Make sure it's plugged in (or not disabled, for a built-in mic), then reload the page — browsers usually only scan for devices once per page load. If you just plugged it in, unplug and replug, then reload.

Close video call apps (Zoom, Teams, Meet), recording software, or other browser tabs that might be holding the microphone open, then try again. Some laptops also have a hardware mute switch worth checking.

Look for a microphone icon in the address bar and choose "Allow." If it's still blocked, check your OS-level privacy settings (macOS: System Settings → Privacy & Security → Microphone; Windows: Settings → Privacy & security → Microphone) and make sure your browser is allowed there too.

Try the "raw microphone" toggle — most browsers apply noise suppression and automatic gain control by default, which can flatten quiet voices on the meter. Also check you picked the right input device if you have more than one.

No. The meter, the waveform, and the playback recording all run and stay on your device. Nothing is uploaded, and the recording disappears the moment you leave or reload the page — it's never saved anywhere.

Make sure it's plugged in (or not disabled, for a built-in camera) and reload the page — browsers usually only scan for cameras once per page load. If another app was just using it, close that app first.

Some laptops have a physical camera cover or privacy shutter — check for one. Otherwise it's usually another app holding the camera open (Zoom, Teams, Meet, or another browser tab); close it and try again.

Some browsers (notably iOS Safari) pause the camera when the tab loses focus and don't resume it automatically. Use the "restart camera" button that appears when you come back — it's expected behavior, not a broken camera.

It's a live measured frame rate, counted from the actual video frames your camera is delivering — not just the number it reports it supports. It can dip under low light (many webcams slow down to gather more light) or if your computer is busy.

No. The live preview and everything you capture stay on your device — a captured photo downloads straight to your computer or phone as a file, it never passes through a server. Turn the camera off and the recording indicator goes with it.

A handful of keys — media/volume keys, Fn, and most Windows/Cmd key combinations — are intercepted by your operating system or browser before a webpage ever sees them. That's normal for every website, not a bug in this tool or your keyboard.

It's the largest number of keys this page has seen held down at the exact same time. Cheaper keyboards can only reliably register 2–3 keys at once (common in games and fast typing); mechanical and gaming keyboards often support 6 or more, or full N-key rollover.

That's key chatter — a classic sign of a dying or dirty switch. This tool flags it when it sees two presses of the same key within about 30 milliseconds with no release in between, which is far faster than a real second tap. Cleaning the switch sometimes helps; persistent chatter usually means it's time to replace that key or keyboard.

On a touchscreen device, the visual key-by-key layout switches to a simplified typing box automatically — tap it and type using your on-screen keyboard. Rollover and chatter detection need a physical keyboard and can't be meaningfully tested this way.

No. Every keystroke is processed on your device just long enough to light up the matching key and update the counters on screen — nothing is stored, logged, or uploaded, so it's safe to use this to check letter and number keys too.

When you click once but the page sees two clicks a few milliseconds apart, that's usually a worn or dirty microswitch bouncing — registering one physical click as two. This tool flags any gap under about 80 milliseconds between clicks as a likely bounce; a healthy mouse and a deliberate double-click both land well above that.

Click it inside the test area above. Extra side buttons (back/forward) aren't standardized across mice and drivers — some report as browser back/forward navigation instead of a button press this page can see, which is a driver/OS quirk, not necessarily a broken button.

Trackpads report smooth, inertial scrolling (many small steps that keep drifting after you lift your fingers) instead of the discrete "clicks" a physical mouse wheel sends — the readout labels which pattern it's seeing. This test reads a mouse wheel's individual steps most reliably.

It's an honest estimate, not a precise measurement — browsers can align or batch pointer-move events to the screen's refresh rate, which caps what a webpage can observe. Treat it as a rough guide (is it closer to 125Hz or 1000Hz), not a spec-sheet-accurate reading.

No. Every click, scroll, and movement here is processed on your device only long enough to draw the trail and update the counters — nothing is stored, logged, or uploaded.