Use Agent Mode to interact with web applications

Beta
View as Markdown

Agent Mode can interact with web applications directly inside Postman using the in-app browser. This enables Agent Mode to navigate web pages, inspect page content, capture network activity, and automate browser workflows without leaving Postman.

You can use this capability to explore APIs behind web applications, generate tests and documentation from user flows, automate repetitive browser tasks, and debug frontend behavior.

The in-app browser is supported in Local View and Cloud View at the workspace level. For workflows that require editing files, such as generating Playwright tests, you must use a Git-connected workspace in Local View.

Use Agent Mode with web applications

Many modern APIs are best understood through the web applications built on top of them. Instead of manually inspecting requests in your browser’s developer tools, Agent Mode can interact with web pages directly inside Postman and help you understand the underlying API behavior.

Agent Mode can navigate pages, interact with UI elements, fill out forms, and run Playwright automation scripts. While interacting with a page, Agent Mode can also capture network traffic and browser console output, helping you understand how frontend behavior maps to backend API activity.

The in-app browser maintains its own cookies and browser session independently of your main Postman session.

Common usage scenarios

You can use Agent Mode with the in-app browser in many different ways depending on your workflow. The following examples highlight common scenarios where interacting with a web application can help you understand API behavior, automate browser tasks, generate tests and documentation, and debug frontend or backend issues.

Generate Playwright tests

Agent Mode can navigate an application and generate Playwright tests based on observed user workflows. Because Agent Mode also observes the underlying API activity, it can generate UI and API tests from the same workflow.

Example tasks include:

  • Sign in to an application
  • Navigate through onboarding flows
  • Create or edit resources
  • Exercise multi-step user interactions

Example prompt

Open https://example.com, sign in, navigate to the dashboard, and generate a Playwright test for creating a new project.

Generate API tests and Postman requests

As Agent Mode interacts with a web application, it can capture the network requests made by the page. This is useful for understanding how a frontend communicates with backend APIs.

You can use captured traffic to:

  • Create Postman requests
  • Generate collections
  • Create API tests
  • Inspect payloads and headers
  • Understand authentication flows

Example prompt

Open the application, create a new user through the UI, and generate a Postman request and API tests for the underlying API call.

Generate documentation from user flows

Agent Mode can navigate an application and summarize workflows, endpoints, and behaviors observed during browser interaction.

This can help generate:

  • API onboarding documentation
  • Authentication flow documentation
  • Endpoint summaries
  • Usage examples
  • Workflow walkthroughs

Example prompt

Navigate through the API onboarding flow and generate documentation describing how users create an API key and authenticate requests.

Debug frontend and API behavior

Agent Mode can inspect browser console output and capture network traffic during page interaction.

You can use this to:

  • Understand failing requests
  • Inspect payloads and headers
  • Investigate frontend errors
  • Compare browser behavior with API requests in Postman

Example prompt

Open the settings page, reproduce the failing action, and summarize any console errors and failed network requests.

Explore third-party dashboards and API portals

Some APIs are best explored through the products built around them. Agent Mode can navigate developer portals, dashboards, and admin interfaces to help you understand how the APIs work.

Example prompt

Open the developer portal, inspect the requests used to generate API keys, and summarize the authentication flow.