Hey Product Hunters 👋 I’m Prantik, and today we’re launching Verity by QAShift. We built Verity after noticing something frustrating about the current generation of AI testing products: Generating a test is becoming easy. Living with that test is still painful. The impressive demo usually ends when a test script appears on the screen. But the real work begins afterwards: • getting it into CI • deciding what should run on every commit • repairing selectors after the interface changes • distinguishing a flaky failure from an actual regression • maintaining the suite as the product grows • knowing whether a green build is genuinely safe to ship We didn’t want to build another recorder with an AI button. We wanted to build the entire testing loop. With Verity, you can describe a user journey—or give it a URL—and receive a runnable Playwright test. From there, Verity runs inside your own CI, proposes locator repairs as reviewable diffs, re-runs suspicious failures, quarantines flakes and builds a history of which files and modules are most likely to regress. There are two intentionally different execution modes: ⚡ Quick Check runs a relevant subset on every commit for fast feedback. 🛡 Full Regression runs the complete suite on merge, schedule or pre-deploy—and is the only mode that can block a release. We make that distinction because a fast, diff-aware test selection is useful, but it is not the same thing as a complete safety net. We would rather label the confidence honestly than turn it into another mysterious AI score. A few principles shaped Verity: Your tests should remain usable without us. Browsers and application secrets should stay in your CI. Self-healing must create a visible, reviewable change—not silently rewrite production tests. Flaky and broken are not the same thing. AI-generated tests are only valuable when they remain healthy six months later. You can try Verity without installing anything or entering a card. Generate your first test from the homepage, then connect your project when you’re ready. We would especially love feedback on three things: • Does the generated Playwright code meet the standard you would accept in your repository? • Do our Quick Check and Full Regression modes match how your team ships? • Which part of test maintenance wastes the most time for you today? We’re here all day, and we’ll answer everything openly—technical questions, architecture, limitations, pricing and what Verity still cannot do. Thank you for taking a look. We hope Verity gives your team something every engineering team deserves: The confidence to ship without holding its breath.

Verity - The AI test suite that writes, heals, triages and predicts

Replies