Hi HN, sharing a SaaS product that I built to automate web UI testing based on natural language test descriptions. The approach we use is a multi-agent system that uses an augmented vision approach for interacting with the UI, rather than relying on the DOM. The goal is to mimic how your users would naturally use the UI, rather than following a predefined path.

I built this because I got tired of doing manual QA for my last SaaS product. I couldn't trust automated UI tests (Selenium/Playwright) because they were flaky, broke whenever the DOM changed, and missed obvious usability bugs.

We are completely bootstrapped, but we want to gather as much feedback as possible in these early days, so we have made the product available for free (for now).

Happy to answer any questions about it here, and would love your feedback.

Thanks!