Which keys cannot be tested in a browser?

Fn, Power and many media or brightness keys are handled by firmware or the operating system and may never reach a web page. The Mac map draws fn with a dashed outline and excludes it from the tested count. Browser and system shortcuts can also intercept keys. An unlit key means this page has not received its event, not that the hardware is necessarily faulty.

Why does F11 not go full screen here?

Because the tool holds it back on purpose. Once you have engaged with it, by pressing a key or clicking the board, it suppresses the default action of F1 to F12, Tab, Space, Backspace, slash and apostrophe, so F5 cannot reload the page mid-test, Space cannot scroll the board off screen and Tab cannot jump focus away from it. Three things are never held back: anything with a Cmd or Win key down, Shift with Tab, which is how focus moves back out, and any key pressed while the cursor is in a text box or on a button. Ctrl is not exempt, so Ctrl with F5 or Ctrl with Space is held back the same as the bare key, and the browser still wins on anything it reserves for itself. Press Esc, anywhere except the typing box, to hand the keyboard back. The table under the tool lists every key it suppresses and every combination it is not allowed to touch.

What does rollover mean?

Rollover is the most keys the browser saw held down at the same moment. Press four or five keys together and the figure climbs; on many membrane keyboards it stops climbing around six, or stalls on particular combinations, because the switch matrix cannot tell some simultaneous presses apart. That ceiling is real and it is what gamers mean by n-key rollover. Read it as a floor rather than a specification, though: the operating system and the browser can each drop a key event, so the number is the most this page actually saw, not the most your hardware can do.

Does it work for a laptop keyboard or a Mac?

Use a laptop, desktop, Mac or MacBook keyboard that sends key events to the browser. Choose ANSI for a US-style map, ISO for a UK or European map, or Mac for Command and Option labels. The number pad is optional. These maps do not reproduce every laptop or compact keyboard; the Last key readout and event log still show the reported code and character.

How do I tell whether a key is actually dead?

Press each key while the test area is active and check Not yet pressed and Last key. If a key does not report, check whether firmware, the operating system or a browser shortcut intercepts it. This page records the events it receives and cannot determine which hardware part caused a missing event. Copy log exports the last 400 events with their timings.

What is the repeat rate, and how is it measured?

Hold a key down and the operating system waits, then starts repeating it. The wait is the repeat delay and the speed afterwards is the repeat rate, and they are set separately in your system settings. This tool measures only the gaps between repeats, never the first gap, because that one includes the delay and would halve the answer. It reports the median of the recent gaps, so a single dropped frame does not distort the figure.

Why does it show KeyQ when I pressed A?

Because the two are different questions. Every event carries a code, which is the physical position on the board, and a key, which is the character your software layout produced there. On an AZERTY keyboard the position labelled A is the one a US layout calls Q, so the code reads KeyQ and the key reads a. Position is what a hardware test needs, so that is what the map marks. Both values appear in the Last key row, and the log records both.

Can I test keyboard ghosting and key rollover?

Hold a combination such as W, A and Shift, then add another key and watch which physical keys stay lit. Repeat with the combinations you use. A missing key event can reveal a rollover limitation, although the operating system or browser may intercept some shortcuts. This test observes browser key events; it is not a measurement of hardware input latency.

Can I test a Chromebook, Windows or Linux keyboard?

Open this page in a browser and press keys while the test area is active. Chromebook, Windows and Linux system shortcuts can intercept some keys, just as macOS shortcuts can. The page reads reported keyboard events; it does not identify a hardware model or bypass system shortcuts.

Can I test a mechanical, compact, USB or wireless keyboard?

Yes, for the key events your connected keyboard sends to the browser. This includes installed mechanical switches and laptop, USB, Bluetooth or other wireless keyboards, including HP, Dell, Lenovo, Asus, Apple and Microsoft devices. A 60%, 65%, 75% or tenkeyless board can leave unused positions on the selected map. This does not measure switch feel, sound, electrical contacts or bare PCBs.

Can I test German, Spanish, Nordic or Japanese layouts?

The map has ANSI, ISO and Mac drawings. Code identifies a physical position, while Key shows the character your active software layout produces, including a German, Spanish, Nordic or Japanese layout when the browser reports it. The page does not offer translated instructions or an exact cap drawing for every layout.

Does this measure keyboard input delay or latency?

No. The repeat-rate figure describes the gaps between browser auto-repeat events. Event-log timestamps are browser event timestamps, not the moment a physical switch closed. Measuring hardware response time needs a separate physical reference, so these timestamps cannot establish input lag.

Can this detect chatter or double presses?

The event log shows the sequence of keydown and keyup events received by the browser. Ordinary auto-repeat is tracked separately. You can inspect unexpected extra events, but this page does not label them as switch chatter or distinguish a hardware fault from software-generated events.

Keyboard Tester

Press a key to see its physical position light on the map and stay marked after it registers. The readout records both the key code and the character your software layout produces, plus simultaneous held keys and auto-repeat rate. Choose ANSI, ISO or Mac and turn the number pad on when needed. Only events that reach this page can be tested.

Press any key. Every key you test stays lit, so a dead key is the one that never comes on.Ready. Press a key or click the board, and browser shortcuts will be held back while you test.

Layout

Numpad

Keys tested

0 / 87

Repeat rate

hold a key to measure

Rollover

nothing held yet

Keys are matched by position, not by letter: the code beside each press is what your keyboard sent, whatever character your software layout turned it into. Nothing is uploaded and nothing is stored.

Last keywhat the browser received
nothing pressed yet
Modifiers heldread from the same event
none
Rollovermost keys reported down at the same moment
nothing held yet
Not yet pressedthe keys still dark on the map
Escape F1 F2 F3 F4 F5 F6 F7 F8 F9 F10 F11 F12 Print Screen Scroll Lock Pause Backtick 1 and 69 more

Type here to see the characters your keys produce

While the cursor is in this box nothing is held back, so Tab, Space and Backspace behave normally. Keys are still recorded on the map above.

What this page holds back, and what it cannot

A web page may suppress some default actions and not others. These are the ones this tool suppresses while it has your keyboard, and the ones no page is allowed to touch. Nothing is suppressed while the cursor is in a text box or on a button.

Keys	What they normally do	In this tool
F1 to F12	Help, find, refresh and full screen. Held back here so F5 does not reload the page and F11 does not swallow the board halfway through a test.	Held back
Tab	Moves focus to the next control. Held back so a Tab press lands on the map. Shift with Tab is always let through, so focus can move back out, and Esc releases the keyboard entirely.	Held back
Space	Scrolls the page down. Held back so the board does not jump off screen every time you test the space bar.	Held back
Backspace	Used to mean back a page, and still does in some builds. Held back so a test cannot navigate away from itself.	Held back
Slash and apostrophe	Open quick find in Firefox. Held back so testing them does not drop a search bar over the page.	Held back
Meta + W	Closes the tab. The browser acts on it before the page is told, so no page can stop it.	Cannot be held back
Meta + T	Opens a new tab. Reserved by the browser in the same way.	Cannot be held back
Alt + Tab and Meta + Tab	Switch windows. The operating system takes these before the browser sees them.	Cannot be held back
Ctrl + Alt + Delete and Meta + Q	Handled by the operating system, which never passes them on to anything.	Cannot be held back
Fn, Power and most media keys	Acted on in hardware or firmware on most systems, so no key event reaches the page at all.	Cannot be held back

What a browser can and cannot see. Every ordinary key reports, and this page holds back the shortcuts it is allowed to hold back, so F5 does not reload the test halfway through. The map is a picture of a layout, not a claim about yours: what counts is the code printed beside each press, because that is what your keyboard actually sent. Every key the browser reports is recorded with the code it sent. Fn, Power and most media keys never reach a web page on most systems, so a dark one there is not a fault. Rollover is the most keys the browser saw down at once, which some keyboards cap.

Common questions

Which keys cannot be tested in a browser?: Fn, Power and many media or brightness keys are handled by firmware or the operating system and may never reach a web page. The Mac map draws fn with a dashed outline and excludes it from the tested count. Browser and system shortcuts can also intercept keys. An unlit key means this page has not received its event, not that the hardware is necessarily faulty.
Why does F11 not go full screen here?: Because the tool holds it back on purpose. Once you have engaged with it, by pressing a key or clicking the board, it suppresses the default action of F1 to F12, Tab, Space, Backspace, slash and apostrophe, so F5 cannot reload the page mid-test, Space cannot scroll the board off screen and Tab cannot jump focus away from it. Three things are never held back: anything with a Cmd or Win key down, Shift with Tab, which is how focus moves back out, and any key pressed while the cursor is in a text box or on a button. Ctrl is not exempt, so Ctrl with F5 or Ctrl with Space is held back the same as the bare key, and the browser still wins on anything it reserves for itself. Press Esc, anywhere except the typing box, to hand the keyboard back. The table under the tool lists every key it suppresses and every combination it is not allowed to touch.
What does rollover mean?: Rollover is the most keys the browser saw held down at the same moment. Press four or five keys together and the figure climbs; on many membrane keyboards it stops climbing around six, or stalls on particular combinations, because the switch matrix cannot tell some simultaneous presses apart. That ceiling is real and it is what gamers mean by n-key rollover. Read it as a floor rather than a specification, though: the operating system and the browser can each drop a key event, so the number is the most this page actually saw, not the most your hardware can do.
Does it work for a laptop keyboard or a Mac?: Use a laptop, desktop, Mac or MacBook keyboard that sends key events to the browser. Choose ANSI for a US-style map, ISO for a UK or European map, or Mac for Command and Option labels. The number pad is optional. These maps do not reproduce every laptop or compact keyboard; the Last key readout and event log still show the reported code and character.
How do I tell whether a key is actually dead?: Press each key while the test area is active and check Not yet pressed and Last key. If a key does not report, check whether firmware, the operating system or a browser shortcut intercepts it. This page records the events it receives and cannot determine which hardware part caused a missing event. Copy log exports the last 400 events with their timings.
What is the repeat rate, and how is it measured?: Hold a key down and the operating system waits, then starts repeating it. The wait is the repeat delay and the speed afterwards is the repeat rate, and they are set separately in your system settings. This tool measures only the gaps between repeats, never the first gap, because that one includes the delay and would halve the answer. It reports the median of the recent gaps, so a single dropped frame does not distort the figure.
Why does it show KeyQ when I pressed A?: Because the two are different questions. Every event carries a code, which is the physical position on the board, and a key, which is the character your software layout produced there. On an AZERTY keyboard the position labelled A is the one a US layout calls Q, so the code reads KeyQ and the key reads a. Position is what a hardware test needs, so that is what the map marks. Both values appear in the Last key row, and the log records both.
Can I test keyboard ghosting and key rollover?: Hold a combination such as W, A and Shift, then add another key and watch which physical keys stay lit. Repeat with the combinations you use. A missing key event can reveal a rollover limitation, although the operating system or browser may intercept some shortcuts. This test observes browser key events; it is not a measurement of hardware input latency.
Can I test a Chromebook, Windows or Linux keyboard?: Open this page in a browser and press keys while the test area is active. Chromebook, Windows and Linux system shortcuts can intercept some keys, just as macOS shortcuts can. The page reads reported keyboard events; it does not identify a hardware model or bypass system shortcuts.
Can I test a mechanical, compact, USB or wireless keyboard?: Yes, for the key events your connected keyboard sends to the browser. This includes installed mechanical switches and laptop, USB, Bluetooth or other wireless keyboards, including HP, Dell, Lenovo, Asus, Apple and Microsoft devices. A 60%, 65%, 75% or tenkeyless board can leave unused positions on the selected map. This does not measure switch feel, sound, electrical contacts or bare PCBs.
Can I test German, Spanish, Nordic or Japanese layouts?: The map has ANSI, ISO and Mac drawings. Code identifies a physical position, while Key shows the character your active software layout produces, including a German, Spanish, Nordic or Japanese layout when the browser reports it. The page does not offer translated instructions or an exact cap drawing for every layout.
Does this measure keyboard input delay or latency?: No. The repeat-rate figure describes the gaps between browser auto-repeat events. Event-log timestamps are browser event timestamps, not the moment a physical switch closed. Measuring hardware response time needs a separate physical reference, so these timestamps cannot establish input lag.
Can this detect chatter or double presses?: The event log shows the sequence of keydown and keyup events received by the browser. Ordinary auto-repeat is tracked separately. You can inspect unexpected extra events, but this page does not label them as switch chatter or distinguish a hardware fault from software-generated events.

Every key the browser reports is recorded with the code it sent. Fn, Power and most media keys never reach a web page on most systems, so a dark one there is not a fault. Rollover is the most keys the browser saw down at once, which some keyboards cap.