Keyboard latency test results describe the accepted browser samples. Press A for each ready target, release fully and wait for the displayed count before continuing. An early manual finish must be described by its actual sample size, not as a completed twenty-sample comparison.
Part of the Keyboard test. Run the full keyboard test first for a complete device check.
The expected sample is absent Two runs disagree The result does not explain the original symptom
Keyboard latency test
Use the keyboard latency test to collect twenty A-key event-to-animation-frame samples. Compare the median and spread under the same setup; this is a browser timing observation, not a hardware key-to-photon measurement.
Also known as: keyboard latency test, key latency test, keyboard input lag test, key delay test, keyboard latency tester, keyboard input delay test. Free, online, and no download required.
No permission neededNothing uploaded
Press the requested key when the target is ready.
01
What the keyboard latency test checks
| Check | How it is measured | Fails when |
|---|---|---|
| Physical event | KeyboardEvent.code identifies the target press. | The requested physical position never arrives. |
| Paint opportunity | Two animation frames bound the next visible update. | The paint callback is interrupted or throttled. |
| Sample summary | p50, p95 and median absolute deviation summarize twenty values. | The run contains fewer than ten valid values. |
How to run the keyboard latency test
- Prepare the device and close any application that could reserve the same input or media stream.
- Start the keyboard latency test, follow the live prompt and complete every named control or direction.
- Finish the run, record the verdict and repeat once before changing one setting or connection.
If something failed
Keep the keyboard, connection, browser, display and foreground state fixed. Press A once when the target is ready, release and wait for the new sample count. Reset before changing one condition. Record twenty accepted samples and compare median plus spread rather than the single fastest press.
This interval begins at the browser event timestamp. It does not include independently observed finger-to-switch or device-to-USB timing, and animation callbacks do not measure display scan-out. Keep keyboard latency test results separate from complete physical key-to-photon delay.
02
How to read the result
What is a good keyboard latency?
Compare like with like. The keyboard latency tester starts at a delivered key event timestamp and waits for two animation-frame callbacks. The page does not observe finger movement, switch contact, USB transport or the moment light leaves the panel, so a review's physical latency value is a different measurement.
Is the A-key target ready? Press A once, release it and wait for the sample to appear. Automatic repeat, other codes and presses while a sample is pending are excluded. Twenty accepted observations finish the run automatically; rapid repeated presses cannot replace that sample requirement.
For a key latency test comparison, keep the display, foreground browser, connection and power state unchanged. Reset before a second setup. A smaller one-off number does not establish a faster keyboard, so read the median and uncertainty rather than selecting the best sample.
Keyboard latency versus polling rate
These are different quantities. The keyboard input lag test measures an event-to-frame interval, while a nominal report frequency describes device-interface scheduling. The keyboard repeat rate test measures automatic repetition and cannot provide the physical report frequency to complete that comparison.
Does background work change the result? Record the current setup and repeat with one condition changed. Browser scheduling and frame timing can alter the observed interval even when the keyboard hardware is unchanged. The key delay test cannot partition the difference into every hidden stage.
A keyboard input delay test needs its observation point stated with the value. A camera can start at physical motion and end at visible light; this page cannot. Use the input latency guide to choose the method that matches the question instead of validating a vendor specification from this browser number.
How the keyboard latency test works
A valid non-repeat KeyA event captures event.timeStamp and schedules two animation-frame callbacks. The second callback records performance.now minus that timestamp. The sample count is bounded to twenty and automatically completes at that count; a callback is not direct evidence of the physical panel's light output. Read input latency explained before comparing measurement methods.
Use the keyboard latency test in a visible foreground tab and keep the device, connection and duration unchanged. Repeat the same action enough times to separate one missed browser event from a persistent pattern. Record the displayed value before changing one setting, then run the same sequence again.
During Keyboard latency test, the live readouts update from bounded counters and recent samples. Complete every requested direction, key, channel or phase before finishing. Reset between devices so events from an earlier run cannot alter the next verdict, and keep the tab focused whenever timing or event frequency is part of the result.
What the result can establish
Keyboard latency test results describe the accepted browser samples. Press A for each ready target, release fully and wait for the displayed count before continuing. An early manual finish must be described by its actual sample size, not as a completed twenty-sample comparison.
Use the keyboard latency test in a visible foreground tab and keep the device, connection and duration unchanged. Repeat the same action enough times to separate one missed browser event from a persistent pattern. Record the displayed value before changing one setting, then run the same sequence again.
Read pass as confirmation of the collected sample, warn as a repeatable concern that needs comparison, fail as a missing or abnormal signal with a direct next action, and inconclusive as too little valid input. The number in the detail explains which threshold or count produced the status for Keyboard latency test.
Measurement limits
This interval begins at the browser event timestamp. It does not include independently observed finger-to-switch or device-to-USB timing, and animation callbacks do not measure display scan-out. Keep keyboard latency test results separate from complete physical key-to-photon delay.
The Keyboard latency test reads only signals that the operating system and browser expose to this page. It cannot inspect physical wear, firmware queues, electrical contacts, analogue circuitry or another application's private processing. A clean run confirms the observed sample, while an intermittent fault may require a longer repeat under the same conditions.
Browser scheduling, power saving, wireless transport, display timing and background work can alter the values shown by Keyboard latency test. Compare results only when those conditions match. Do not treat a small difference as a hardware change unless it repeats across several controlled runs.
03
Keyboard latency test questions
What is included in this keyboard input lag test result?
The interval begins at the delivered key event timestamp and ends at the second animation-frame callback used by the page. It includes waiting and processing within that observed span. It does not observe physical actuation, USB reports or panel light output, so it is not complete hardware key-to-photon latency.
Why do results change between runs of the same key test?
Browser frame scheduling, foreground state, background work and power settings can change the measured interval. Compare twenty accepted samples with the same A-key sequence and setup. A median and spread describe the run more usefully than one minimum; a small difference still needs a controlled repeat.
Can I compare this key delay test with a camera measurement?
Not directly. A camera can observe physical movement and visible light, while this page starts at a delivered event timestamp and uses animation callbacks. Keep the method with each number. A browser comparison on the same machine cannot independently validate the keyboard's advertised physical latency.
What is a good keyboard latency?
There is no universal good hardware value established by this page. Read the median and uncertainty for twenty accepted A-key observations, then compare an unchanged setup. A shorter browser interval does not prove a faster switch, and the page cannot set a manufacturer tolerance from that number.
Does keyboard polling rate change latency?
Device reporting is one possible stage in an input route, but this page cannot isolate it or quantify a hardware setting's effect. Automatic repeat timing is also a different task. Use a suitable device-interface or physical measurement before attributing a browser interval change to keyboard polling alone.
How should I repeat the keyboard latency test fairly?
Keep the keyboard, connection, browser, display and foreground state fixed. Press A once when the target is ready, release and wait for the new sample count. Reset before changing one condition. Record twenty accepted samples and compare median plus spread rather than the single fastest press.
What can the keyboard latency test not prove?
It cannot prove physical switch delay, USB transport, complete display response or a manufacturer's latency claim. The keyboard latency test uses event timestamps and animation callbacks. Label the result with that method and consult the input latency guide before comparing it with another measurement.