About this site

Why this exists

Most hardware questions have a simple answer that is annoyingly hard to get. Is that key actually dead, or is the keyboard fine and the layout wrong? Is the mouse double clicking because the switch is worn, or because a setting is off? Is that dot on the screen a dead pixel or a speck of dust?

Those questions are answerable in about ten seconds with the right tool. Without one, people end up in decade-old forum threads, or replacing hardware that was never broken.

Every test here runs in your browser because that is the only place the answer exists. Your hardware is on your desk. No amount of writing about dead pixels tells you whether your screen has one.

How the tools work

They read your device directly through standard browser interfaces: keyboard and pointer events, the Gamepad API, getUserMedia for camera and microphone, the Fullscreen API for screen testing, and animation frame timing for refresh rate.

Everything is processed on your own device. There is no upload step, no server doing the analysis, and no account. This is not a privacy feature we added, it is simply how the tools work, and it is why the privacy policy can be short about the tools themselves.

The rules we hold ourselves to

We say when a browser cannot test something. The Fn key never reaches a web page. Mouse DPI is not exposed to any website. Several tools you will find elsewhere report these anyway, by guessing or by making something up. We would rather tell you the limit than invent a number.

We turn off the processing that would flatter your hardware. The microphone test disables noise suppression, echo cancellation and automatic gain control, which browsers enable by default. Leave them on and a failing microphone sounds fine, because the browser is amplifying and cleaning the signal. You hear what your microphone actually captures, which is sometimes worse than you expected and always more useful.

We distinguish a fault from a limit. A keyboard that cannot register six keys at once is not broken, it has a rollover limit. A controller that shows 0.03 stick deflection at rest is within normal tolerance, not drifting. Reporting normal behaviour as a fault sells hardware nobody needed to buy.

Model pages cite their sources. Where we list a specific switch or specification, the source is linked at the bottom of the page. This matters more than it sounds: the Logitech G502 Hero uses an Omron D2FC-F-K, while most search results confidently say D2FC-F-7N. If you order parts based on a wrong page, you get the wrong parts.

We tell you when a test cannot be trusted. The refresh rate test discards its own result if the tab lost focus, because browsers pause animation frames in background tabs and the number would be meaningless. It is better to ask you to run it again than to report something wrong.

What we are not

We are not a repair service, and we are not affiliated with any hardware manufacturer. Brand and product names are used to identify hardware, nothing more.

We are also not a substitute for a manufacturer’s own diagnostics when something is at stake. If you are inside a return window or deciding on a warranty claim, use these tools to understand the problem, then confirm it through the channel that actually matters. The terms go into more detail about the limits of a browser based test.

How this site is funded

Advertising. That is the whole model. There is no paid tier, nothing is upsold, and no test result is used to recommend a product.

It does mean advertising partners collect data even though the tools do not, and the privacy policy explains that split rather than glossing over it.

Corrections

If something here is wrong, we want to know, particularly on the hardware model pages where a wrong switch type sends someone to buy the wrong part. Get in touch via the contact page.