What this site does
KeyboardTesterOnline is a set of focused browser tests for one question at a time: is this key registering, is it registering twice, how many keys register at once, and how fast. The main tester on the homepage covers the general case — press every key, watch each one light up. Ten dedicated tools cover the problems a full-board check can’t isolate on its own.
Alongside the tools there are twelve repair guides. Each one starts from a test result rather than a guess, because the single most useful thing you can know about a keyboard problem is whether it’s hardware or software — and a tester answers that in about ten seconds.
How the tests work
Every tool listens for the browser’s raw keydown and keyup events and reports exactly what arrived. Nothing is inferred and nothing is smoothed. When a key lights up, the browser genuinely received a signal for it. When a key stays dark while you’re holding it down, the browser genuinely received nothing.
That matters for diagnosis. A text editor shows you the character a key produced, which tells you nothing about Shift, Ctrl, Alt, the function row, or any key whose signal was suppressed before it became text. A tester shows the event itself, which is the layer where keyboard faults actually live.
What the tests can’t tell you
Being clear about the limits is more useful than overselling the tools:
- Switch feel — actuation force, travel distance, wobble and how a switch sounds are mechanical properties. A key can feel mushy or scratchy and still register perfectly.
- Keys the OS intercepts — PrintScreen, Fn, Touch ID and Windows key shortcuts never reach the browser at all.
- The cause of a fault — a test tells you a key isn’t sending a signal. Whether that’s debris, a worn contact, a torn ribbon cable or a driver failure comes from the pattern, which is what the guides are for.
- Absolute hardware latency — the latency and polling tests measure what the browser observes, which includes your OS input stack and display pipeline, not just the keyboard.
Privacy — the short version
No keypress is recorded, stored, or transmitted anywhere. The tests run entirely in your browser, there’s no account to create, and there’s nothing to install. The privacy policy sets out exactly what is and isn’t collected.
Corrections and suggestions
If a tool misreports something on your hardware, a guide gives advice that didn’t work, or a layout doesn’t match your keyboard, that’s worth knowing about — keyboard behaviour varies more across models than any single test can anticipate. The contact page has the details.
Anyone with a keyboard that isn’t behaving — someone checking a used board before paying for it, a gamer whose combos drop inputs, a typist chasing double letters, or a builder verifying switches after a hot-swap. No prior hardware knowledge is assumed anywhere on the site.
Once a page has loaded, yes. Every test runs in JavaScript in your browser with no server round-trip, so you can disconnect from the internet and the tools keep working exactly the same way. That is also the simplest way to confirm for yourself that nothing is being transmitted.
Browsers deliberately never see some keys. PrintScreen, the Fn key on most laptops, Touch ID on a Mac, and Windows key combinations such as Win+L are handled by the operating system or the keyboard firmware before any web page is notified. That is a security boundary, not a limitation we can code around — no browser-based tester can read those keys.
The chatter, latency and polling rate tests read the browser’s high-resolution event timestamps, which are accurate enough to separate a 5ms contact bounce from a 300ms double tap — the distinction that actually changes your diagnosis. They are not lab instruments. Browser timer resolution, display refresh rate and background CPU load all add small amounts of noise, so treat a single reading as indicative and a repeated pattern as reliable.