On this page
There is a specific kind of bad day that I kept having.
The status page would show green. Response time charts would be flat. No alerts. No pages. And then someone would message me (a user, a friend who'd signed up, occasionally my own test account) and they'd say: "hey, I couldn't check out." Or: "the login screen just keeps loading." Or the quieter version: nothing. They just stopped coming back.
The server was up. The endpoint returned 200. The product was broken.
That gap, between "the infrastructure is responding" and "the thing I built actually works for people," is the one I couldn't close with anything that already existed. Uptime monitors check whether a machine answers. They don't check whether your checkout completes, whether your onboarding finishes, whether the one test that drives your whole business is doing what it's supposed to do.
The lie
Every monitoring tool I tried told me the same lie. "All systems operational" is a statement about servers. It says nothing about tests.
I'd set up Playwright tests and they'd break the next time someone moved a button. I'd hire someone to write manual test cases and they'd become stale the week after the next deploy. I'd turn on synthetic monitoring and it would check whether a URL loaded, not whether the thing on the URL worked.
The problem wasn't a lack of tools. The problem was that every tool was solving a slightly different problem (usually an infrastructure problem) while the product-level failures kept happening quietly.
What we built
Smoketest runs your critical user tests against a real browser after every deploy. Not a simulated HTTP request. Not a ping. An actual browser, completing the actual test, the way a real user would.
You describe the test the way you'd explain it to a teammate. "User signs up, verifies email, connects their first integration." We run it. We tell you whether it worked, and we show you exactly what happened when it didn't.
No selectors to maintain. No test code to write. No Playwright to debug at midnight because someone renamed a CSS class.
Who this is for
If you have a product that real people use and you don't know right now whether the most important test in it is working, this is for you.
That's the whole thing. I'm not going to tell you about our architecture or our funding. I'm going to say: if the thought "what if checkout is broken and I don't know it" has crossed your mind in the last week, we built this for that feeling.
The monitor has been lying to you. We're trying to fix that.