Your AI Agents Can Now Use Postman within Kiro IDE

Quick Overview

Postman Kiro Power allows AI agents within the Kiro IDE to interact directly with Postman resources, enabling complex workflows like discovering an API, creating necessary Postman workspaces and collections from source code, and validating the generated collection against the Postman Collection Format schema, demonstrated by setting up and testing a Planet Information API.

Key Points: Postman Kiro Power integrates Postman functionality directly into the Kiro IDE, enabling AI agents to manage Postman resources. The demonstration shows the agent automatically reading local source code (server.js) to infer API structure and generate a Postman collection. The agent checks for existing Postman resources (like workspaces) and prompts the user for selection or creation (02:47). The agent successfully created a collection containing all 17 API endpoints, including test scripts and environment variables, based on the application code (3:13). A key feature is the Postman MCP Server Steering Guide documented in the code, outlining patterns for creating workspaces/collections and running collection tests (03:22). The generated collection tests executed successfully, covering planets, weather, moons, orbital mechanics, and search functionality with a 100% success rate (4:30).

Context: Quinton Wall, Head of Dev. Relations at Postman, demonstrates the integration of Postman's capabilities into the Kiro IDE via Kiro Power. This integration allows AI agents operating within the IDE environment to programmatically interact with Postman's features, such as creating collections and running tests, by leveraging a Postman MCP (Meta-Control-Plane) Server connection.

Detailed Analysis

The video showcases Kiro Power enabling an AI agent to perform complex API setup tasks within the Kiro IDE by connecting to a Postman MCP Server. The process begins with the agent reading the local application code (server.js) to understand the API structure, including endpoints like GET /planets and GET /planets/:name/weather (00:11). The agent executes a series of scripted steps defined in a Postman MCP Server Steering Guide (visible in the code editor), which outlines patterns for creating workspaces and collections from API specifications (03:22). When prompted to create a collection from the app, the agent first authenticates with the Postman MCP server (02:04), retrieves existing workspaces (02:47), and opts to use an existing workspace named 'Quinton's Playzone' (02:52). Since no collections exist, the agent proceeds to create a comprehensive collection covering all 17 API endpoints, complete with test scripts and environment variables (03:10). Finally, the agent executes the generated collection tests against the running local API (http://localhost:3000) (04:15), resulting in a 100% success rate, validating status codes, response structures, data integrity, and business logic (04:30). The entire process highlights how Kiro Power automates infrastructure setup for API testing directly within the developer's IDE.