Donobu Logo
API testing

Your API, tested on every release.

Donobu's HTTP testing plugin adds deterministic calls, assertions, and polling to Playwright's request fixture, with AI only where you want it. Every run shows each request and response, so a failure points at the exact call that broke.

Book a demoAsk about your API

Orders API · 6 requests

Running…

orders.api.spec.tsstaging
    Reviewed findings land here when the run ends
    The problem

    UI tests catch the symptom. API tests catch the cause.

    When an order total is wrong in the browser, the bug usually lives one layer down. Teams that test only through the UI find it late. Teams that keep API tests in a separate tool end up with two suites, two reports, and two places to look when something fails.

    How it works

    One suite for the browser and the API.

    01

    Install the plugin.

    Add @donobu/donobu-http-testing to a Playwright project and import httpTest. Your tests run on the same Playwright runner and CI setup you already use.

    02

    Write calls and checks that stay deterministic.

    request.call, assert, extract, and poll cover the usual API test: send a request, check the response, carry a value into the next call, and wait until a job finishes. No model is called unless you ask for one.

    03

    Bring AI in where it helps.

    request.ai puts a model in the loop for the steps where you want one. A coding agent can write the tests too: donobu skills install gives it grounded guidance for the HTTP plugin.

    04

    Read the run as requests, not screenshots.

    On donobu.com, an API run shows each request and response with the method, URL, status, and a highlighted JSON body. Poll steps keep their summary, and secrets in URLs and headers are redacted.

    One system

    Same project, same dashboard as your web tests.

    API runs use the same record shape as browser tests, so they live in the same project, the same test runs, and the same reports on donobu.com. API tests carry their own icon in the tests list, and a failed run points at the request that failed.

    Ownership

    Real Playwright. In your repo.

    API tests are Playwright tests in TypeScript, checked into your repo next to your browser tests. Donobu runs local-first, on your machine or in your CI, and nothing about your tests is locked into a Donobu-only format.

    More ways teams use Donobu

    Functional testing

    Turn critical user journeys into a regression suite that never goes stale.

    See functional testing

    Exploratory testing

    Point an agent at your app and see what's actually broken, no script required.

    See exploratory testing

    Mobile app testing

    Test native Android and iOS builds the same way you test the web.

    See mobile app testing

    Localization QA

    Continuous QA across every market and locale you ship.

    See localization QA

    Content QA

    Every page and document you publish, checked against your guidelines for tone, accuracy, and freshness.

    See content QA
    FAQ

    Frequently Asked Questions

    Do we need Donobu's browser testing to use API testing?

    No. The HTTP plugin works on its own in any Playwright project. If you also test the browser with Donobu, both kinds of test share one project and one dashboard.

    Does every API test call an AI model?

    No. request.call, assert, extract, and poll are deterministic and run without a model. request.ai is there for the steps where you want one.

    What happens to tokens and other secrets in the report?

    The run view on donobu.com redacts URL query values, user info, and header values.

    Can Donobu write and run the API suite for us?

    Yes. As a managed service, our engineers author the suite, AI agents run it on every release, and test engineers review every result before it reaches you.
    Get started

    Ready to see Donobu test your API?

    Book a demo