Step 1
Write the tests
Add pm.test assertions to a request, folder, or collection. Check the status, validate the body against a JSON Schema, and confirm bad input fails the way it should.
Your API returns a 200. Now prove it returns the right thing.
Postman Functional Testing checks that each endpoint returns the right status, schema, and field values, using the collections you already have, so developers, QA teams, and coding agents catch broken responses before they ship.
Run it in the Collection Runner in the Postman app, or let your favorite coding agents like Claude Code, Codex, and Cursor run it for you with the Postman CLI.
Trusted by teams at
Install the Postman plugin and your agent gets testing workflows it can run on its own: writing assertions from your spec or code, running one request or a whole collection, and gating CI on failed tests. Your agent can also test against a mock when the live backend isn't ready.
Every command runs through the Postman CLI, the same tool your CI pipeline uses.
Marketplace
get the Postman plugin
claude plugin marketplace add anthropics/claude-plugins-officialclaude plugin install postman@claude-plugins-officialWrite and run assertions for any endpoint
Gate every PR on your Postman tests
Test failure paths against a mock
Run the same tests over every row of data
Keep tests as reviewable files in Git
Try these prompts
WHY POSTMAN
Postman Functional Testing runs the pm.test assertions saved in your collections against each response's status, schema, and field values. The same tests run in the Collection Runner, in your coding agent's terminal, and in CI, so there's no new framework to learn. A 200 only says the server answered. A failed assertion fails the run anyway.
Check status codes, field values, and headers, validate the body against a JSON Schema, and send incomplete data or wrong parameters to confirm the endpoint fails the way it should.
Assertions live at the request, folder, or collection level, so every run checks the same things: in your agent's terminal, in CI, and in the Postman app.
A failed assertion exits non-zero, even when the API returned a 200. Add --output for a YAML run report your agent can read back to see which tests failed.
Add the Postman CLI GitHub Action, or copy a CI config for your provider from the Postman app. A failing test fails the job. JUnit, JSON, and HTML reports are available for v2 JSON collections.
The collection your agent tests with also runs as a performance test with the same pm.test assertions, and as a scheduled monitor that alerts you when a test fails.
Working in the app instead of your editor? Ask Agent Mode to add new tests to a request or fix the ones that fail.
Step 1
Add pm.test assertions to a request, folder, or collection. Check the status, validate the body against a JSON Schema, and confirm bad input fails the way it should.
Step 2
Run it in the Collection Runner, or let your agent run postman collection run from its terminal.
Step 3
A failed assertion exits non-zero, even when the API returned a 200, so your agent knows which test broke, fixes the code, and reruns.
Step 4
Run the same command in your pipeline with the Postman CLI GitHub Action or your provider's config. A failing test fails the job.
Postman Collections have transformed API testing, enabling us to experiment quickly and efficiently when developing new features or implementing changes.”
Benny Yamagata
Senior Software Engineer, NetDocuments
API functional testing checks that an endpoint returns the correct response for a given input: the right status code, the right response schema, and the right field values. In Postman, those checks are pm.test assertions saved in a collection, so the same tests run from a coding agent, a CI pipeline, or the Postman app.
Yes. Install the Postman CLI with npm install -g postman-cli, then run postman init to install Postman's agent skills in your repository. Claude Code, Codex, Cursor, and other agents can then add pm.test assertions to a collection, run it with postman collection run, and read the result from the exit code. postman init needs no Postman account.
Run postman collection run in your pipeline. The command exits non-zero when a test fails, so a failing test fails the job with no extra setup. Use the Postman CLI GitHub Action, or generate the config for your CI provider from the collection in the Postman app. When the CLI is signed in, results upload to Postman, and CI runs include the commit and pipeline.
More than the status code. A good functional test checks the status code, the response schema, and the field values that matter, and sends incomplete data or wrong parameters to confirm the endpoint fails the way it should. In Postman, pm.response.to.have.jsonSchema validates a response against a JSON Schema, and pm.expect asserts on individual fields. See Write scripts to test API response data for the script reference.
Use Agent Mode when you work in the Postman app, and your coding agent when you work in your repository. Agent Mode is built into Postman and can add new tests to a request or fix failing ones. Coding agents such as Claude Code, Codex, and Cursor work through the Postman CLI and agent skills, so they write tests next to your code and run them in the terminal and in CI.
Newman is Postman's open-source command-line runner for v2.1 JSON collections, and it won't receive updates or fixes. The Postman CLI is Postman's official, signed command-line tool: it runs collections in every format, including the v3 YAML collections used by Native Git, and accepts most of Newman's options with the same syntax. It also works with mock servers, monitors, and API governance, which Newman doesn't. To switch, see Migrate scripts from Newman to the Postman CLI.
Postman and Bruno both keep collections as files in your repository. With Postman, those files are v3 YAML collections that your agent can write and you can review in a pull request, and the Postman CLI runs them locally and in CI. The same collection also runs as a performance test and as a scheduled monitor, and signed-in runs upload to Postman so your team can open the results without cloning the repo. Postman vs Bruno has the full comparison.
Functional testing is on every Postman plan, including Free. Automated testing with the Collection Runner is unlimited on all plans, and GraphQL and gRPC collection runs need a paid plan. Running packages from your team's Package Library through the CLI needs a Solo, Team, or Enterprise plan. The CLI runs on macOS, Linux (x64 and ARM64, not Alpine), and Windows x64, and doesn't support OAuth 2.0 authentication. See Postman pricing for what each plan includes.