How to run the test correctly
Choose the suspect key and press it normally 20–30 times. Release it fully each time and avoid deliberately tapping as fast as possible. The first key you press becomes locked for that sample.
Double-press diagnostic
Keyboard chatter can make one physical press appear more than once. Test a suspect key for unusually close duplicate events while excluding normal held-key repeat.
Quick test
Press the key you want to check The first detected key becomes the selected key for this sample.
Press the selected key normally 20–30 times without holding it down.
Detected presses will appear here.
Understanding the result
Chatter is an unwanted series of registrations from what the user intended as one press. It is often associated with a switch or contact that changes state more than once, but double typing can also come from software, firmware, macros, remapping, or accessibility tools.
Choose the suspect key and press it normally 20–30 times. Release it fully each time and avoid deliberately tapping as fast as possible. The first key you press becomes locked for that sample.
The tool flags separate, non-repeat keydown events for the selected key that arrive within a short timing window. This is a screening signal, not proof that a mechanical switch is defective.
Keyboard firmware, macro or remapping utilities, browser extensions, accessibility software, and operating-system input settings can change what reaches the browser. Very fast intentional tapping can also resemble a duplicate.
Repeat the test with another key, then compare in another browser and on another computer. For an external keyboard, reconnect it or try another appropriate USB port. Disable macro or remapping software temporarily where safe.
Dust, wear, switch construction, firmware debounce, and replaceable switches can affect behavior. Follow the manufacturer's cleaning or service guidance; do not disassemble powered equipment or attempt repairs outside your experience.
Browser timing is affected by the operating system, background load, and event scheduling, so it cannot measure exact switch latency. Event labels, typed characters, and timing history remain in temporary page memory and are cleared on reset or reload.