Begin with one deliberate press
Open the lab in a foreground browser window and start a fresh live session. Focus the outlined capture area, press one key, release it fully and wait before pressing it again. Keep the first round slow. A repeated character seen while writing is a symptom, but this test needs a clean event sequence before you can compare possible explanations.
The record contains physical-position codes and relative event times. It deliberately omits typed text and device identifiers. A highlighted key shows that the browser exposed its code during the test; it does not establish the condition of every electrical contact inside the keyboard. Keep your test separate from passwords, messages or normal writing.
Separate three different repeat patterns
A held key commonly generates automatic repeated keydown events. The browser provides a repeat flag, and the lab counts those events separately rather than inflating the number of fresh presses. A second down event while the same code is already held is another pattern, recorded as a down without a release for review.
The rapid-repeat review window examines a new down after a release of that same code. A short interval can be worth investigating, but fast intentional tapping can produce the same observation. The configurable threshold is not a universal fault boundary. Increase the time between deliberate presses, repeat the session and compare whether unexpected pairs continue to appear.
Preserve a useful support record
Stop the session before discussing or changing the setup. Export the JSON report to keep the sequence, the selected window and the end reason together. CSV is useful for reviewing individual event times. Neither format automatically uploads to a support desk or another service; you decide where to share the downloaded file.
When comparing two sessions, keep the browser, connection method and deliberate pressing routine as similar as possible. Describe changes such as switching a USB port or reconnecting a wireless receiver in your support message. Browser evidence can help reproduce a symptom, but it cannot by itself separate switch behavior, keyboard firmware, operating-system settings and input utilities.