👋 Hey Product Hunt! We built Testronaut to make end-to-end testing less painful and more adaptive. Instead of writing brittle scripts that break every time the UI shifts a pixel, you define your test as a "mission" in plain English — and an AI agent runs it through a real browser using Playwright. Why we built this: as devs, we kept running into the same problem — E2E tests are slow to write, hard to maintain, and fail for silly reasons. We wanted something that understands the workflow while still giving a deterministic pass/fail report. ✨ What it does today: - Run real-world UI flows using AI + Playwright - Define tests as simple missions (strings or functions) - Step-by-step JSON + HTML reports with screenshots - CLI-based (`npx testronaut`) — runs locally, no hosted dependency 🚀 Current status: early but working. Great for login flows, onboarding, or quick sanity checks. Open source on npm and GitHub. We’d love your feedback: - Where would this fit into your workflow? CI/CD? Replacing flaky tests? Exploratory testing? - What worries you about AI-driven test runners? - Anything we should prioritize on the roadmap? Thanks for checking it out! 🙏

Testronaut

Testronaut – AI-powered mission-based browser testing

Testronaut – AI-powered mission-based browser testing