If you use ChatGPT or Claude to write test cases, the AI does the 30 second part. You do the 10-day part: copy into Confluence or Excel, reformat, tag, share. And the output is always a little off, because a chat window only sees what you paste. It doesn't know your payment flow depends on a Stripe webhook, or that onboarding was redesigned last sprint.
The difference isn't generation. It's connection.
Hey Product Hunt 👋
I'm Deniz, PM at Text2Test.
AI writes our code in minutes now. That's the easy part. The hard part: you can't fully trust it, and you can't easily validate it. Fast code you can't verify isn't actually faster, it just moves the risk downstream.
Testing should close that gap. But it stayed manual, slow, and brittle, breaking on every UI change. So we built Text2Test.
You describe a test in plain text, and it compiles into a deterministic test that runs the same way every time on real browsers. Consistency you can count on, not flaky results you re-run. When your app changes, tests self-heal, so you stay in control instead of maintaining them.
We are ambitious about hitting the needs. Try it, tell us what's missing, we welcome all of it and read every comment. Thanks for checking us out 🙏
Mantra Timer
@deniz_colakoglu1 - Interesting product, would this work for native ios and macOs swift test using xcode?
@codeandsea Thanks!
Native iOS: yes. We drive both real devices and simulators, so a native Swift or SwiftUI app works. It's not in the self-serve product yet (mobile is our next launch), and it comes with a local runner app that connects your own Mac plus devices and simulators.
Worth noting: we don't generate XCTest or XCUITest code inside Xcode. You write the test in plain English and our AI runs it against your built app, so it sits alongside your Xcode suite and can still trigger from CI.
Since mobile isn't public yet, happy to give you a live demo on a real app. Grab a slot at text2test.ai/demo or just reply here.
A nice bonus: you write the test case once and run it on both iOS and Android, since the tests describe behavior in plain English rather than platform-specific code.