⚡ Keyboard Latency Test — Event Delay, Jitter & Response
Keyboard Timing Lab

Free Keyboard Latency Test for Event Delay, Jitter, P95, and Consistency

Collect repeated browser key-event timing samples, compare median and P95, inspect jitter and outliers, and separate event scheduling from a human reaction challenge.

100% freeNo emailNo signupLocal browser processingMobile friendly
Interactive diagnostic

Keyboard Latency Test for Event Delay, Jitter, P95, and Consistency

0/20Timing samples
Median event delta
P95 event delta
Jitter
Press Start, then press any non-modifier key 20 times.
#KeyEvent delta msNext paint msRepeat
Timing samples have not started.
12 practical capabilities

Built for a complete diagnostic workflow

01

20-sample timing challenge

Keyboard Latency Test workflow feature designed for practical browser use.

02

Event timestamp delta calculation

Keyboard Latency Test workflow feature designed for practical browser use.

03

Median timing

Keyboard Latency Test workflow feature designed for practical browser use.

04

Average timing

Keyboard Latency Test workflow feature designed for practical browser use.

05

P95 timing

Keyboard Latency Test workflow feature designed for practical browser use.

06

Standard-deviation jitter

Keyboard Latency Test workflow feature designed for practical browser use.

07

Outlier count

Keyboard Latency Test workflow feature designed for practical browser use.

08

Event-to-next-paint delay

Keyboard Latency Test workflow feature designed for practical browser use.

09

Per-key sample history

Keyboard Latency Test workflow feature designed for practical browser use.

10

Hold-repeat interval monitor

Keyboard Latency Test workflow feature designed for practical browser use.

11

Separate human reaction challenge

Keyboard Latency Test workflow feature designed for practical browser use.

12

CSV export and local history

Keyboard Latency Test workflow feature designed for practical browser use.

Practical guide

How to interpret this keyboard latency test

A browser keyboard latency test cannot isolate the complete physical path from switch actuation to pixels on the display. The page only begins measuring once the browser provides an event timestamp, so hardware scanning, wireless transport, operating-system processing, and other upstream stages are not independently visible.

The event timestamp delta compares the event timestamp with the page’s high-resolution clock. Depending on browser timestamp semantics, this can be useful for spotting unusual scheduling behavior, but it should not be treated as a universal hardware latency benchmark across browsers.

The next-paint sample measures how long the page waits from handling the keydown event until the next animation frame callback. That is a browser/render scheduling signal and can vary with display refresh rate, page load, background activity, and power-saving modes.

Use repeated samples and compare distributions rather than one number. The median is resistant to isolated spikes, P95 highlights slower events, and the standard deviation indicates consistency.

Interpretation table

What the browser can and cannot tell you

SignalWhat it representsHow to use it
Event deltaperformance.now() − event.timeStampBrowser timing signal; semantics vary
Next paintKey handler to requestAnimationFrame callbackRender scheduling, not upstream hardware
MedianMiddle sorted sampleRobust central estimate
P9595th-percentile-like sampleHighlights slow tail
JitterSample standard deviationConsistency indicator
Six-minute learning path

Build a better troubleshooting habit

01

Start with a baseline

Run the primary diagnostic before changing settings so you have a comparison point.

02

Repeat the test

One sample can be misleading. Repeat the check and compare the pattern, not only the best number.

03

Change one variable

When troubleshooting, change one device, browser, permission, cable, or setting at a time.

04

Separate browser from hardware

A browser result reflects only the parts of the input or output path that the page can observe.

05

Document the environment

Keep browser, operating system, device connection, scaling, or permission context with the result.

06

Use the next diagnostic

Open a related ready page when the result suggests a neighboring part of the device or browser stack.

Knowledge check

Three quick questions

Is one isolated sample enough for a confident diagnosis?
Should you change several troubleshooting variables at the same time?
Does a browser test always isolate every hardware stage?
Online checklist

Finish with a quality review

Local diagnostic history

Recent saved milestones on this device

Related ready pages

Continue without a dead end

Return to TestsAndTools home

Frequently asked questions

Free Keyboard Latency Test for Event Delay, Jitter, P95, and Consistency: FAQ

Does this measure the keyboard’s true hardware latency?+

No. Browser JavaScript cannot isolate every upstream hardware and operating-system stage.

Why can event delta be close to zero?+

Modern browsers often use compatible high-resolution time origins for event timestamps, so the handler may run shortly after the timestamp is assigned.

What does next-paint delay mean?+

It measures browser scheduling from the key event handler to the next animation frame callback, which includes frame timing and page scheduling.

Should I compare results across different browsers?+

Use caution because timestamp semantics, scheduling, power management, and rendering behavior can differ.

Why use median and P95?+

Median describes the center without being dominated by spikes, while P95 helps expose the slower tail of repeated samples.

Source and scope

Use the result responsibly

This tool reports what a browser can observe or calculate in the current environment. Device firmware, drivers, operating-system processing, permissions, transport, display refresh, acoustic conditions, and browser implementation can add behavior that this page cannot isolate.

External technical reference: MDN: Event.timeStamp ↗.

Free to use. No email, phone number, account, or payment required.