Build an API Status Checker Agent with Postman (Full Tutorial)

Quick Overview

The tutorial successfully demonstrates how to build an AI Agent using Postman Flows to automatically check the status of all APIs within a Postman collection against a production environment, reporting successes and failures, which is then deployed as an MCP server agent callable via a URL.

Key Points: The tutorial builds a Postman Flow to act as an AI Agent that checks the status of all APIs in a Postman collection against a production environment. The flow successfully retrieves the collection details using the 'Get Postman Collection' API, extracts the base URL and API key, and flattens the nested list of APIs using in an 'Evaluate' block. The agent uses the 'API Status Checker' tool (which relies on a specific production URL) to check the status of 25 endpoints, reporting 21 successful and 4 failed (all 404 errors on authentication endpoints). The final step involves deploying the Postman Flow as an MCP server agent, accessible via a specific URL, which can then be called by the AI Agent tool within the flow. The resulting agent summarizes the health check, noting that most services are healthy but identifying four specific endpoints (auth login, auth/logout, control/users, control/powerservers) returned 404 errors, likely due to incorrect HTTP methods (GET instead of POST) or missing configurations.

Context: This video is part two of a tutorial series focused on leveraging Postman Flows to create custom, automated AI Agents capable of interacting with Postman APIs. The presenter, Sterling Chin (Sr. Developer Advocate at Postman), constructs a flow designed to iterate through all requests in a specific Postman collection, execute them against a defined production environment, and summarize the success or failure status of each endpoint, effectively creating an automated API status checker agent.

Detailed Analysis

The tutorial builds a complex Postman Flow designed to function as an automated API status checker agent. The process begins by defining inputs in the 'Start' block: , , and (the production endpoint URL). An 'Evaluate' block then uses JavaScript ( ) to process the list of endpoints fetched from the Postman API, flattening the nested structure into a simple list of URLs. This flattened list is fed into a 'For' loop block, which iterates over each API endpoint URL. Inside the loop, an HTTP Request block calls the 'API Status Checker' tool using the production base URL and the Postman API key (retrieved from configurations or environment variables) to check the endpoint status. The results from the HTTP Request are collected, and a final 'Evaluate' block summarizes the findings, noting 21 successful checks and 4 failures (all 404 errors related to authentication/control endpoints). The presenter then configures the flow's output and deploys the flow as an agent on an MCP server, accessible via a specific URL, which confirms the tool is working as expected by successfully checking the API statuses.