Mechasm.ai Open Alpha – AI-powered Web Testing with Zero Setup

Hey everyone,

I’m excited (and a little nervous) to open up the alpha for a tool I’ve been building:
Mechasm.ai

Why I built this:

As a developer, I got tired of dealing with flaky selectors, slow CI runs, and rewriting tests every time the UI changed. I wanted a tool that just worked, something that let me focus on testing features and edge cases, not maintaining scripts.

So I built Mechasm:
A platform for fast, self-healing, browser-based tests that require no setup, no test code, and no infrastructure. Just describe what you want to test in plain language and run it against a public URL.

Who it's for:

Product-minded developers who want high-level coverage fast
QA engineers who spend too much time debugging brittle E2E tests
Teams without dedicated QA who want to test critical flows reliably
Even non-technical folks who want to validate their website behavior

How it works:

1. Go to mechasm.ai

2. Create a team

3. Create a project

4. Create a test

5. Paste in a public URL

6. Write what you want to test in plain language (e.g. “Sign in with x and y, go to dashboard, check for balance text”)

7. Mechasm spins up a headless browser, runs your steps, adapts to UI changes, and gives you:

   • Video replay ( on failure by default, on passed can be enabled in project settings )

   • Execution logs and console errors

8. Optionally:

   • Trigger tests from your CI ( check project settings )

   • Get notified when something breaks ( via email )

Open Alpha notes:

This is still early so expect bugs, but the core loop works. I’d love your feedback on:
What works or what’s missing
Where it breaks for your stack
What would make this useful for your team

You can try it now at https://mechasm.ai

Happy to answer any questions or chat if you're curious.

Thanks for reading and building better software

Silviu