What the keyboard tester observes
When you press a physical key, the tools use the standard keyboard events delivered by the browser. Most visual layouts identify the physical position through KeyboardEvent.code, keep current and tested state in temporary page memory, and present only observations the browser actually supplied.
What browser events cannot inspect
The site cannot inspect switches, contacts, circuit boards, firmware settings, cables, wireless radio state, scan rate, USB descriptors, or electrical signals. Operating-system shortcuts, browser controls, keyboard layouts, Fn layers, remapping, accessibility features, extensions, and application behavior can all affect which events become visible to a page.
Why the diagnostics are conservative
A detected event confirms that an event reached this browser; an unobserved event is not automatically a failed key. Chatter screening excludes keydown events marked as normal held-key repeat and requires a tightly defined completed press-and-release sequence before it reports a possible pattern. Skipped keys remain not tested, and no browser result is presented as hardware certification.
Focus and interrupted observations
A page may receive keydown and then lose focus before keyup arrives. Live held state is therefore cleared when the page blurs or becomes hidden, and guided release checks are treated as interrupted rather than converted into hardware failures. This prevents a browser navigation or reserved shortcut from creating a confident false diagnosis.
Why comparisons matter
Testing the same symptom in another application, browser, connection, or computer changes one part of the input path. A result limited to one application points toward a different investigation than behavior that follows an external keyboard across computers. These comparisons narrow possibilities; they still do not prove a component-level cause.
How the implementation is maintained
Diagnostic state rules are kept in shared logic so the main and specialized tools apply the same definitions. The repository includes automated tests for key press and release cycles, held-key repeat, conservative chatter sequences, simultaneous-key state, focus cleanup, resets, and layout changes. Project lint, TypeScript, test, and production-build checks are used when changing the implementation. These checks verify software behavior, not physical keyboards or every browser and operating-system combination.
Responsibility and purpose
Key-Test.online is independently developed and maintained. Its purpose is to provide understandable, privacy-conscious diagnostic evidence without overstating what a browser can know. No account is required to run a test.