Keyboard Latency Checker
Latency test
Press any key while the test is active. Each sample estimates browser event processing delay (not true hardware latency).
Status: Ready
Results
Current
—
Average
—
Best
—
Worst
—
Samples
0
Press Start, then press keys to capture samples.
How to use
- Click Start test to begin sampling (inactive presses are ignored).
- Press keys normally on your keyboard while the session is active.
- Watch Current update for the latest press, and Average, Best, Worst over the whole session.
- Skim the log for the last several key codes with their millisecond values.
- Click Stop when finished, or Clear to wipe samples and the log.
FAQ
What does this latency number mean?
Each sample is roughly performance.now() minus the key event timeStamp — an estimate of how long the browser took to deliver that keydown to your page. It is not full end-to-end hardware+display latency.
Why do results vary between presses?
Browser scheduling, system load, background tabs, and USB polling all affect individual samples. Take several presses and look at the average.
Is anything uploaded?
No. Samples stay in memory in your tab only.
Can I test mouse latency too?
This tool listens for keyboard keydown events only. Mouse-specific tests need pointer event tooling.
Introduction
The Keyboard Latency Checker records a delay estimate on each key press while the test is active: roughly how long after the key event timestamp your JavaScript handler sees the keydown. Use it to compare browsers, power profiles, or wired vs wireless on the same machine — not as an absolute hardware spec.
Purpose
- Spot unusually noisy or slow event delivery in the browser layer.
- Build intuition that a few milliseconds spread is normal.
Key Features
- Start / Stop / Clear session controls with latest sample plus aggregate stats.
- Rolling log of recent key codes and delay estimates in milliseconds.
- Transparent disclaimer: measures browser scheduling delay, not display chain latency.
Common Use Cases
- Comparing Chromium vs Safari vs Firefox on the same machine.
- Seeing whether “game mode” or plugged-in power changes event jitter.
Best Practices
- Collect 20+ presses before trusting averages; single spikes happen.
- Avoid running while heavy downloads or thermal throttling is active.
Comparison metrics
| Metric | Meaning |
|---|---|
| Current | Latest sample — most volatile. |
| Average | Mean of all samples in the session — best single number for A/B. |
| Best / Worst | Min / max observed delay — shows spread. |
| Log | Recent key codes with ms values — helps spot outliers by key. |
Treat values as relative browser timing, not monitor input lag or esports-grade latency.