Keyboard Tester: Check Every Key, Rollover, and Layout Online
A keyboard tester answers one question quickly: does every key on this board still reach the computer? Press keys, watch them light up, and any key that stays dark is the one to look at. That takes about a minute on a full-size board. The interesting part starts after that, because a second question usually follows the first. When several keys are held down at once, does the keyboard report all of them? The tester below measures both, and it draws the layout of another system on the same keys when you need to compare.
What a Keyboard Tester Measures
Four different faults show up during a keyboard test, and they do not look alike on screen.
A dead key never lights up. Nothing reaches the browser when you press it, so the cause sits somewhere between the switch and the page: a worn contact, a broken trace, a damaged cable, or a program that claimed the key before the tab saw it.
A stuck key lights up and refuses to go out. The switch is not releasing, so the computer keeps receiving the press.
A chattering key fires again and again from a single push. The contacts inside the switch bounce instead of closing once, and the board reports several presses where you made one.
The fourth fault is not a single key at all. Hold three or four keys together and watch what happens. On a cheap keyboard one of them never registers, or a key you are not touching lights up instead. That is ghosting, and it has nothing to do with any individual switch.
Two states on the board separate these cases. A key glows while you hold it. Once it has registered, it keeps a quieter mark for the rest of the session. So a key that never lit is dead, a key that stays bright after release is stuck, and a key that keeps flickering is chattering. The counter above the board tracks how many distinct keys you have covered, which turns a long keyboard test into a simple score.
How to Test Every Key in One Pass
Work through the board instead of pressing keys at random. A single pass, left to right, covers everything.
- Start with the function row. Press F1 through F12 once each. On some laptops those keys double as volume and brightness controls, so check whether the plain function key is what arrives.
- Take the main block row by row. The number row, then QWERTY, then the home row, then the bottom row. Watch the tested counter move as you go.
- Do not skip the modifiers. Left and right Shift, both Ctrl keys, both Alt keys, and the Meta or Windows key each send their own code, so a fault in one does not show up in the other.
- Cover the navigation cluster and the arrows. Insert, Home, Page Up, Delete, End, and Page Down sit above the arrow keys on a full board.
- Finish with the number pad. Num Lock changes what those keys send, so test them with it on and off.
Then read the score. A full ANSI board draws 104 keys, so the counter should reach 104 of 104. Every key left dark narrows the fault to a switch, a membrane contact, or the software in between.
One habit saves time. Press each key once and move on. Repeat presses tell you nothing about coverage, and the history line below the board only keeps the last fifteen entries, so a burst of the same key pushes everything else out.
Key Rollover and Ghosting Explained
Rollover is the number of keys a keyboard can report at the same time. Two counters on this page measure it. The first shows how many keys are held down right now. The second keeps the record for the session, the largest number you have held at once. That second number is the one worth quoting.
The test is easy to run. Hold A, S, D, and Space together, then add Shift and more letters. If you are physically holding six keys and the live counter stops at three, the other presses never reached the computer. Nothing is broken. You have simply found the board's rollover ceiling, which is a property of its wiring rather than a defect.
| Rating | Keys reported at once | Typical hardware |
|---|---|---|
| 2KRO | Two guaranteed, often more in common combinations | Basic office and membrane keyboards |
| 6KRO | Six, plus the modifiers | Most mid-range and many mechanical boards |
| NKRO | Every key independently | Gaming keyboards with a diode per switch |
Ghosting is the uglier cousin. Past the rollover limit a keyboard does not always drop the extra press. Sometimes it reports a key you never touched, which is why the effect got its name. The cause is the matrix: a keyboard reads keys by scanning rows and columns, and without a diode on each switch, three keys pressed together can complete a circuit that looks exactly like a fourth key. Diode isolation, which is what N-key rollover really means, removes the false readings.
Two practical notes. Rollover is measured with the keys you actually hold, so test the combinations you use: W, A, Shift, and Space for games, Shift, Ctrl, and a letter for shortcuts. And a keyboard that passes at six keys is perfectly fine for typing; the ceiling only matters once chords get wide.
Why the Tool Reads Key Codes, Not Characters
Every press this tester records is identified by its key code, the label for the physical position of the key: KeyA, Digit1, Space, Numpad1, ShiftLeft. It does not matter which character the key produces on your system. The switch that fired is what lights up on the board.
That choice has a consequence worth knowing. If you test a QWERTZ keyboard while your system is set to QWERTY, the letters you get on screen will not match the ones printed on the keycaps. The board is not wrong. It is showing physical keys, and the characters are a layer that sits on top of them.
The character itself is not discarded. The history line records each press as a pair: the code, then the character the operating system delivered. Read it as KeyQ followed by the letter q. When a shortcut refuses to fire, that pair tells you which of two things is wrong. Either the key you are pressing is not the key the application expects, or the application is looking for a character that your current layout never produces. Developers hit this constantly with bracket and slash keys, which move between layouts and national variants.
Overlaying a Layout to Compare Systems
The picker on this page carries roughly a thousand keyboard layouts drawn from three families: Linux layouts from the XKB set, Windows layouts from the KBD files, and macOS keylayouts. Choose one and it is drawn onto the board, key by key.
Each key can carry up to four legends, and the overlay shows all of them: the base character, the Shift character, the AltGr character, and Shift with AltGr. Base and Shift are printed directly on the keycap, the way they are engraved on real hardware. The other two live in the key tooltip, which appears when you rest the pointer on a key.
| Layout | Common on | Keys that move |
|---|---|---|
| QWERTY | US and UK hardware | Nothing; this is the reference |
| QWERTZ | German and Central European boards | Y and Z swap places |
| AZERTY | French boards | A and Q swap, digits need Shift |
| Dvorak | Typing-optimised setups | All letters, home row holds the vowels |
| Colemak | Modern alternative layouts | Most letters, with punctuation promoted |
The catalog is searchable by name or identifier, grouped by source, and the whole list is browsable when you do not know what you are looking for. On load the page picks the layout that matches your browser language and marks it as automatic. If you pick something else, one entry in the list returns you to the detected layout without reloading.
This is the part of a keyboard test that other pages usually skip. Comparing legends is how you check a keyboard for a customer whose system runs a different layout, or work out why a symbol you can see on the keycap never appears on screen.
What a Browser Keyboard Test Cannot See
Two limits are worth stating plainly, because both get reported as faults when they are not.
The Fn key on a laptop almost never appears. Its job is done by the keyboard firmware before the operating system is involved, so there is no event for a page to read. The same applies to many Fn combinations: pressing Fn with a function key usually produces the action, such as a brightness step, without sending a key code at all.
System shortcuts behave the same way. Ctrl+Alt+Del and Alt+Tab are claimed by the operating system, and no browser will ever see them. Their absence in a keyboard test says nothing about your hardware.
Everything else arrives. A browser reads key events for the whole window, so no input field needs to be focused before the first press registers. The page also runs entirely on your device. Key events are used to colour keys and update the counters, then discarded. Nothing is uploaded, stored, or shared, which matters when the keyboard you are testing belongs to someone else.
Frequently Asked Questions
What is the fastest way to check a keyboard?
Press every key once, row by row, and watch the counter climb. A key that never lights up is the one that failed. The whole pass takes about a minute on a full-size board.
How many keys should register at once?
That depends on the keyboard, not on the tester. Office boards often report two to six keys together, while gaming boards with N-key rollover report every key. Hold a wide chord and read the record.
Why do my letters come out wrong while testing?
You are almost certainly testing a keyboard whose legends do not match the layout your system uses. The tester shows physical keys, so switch the overlay to the layout printed on the keycaps and the two line up.
Can the tester detect a key that repeats by itself?
Yes. A key that fires many times from one push keeps flickering on the board and fills the history line with repeated entries. That pattern is switch chatter, not a software problem.
Does a keyboard test work on a wireless keyboard?
Yes. Bluetooth and USB receivers both deliver standard key events, so a wireless board behaves exactly like a wired one as long as the connection is alive.
Is it safe to press a password on this page?
Nothing you type is captured, stored, or transmitted, so there is no password to leak. That said, the tester only reads key events; you never need to type any real text into it.
Why does a key stay green after I let go?
A key that keeps its held state after release is stuck at the hardware level, usually a switch that is not springing back. Press it several times. If it stays lit, the switch or its contact is the problem.
Can I compare two layouts side by side?
Not in one view, but the switch is quick: pick one layout, note the keys that move, then pick the other. The tooltip shows all four levels on any key you point at, so a single pass gives you the full mapping.
London · SO reputation 7991 · Badges: 5🥇40🥈48🥉 · SO member since 2010