Stop shipping broken flows. AutoSmoke is AI-powered smoke testing that actually uses your app like a real user. An LLM agent drives a real browser to verify your critical userflows(login, checkout, onboarding) after every deploy. No brittle selectors. No test scripts to maintain. Just describe what to test in plain English, and AutoSmoke clicks, types, and verifies it works. Can be scheduled but it integrates with your CI/CD too. Set up in minutes, catch regressions before users do.

Framer — Launch websites with enterprise needs at startup speeds.

Launch websites with enterprise needs at startup speeds.

Hey everyone! I'm Laszlo, the maker of AutoSmoke. I built this because I kept living the same nightmare. Someone ships a "small change" and then Slack blows up because login is broken or checkout silently stopped working. We had E2E tests, but they were so brittle that maintaining them felt like a second job. Half the time we'd just skip them. The other thing I kept seeing is teams moving incredibly fast with AI coding tools, shipping multiple times a day, but nobody is actually checking if the core flows still work. I talked to a founder whose signup was broken for three days before a user emailed them about it. No tests, no monitoring, just vibes. Both problems pointed to the same question: what if the test could just *use* the app like a human does? That's AutoSmoke. You describe your flow in plain English and an AI agent opens a real browser and actually does it. Clicks buttons, fills in forms, verifies things work. When something breaks, you get a video of exactly what went wrong. It plugs into your CI/CD with a GitHub Action so it runs on every deploy. What flows would you test first?

Framer — Launch websites with enterprise needs at startup speeds.

Launch websites with enterprise needs at startup speeds.

AutoSmoke

Catch broken flows before your users do

Catch broken flows before your users do