required arguments and click Run to get a response from the MCP server.