What does the chatter detector measure?

It flags possible key chatter when two keydown events of the same key occur within a short threshold (default 35 ms) without an expected gap. Fast intentional double-taps can appear similar; lower the threshold to be stricter.

Why do repeat delay/rate results vary between machines?

OS keyboard settings, firmware and browser scheduling affect repeat timings. Use the same key and browser for comparison and treat results as approximate.

Can this page prove true NKRO (n-key rollover)?

It can reveal practical rollover limits by testing chords, but OS and browser event layers can mask or coalesce inputs. Hardware-level NKRO cannot be guaranteed from the web.

Does this collect or upload keystrokes?

No. All processing happens locally in your browser. The optional Share Link encodes a summary into the URL hash; nothing is uploaded by this page.

Why are some modifier keys not repeating?

Many systems do not generate auto-repeats for modifiers (Shift/Ctrl/Alt/Meta). Use a regular character key when measuring repeat delay and rate.

Advanced Keyboard Tester – Chatter, Repeat Rate & Event Inspector

Event Inspector

Max rows:

#	ts (ms)	type	key	code	loc	repeat	mods	isComp

Shows raw keydown/keyup data (scancode-ish via event.code). Use to debug layout, modifier state and repeat behaviour.

Chatter / Double-Press Detector

Threshold (ms):

Type normally. We flag a possible chatter if two keydowns of the same key occur within the threshold without a meaningful gap.

code	presses	suspected	minΔ (ms)	p95Δ (ms)	status

Heuristic only: very fast intentional double-taps can look like chatter; lower the threshold to be stricter.

Key Repeat Delay & Rate

Press Start, then pick a key and hold it down until results appear. We measure:

Initial repeat delay — gap until repeats begin.
Sustained repeat rate — average frequency of repeated keydown events (Hz).

Key:

—

Initial delay:

—

Sustained rate:

—

Samples:

Notes: OS settings and firmware affect repeat timing. Some apps suppress repeats for modifiers (e.g., Shift). Use a character key for consistent results.

Combo / Chord Tester

Click Record, press your shortcut (e.g. Ctrl+Alt+K), then replay it a few times. We’ll log success if all required keys are down together.

Target chord: —

Currently pressed: —

Attempts:

Successes:

Success rate:

—

If success rate is inconsistent, your board may lack sufficient rollover for this chord or certain rows/columns conflict (matrix ghosting).

How to use this page

Click the page to give it focus, then press Start Capture.
Open Event Inspector and press a few keys. You should see keydown/keyup with codes and modifiers.
Type normally under Chatter Detector. If you see “suspected” counts rising for a key, try cleaning or lowering the threshold.
Under Repeat, press Start, then hold a character key to measure initial delay and sustained rate.
Use Combo Tester to record a shortcut and replay it to check rollover limits.
Export JSON/CSV for support tickets or to compare boards.

Safety: avoid running global shortcuts while testing (e.g., OS hotkeys). Keep this tab focused.

Understanding your results

Chatter / Double-Press

presses: total keydown count seen.
suspected: times we saw rapid back-to-back keydowns < threshold. Rule of thumb: one-off is fine; repeatable on the same key likely indicates switch bounce or debris.
minΔ / p95Δ: smallest and 95th-percentile gap (ms) between keydowns. Very low minΔ with non-zero suspected increases suspicion.

Repeat delay & rate

Initial delay: commonly ~200–600 ms depending on OS settings.
Sustained rate: typical 10–30 Hz (keys per second). Compare within the same system/browser for best consistency.

Combos / Rollover

Success rate: if some attempts fail, your keyboard matrix may block that chord or the browser/OS filtered events.
Tip: try the same chord on different rows (e.g., Q+W+E vs J+K+L) to see pattern differences.

Browser-level limitations

Timing resolution & scheduling: event timestamps use performance.now(); busy CPUs, background tabs or power-saving can introduce jitter.
OS & firmware influence: repeat behaviour and debounce occur before the browser; results are approximate.
Modifiers & IMEs: many systems don’t auto-repeat modifiers; Input Method Editors (e.g., for East-Asian languages) may alter key sequences.
Rollover proof: the web stack can’t guarantee hardware NKRO; we infer practical limits from what the OS delivers.
Focus & global hotkeys: the page must have focus. OS/global shortcuts may consume keys before they reach the page.
Security sandbox: we can’t read raw scan codes from USB/HID without user-approved WebHID, and even then access is limited.