Why use this tool?
- Measure repeated keyboard-to-browser frame response without installing software.
- See typical delay separately from occasional slow outliers.
- Compare browsers, display refresh rates, power modes, and background-load conditions.
Keyboard Latency Test
See every key press and plot keyboard event-to-frame latency on a live line graph, with median, average, fastest, slowest, P95, and consistency results.
Timing samples stay in this browser tab
This is a browser event-to-next-frame estimate, not a laboratory measurement of total hardware input lag. Use it for controlled comparisons on the same computer, display, and browser.
Press-to-frame test
Collect at least 20 samples for a useful comparison. Up to 100 recent samples are retained.
Latest key press
—Click here, then press any key repeatedly
Held-key repeat is ignored. Press Tab to leave the test area.
Recent key presses
Newest firstPressed keys will appear here.
Waiting for samples
Use the test area above to begin.
Latency results
All figures describe DOM input-event to next animation-frame timing in this page.
- Samples
- 0
- Latest
- —
- Median
- —
- Average
- —
- Fastest
- —
- Slowest
- —
- P95
- —
- Consistency
- —
Maximum 100 recent samples
Key: —
Best typical-response comparison
Mean of all retained samples
Lowest observed sample
Highest observed sample
95% of samples are this fast or faster
Standard deviation; lower is steadier
Latency over recent inputs
Each point is one measured key press.
The latency line graph will appear after your first input.
Complete tool guide
Keyboard Latency Test guide
Use this online keyboard latency test to estimate how long the current browser takes to carry a keydown event into the next painted frame, then compare the median, average, 95th percentile, and consistency across repeated presses.
Your input is processed in the current browser tab and is not sent to the application server.
Common use cases
- Checking browser responsiveness before a rhythm game or competitive game
- Comparing a wired, wireless, or Bluetooth keyboard under the same conditions
- Finding browser, extension, power-saving, or main-thread performance problems
How to use it
- 1
Click the keyboard test area so it has focus.
- 2
Press and release any ordinary key at least 20 times; held-key repeats are ignored.
- 3
Review the median for typical response and P95 for slower outliers.
- 4
Reset, change one condition, and repeat with the same browser and display.
Controls and results explained
Keyboard test area
The page records keydown events only while this area is focused. Tab is never blocked, so you can leave the tester using standard keyboard navigation.
Latest sample
The newest event-to-next-frame estimate. One result can vary with the exact point in the display refresh cycle, so judge a run from at least 20 samples.
Median and average
Median is the middle result and the best headline for typical response. Average includes every sample and moves more when a background task causes a large delay.
Fastest and slowest
These show the range observed in the current run. The fastest result is not a guaranteed hardware minimum, and one unusually slow result may be a scheduling interruption.
P95 latency
The 95th percentile is at least as fast as roughly 95% of recorded samples. It exposes recurring slow frames that a good median can hide.
Consistency
This is the standard deviation of the samples. A smaller number means response is steadier; a larger number indicates variable browser or system scheduling.
Pressed-key display and history
The large keycap flashes the latest key as soon as its keydown event arrives. Recent key presses then list the measured delay beside each key so you can connect every point on the graph to an input.
Latency line graph
The line follows recent samples in order, each dot represents one key press, and the dashed reference shows the run median. Sharp peaks reveal slow frames that summary averages can hide.
Reset samples
Clears the current statistics and chart so you can make a controlled comparison after changing the keyboard connection, browser, refresh rate, or power mode.
What is measured
The estimate starts from the DOM keydown timestamp and ends at the next requestAnimationFrame callback, approximating browser event-to-next-paint delay.
What is not measured
A web page cannot isolate the keyboard scan, USB or Bluetooth transport, operating-system processing, display pixel response, or total physical key-to-photon latency.
Refresh-rate effect
At 60 Hz a frame lasts about 16.7 ms; at 120 Hz it lasts about 8.3 ms. The point within that cycle strongly affects each individual sample.
Practical tips for better results
- Collect at least 20–30 presses and compare medians rather than the single fastest sample.
- Keep this tab visible and close games, video, and other busy tabs before testing.
- Disable battery-saving modes and keep the same display refresh rate for comparisons.
- Change only one variable at a time, such as wired versus Bluetooth.
Frequently asked questions
Does this show my keyboard’s true hardware latency?
No. It estimates DOM-event-to-next-frame delay inside the browser. True key-to-photon hardware latency requires external high-speed measurement equipment.
Why do results jump by about one frame?
A key can arrive anywhere within the display refresh cycle. Browser scheduling can place feedback in the current or following frame, producing steps close to one frame duration.
Which number should I compare?
Use the median as typical response, P95 for common worst-case response, and consistency to judge variability. Use the same device, display, browser, and sample count.
Are the keys I press saved or uploaded?
No. Only temporary timing values and the latest key label are kept in this tab. Nothing is uploaded, and Reset or a refresh clears the run.
Keep working
Related device & browser tests
Continue with another focused tool from the same collection.
Keyboard TestTest every keyboard key on a visual layout, find dead or repeating keys, and inspect keydown, keyup, code, and modifier events.Mouse TestTest five mouse buttons, wheel directions, cursor movement, report rate, and double-click timing, including possible switch bounce.Mouse Latency TestSee each mouse click and graph button-event-to-frame latency with median, average, fastest, slowest, P95, and consistency results.Touchscreen TestDraw across a touchscreen to find dead zones or ghost touches and check multi-touch contacts, pressure, contact size, and pointer type.