Average, minimum and maximum
Compare the centre and range of the browser-processing sample instead of trusting one unusually fast key event.
Run a quick browser-based keyboard delay test. See average, minimum and maximum event-processing delay, jitter, consistency, observed event rate and recent key timing.
Start the timer and use several keys at a steady pace. The page measures when keyboard events reach and are handled by this browser, then summarizes the sample.
Start the timer, then press different keys at your normal pace.
Only keyboard event timing is measured. Nothing is uploaded.
No key event measured yet
A website can estimate browser-level event delay and the gaps between received key events. It cannot see the exact instant a physical switch moves, so it does not measure total hardware latency. The result is still useful for a playful check or for comparing the same keyboard, browser and computer before and after a change.
The test keeps the interaction simple while showing enough context to explain why one run may differ from another.
Compare the centre and range of the browser-processing sample instead of trusting one unusually fast key event.
See whether the measured event delay stays close together or varies while the system is busy.
Review the key, browser delay and gap from the preceding accepted keypress. Held-key repeat events are excluded.
The shortest received gap is converted to a playful event-rate estimate; it is not labelled as a verified USB polling rate.
Use the figures for comparison, not as a certificate for a keyboard model. Repeat the same method when comparing results.
The difference between the event timestamp and the moment the page handles that event. System load and browser scheduling can affect it.
The time between consecutive accepted keydown events. Human tapping speed is part of this number, so it is not pure hardware latency.
The spread of the measured browser-delay samples. Lower jitter means the page received events more consistently during that run.
One second divided by the shortest received event gap. Rapid alternating keys produce a more useful playful estimate than holding one key.
A complete keyboard latency measurement needs an external trigger, high-speed camera or hardware analyser to record the physical key movement and visible response. This page measures only browser keyboard-event timing. Results can change with the browser, operating system, background load, power settings and connection.
These public web standards document the keyboard events and timestamps available to an online test.
Reviewed 29 July 2026Answers about accuracy, good results, gaming, wireless keyboards, polling rate, changing scores and reducing keyboard delay.
Start the 10-second sample, then press different keys at a steady pace. Alternating A and S avoids held-key repeat and gives a cleaner sequence of browser events.
It accurately reports the timing available to this page, but that signal is browser event processing—not the full physical switch-to-screen delay. Use it for fun or controlled comparisons, not a hardware certification.
There is no universal pass mark for this browser estimate. Lower and more consistent numbers generally feel better, but compare repeated runs on the same browser and computer rather than unrelated online scores.
Browser scheduling, open tabs, background programs, battery mode, wireless interference and your tapping pattern all vary. Close heavy tasks and repeat the same 10-second method.
Yes. Run several samples with the receiver or Bluetooth connection unchanged. The page cannot separate wireless transmission delay from browser and system processing.
No. Polling rate is how often a device can report to the computer, while latency is a delay through the full input path. The observed event rate here is not a verified USB polling rate.
It can matter in competitive or rhythm games, but display latency, game processing, frame rate and network delay can be larger parts of the experience. One browser number should not decide whether a keyboard is good.
Try a direct USB port, a charged wireless keyboard, current firmware or drivers, fewer background tasks and a consistent power mode. Test one change at a time and repeat the same sample.