Qualifix generates Playwright end-to-end tests from a URL or plain English, then self-heals them when your UI changes by re-locating elements the way a person would (label, ARIA, visible text) instead of a brittle CSS selector. It also explains test failures with AI instead of a raw stack trace, runs automatically on every GitHub PR with risk-scored diff analysis, and remembers past fixes so recurring failures aren't re-solved from scratch. Built solo in three months, live today, on zero budget.

Hey Product Hunt 👋 I built this alone because I kept watching E2E suites break for the wrong reasons — not real bugs, just selectors that couldn't survive a redesign. This tries to fix that: AI generates the tests, self-healing keeps them alive through UI changes, and it plugs into GitHub PRs to catch regressions before they ship. It's live and working, not a prototype — frontend on Cloudflare, backend on Railway, real Postgres, real auth, real security basics (rate limiting, hardened headers, bcrypt). No customers yet, and I'm not going to pretend there are. I'm here to get it in front of teams who have this exact pain and find out what's broken or missing. No self-serve signup right now — comment, DM me, or email qualifixsoftware@gmail.com and I'll walk you through a live demo.

honestly the self-healing sounds like the killer part here. one thing that would actually save me though is letting it record a short video clip of the failing step so i can see what the page looked like when it broke, since sometimes the ai explanation still leaves me guessing about the visual state

Qualifix

AI that writes and self-heals your E2E tests

AI that writes and self-heals your E2E tests