Alumnium

If you’ve ever written browser tests and found yourself stuck between the precision of code and the clunkiness of manual steps, Alumnium might feel like a quiet relief. It’s a tool built for test engineers and developers who want to speed up test creation without giving up control. You still write the logic, but Alumnium helps translate your plain-language instructions into actual browser actions. It’s not trying to replace your framework or write tests for you – it’s more like a bridge between your brain and the browser.

I gave it a try while working on a small web app that needed some basic UI testing. Normally, I’d write everything out in JavaScript using Playwright or Selenium, and it would take a while to get the selectors right, handle edge cases, and make sure the test didn’t break every time the layout shifted. With Alumnium, I just wrote out what I wanted in plain English – something like “do click the login button” or “check that the welcome message appears” – and it handled the translation. It wasn’t magic, but it saved me a lot of time fiddling with selectors and waiting for elements to load.

The way it works is pretty straightforward. You write your test instructions using simple keywords like “do,” “check,” and “get.” These tell Alumnium what kind of action you’re describing – whether it’s something to perform, verify, or extract. Behind the scenes, it uses AI to interpret those instructions and generate the browser interactions. It leans on the accessibility tree and screenshots to figure out what’s on the page, which means it’s not just guessing – it’s actually looking.

What I liked most is that it doesn’t take over your workflow. You can use it with whatever test framework you prefer, and it doesn’t force you into a new ecosystem. I stuck with Python, and it worked just fine. You’re still in charge of the test logic, the assertions, and the structure. Alumnium just handles the part that usually eats up your time – clicking buttons, filling forms, waiting for things to load.

It also works with a bunch of AI providers, like OpenAI, Anthropic, and Google Gemini. You can plug in your own API key if you want, which gives you some flexibility depending on what you’re comfortable with. I used OpenAI for my tests, and the results were consistent enough that I didn’t feel the need to switch.

There’s a GitHub repo with examples, which helped me get started. I copied a few test cases and tweaked them for my own app. It didn’t take long to get the hang of it, and once I did, I found myself writing tests faster and with less friction. It’s not trying to be a full testing suite – it’s just a helper that sits quietly in the background and makes things smoother.

If you’re someone who writes browser tests regularly and wishes the process felt less tedious, Alumnium is worth exploring. It doesn’t try to be clever or flashy – it just helps you get from idea to execution with fewer steps in between. And sometimes, that’s exactly what you need.