Keyboard Latency Test

This test times the part of a keypress that a browser is actually able to observe, and it is careful not to claim more than that. When you press a key, the browser creates an event and stamps it with the moment it was generated. This page then compares that stamp against the clock at the instant its own code runs, and the difference is the event handling delay — how long the keypress spent queued between the browser and the page.

The second figure is key to next frame. As soon as the keypress arrives, the page asks the browser to call it back before the next repaint, and times how long that takes. This is bounded by your refresh rate: on a 60 Hz screen the answer can be anything from almost nothing to about sixteen milliseconds, depending purely on where in the refresh cycle your press happened to land. On a 144 Hz screen the ceiling drops to roughly seven. Together the two numbers describe how quickly anything you type can possibly show up.

Pay as much attention to the spread as to the average. A machine that answers in two milliseconds every single time feels crisp, while one that averages the same figure but swings between one and twenty feels unreliable, because your hands notice inconsistency more readily than a slightly higher constant delay. Background compilation, busy tabs, browser extensions, and aggressive CPU power saving all widen that spread.

What this cannot tell you is your keyboard's hardware latency — the time from the physical switch closing to light changing on the display. That measurement requires instrumenting the hardware, typically with a high-speed camera or a microcontroller wired to the switch, and it is genuinely out of reach for any web page. If you are comparing keyboards rather than software, published measurements from reviewers with that equipment are the right source.

Everything is measured and kept in your browser. No keystroke data is transmitted, and the tool records only timings and the name of the key.