Keyboard Latency Test
Press any key to see how long it takes to show up in your browser.
This browser rounds key times to about ms
Keys with possible chatter
Ready.
Every test runs inside your browser. Nothing you type, say, show or press is uploaded.
How to use this test
This keyboard latency test times every key press inside your browser, from the moment the browser receives the key to the start of its next animation frame. It also shows how long you hold each key, flags keys that go down again a few milliseconds after you release them, a common sign of chatter, and measures your computer's key repeat delay and rate.
Click the box above so it has focus, then press keys the way you normally would. Each press adds a point to the chart. To check for chatter, tap one key at a steady pace for half a minute and watch the double-type counter. To measure key repeat, hold a key down for two seconds. Reset clears the readings; the threshold stays where you set it.
What the results mean
Event to frame is the time from when the browser stamped the key event to the start of the next animation frame. With nothing else running it usually stays within a frame or two; one frame is about 17 ms on a 60 Hz screen and 7 ms at 144 Hz. Processing delay is how long the event waited before the page's code ran. Hold time is how long the key stayed down. A double-type is counted when the same key goes down again within the threshold after you released it.
What a browser can and can't measure about input lag
A keyboard's response time is the sum of several stages. The switch closes, the keyboard's controller filters out bounce and waits for the next USB poll, the operating system handles the input, and only then does the browser get an event. This test starts its clock at that event and stops it at the browser's next animation frame, so it covers the last stage only.
The last stage still tells you something. If the event-to-frame time jumps around or sits well above one frame, the delay comes from the computer, and a faster keyboard won't fix it. Close heavy tabs, and on a laptop test with the charger connected, before you blame the hardware.
Hold time, key repeat and the double-type threshold
Hold time is how long each key stays down. In normal typing it is often around a tenth of a second, and gamers holding movement keys see much longer times.
Key repeat belongs to your operating system. Hold a key and it waits for the repeat delay, then sends repeats at the repeat rate. On a Mac you set both in System Settings > Keyboard; on Windows, in the Keyboard control panel. The test shows repeats separately and never counts them as chatter.
The threshold decides how quickly a second press has to follow the release to count as a double-type. The default is 30 ms, because a deliberate double tap rarely comes back that fast. Lower it if you tap very quickly; raise it toward 80 ms if a key still looks suspicious.
Common questions
Can a website measure my keyboard's real input lag?
Not all of it. The switch, the keyboard's controller, the USB polling interval and the operating system all add delay before the browser hears about the key, and a web page can't see that part. Polling alone can add up to 8 ms on a 125 Hz keyboard and up to 1 ms at 1000 Hz. Measuring the whole chain takes a high-speed camera or a dedicated latency tester.
A key types twice when I press it once.
That is usually chatter: a worn or dirty switch can bounce and send a second press a few milliseconds after the first. The test can't see the switch itself, but if the double-type counter climbs while you tap one key at a normal, steady pace, chatter is the likely cause. Compressed air sometimes clears it. On a hot-swap mechanical keyboard, replacing the switch usually fixes it, and some keyboards let you raise the debounce time in their software.
My numbers jump to 50 ms or more.
Spikes like that usually come from the computer, not the keyboard: a busy tab, a background task, or a laptop saving power. Close other tabs, plug the laptop in and test again. If a note under the readouts says the browser rounds key times, part of the difference between two readings is rounding, so compare medians rather than single presses.