Press and hold any key. The first auto-repeat gives your repeat delay, the ones after it give the repeat rate, and the held-keys list shows exactly what the operating system thinks is still down. Let go of everything and click "I let go": anything still listed is a stuck key. Try the keyboard ghosting test next if you want another angle on how your keyboard behaves.
Hold one key for at least three seconds
Holding
--
Press and hold any key. Let go, then click "I let go" if the key still shows as held.
Repeat intervals (ms per auto-repeat; dashed line is the median)
Keys the OS reports as held
Stuck keys
none
The repeat delay and rate come from the OS, not the keyboard. If the readings look off, this is where to change them.
| System | Repeat delay range | Repeat rate range | Where to change it |
|---|---|---|---|
| Windows 10 / 11 | 250 ms (short) to 1000 ms (long); default about 500 ms | about 2 to 30 per second; default near 30 | Control Panel > Keyboard > Speed |
| macOS | 225 ms to 1800 ms on the slider | about 0.5 to 33 per second; default around 11 | System Settings > Keyboard > Key repeat rate / Delay until repeat |
| Linux (GNOME / KDE) | 100 ms to 2000 ms; default 500 ms | default about 30 per second | Settings > Accessibility > Repeat Keys |
A key that repeats, stays down or goes silent is hard to blame on hardware or software by feel alone, so the stuck key test shows you what every press really does. Press the suspect key in your browser, watch the response, and learn within a minute whether a jammed or unresponsive key is a fault in the keyboard itself on your Windows or Mac computer.
Use this stuck key checker by clicking into the live tester and tapping the key you doubt. Press any key and the matching key on the screen lights up, turns green or stays highlighted, while each keydown event is logged. Work through letters, number keys, function keys and every modifier key so each physical key gets the same treatment. As an online keyboard tester it also doubles as a quick keyboard checker for a whole office of machines, and a key that never responds is your first red flag. A quick pass through the mechanical keyboard switch tester shows whether the problem is your keyboard or your settings.
The virtual keyboard lights up in the position of the physical key you pressed, so the on-screen keyboard works like a mirror of your hands. Compare what you did with what the display shows: a healthy key should turn green for the length of the press, a missing highlight means the press never arrived, and a highlight that lingers means the key is still reporting itself as down.
This free online keyboard test runs entirely in the browser, so there is nothing to install and your privacy stays intact because nothing you type leaves the page. It works on a laptop, a gaming keyboard or a plain office board, and hardware managers can clear a shelf of spare units in minutes instead of hours with the same key checker.
The display follows a standard QWERTY keyboard layout, yet it records the physical position of each key, so an AZERTY or other layout still lights the right spot. Typing in another language will not confuse the result, and a jammed key lights the same spot on any layout.
A stuck or repeating key usually announces itself through everyday typing problems. Look for these signs: The function key test runs entirely in your browser, so nothing you press is stored or sent anywhere.
Holding a key triggers auto-repeat, which is normal. Your key repeat rate and repeat delay come from system settings, so a held key printing many characters proves nothing. Repeat firing is different: it means extra events after you have lifted your finger, or several rapid events from one tap. The repeat rate itself is not the fault.
Press the key once, lift your finger completely and watch the display for three seconds. If the key is released in reality but still glows, the switch or its contacts have not reset, and software has been ruled out for that key.
A stuck key still sends signals, too many or never-ending ones. A dead key sends none, so it stays dark in the tester. Both look unresponsive while typing, yet a dead key points to a broken contact or software mapping, while a stuck one points to mechanical binding.
Every keycap sits on a switch that must close a circuit on the way down and open it on the way up. When anything disturbs that cycle, the key sticks. The table lists the usual causes.
| Cause | What you notice | Typical fix |
|---|---|---|
| Debris under the keycap | Key binds or returns slowly | Lift the cap and clean |
| Liquid contamination | Key repeats or stays down | Unplug, drain, dry, clean residue |
| Worn switch spring | Key returns only partway | Replace the switch |
| Oxidized contacts | Key works on some presses only | Clean contacts or replace the board |
| Damaged membrane | A key or area never registers | Replace the membrane or keyboard |
On a mechanical keyboard, the switch stem slides on a spring, and fatigue shows up after millions of strokes. A hot-swap board lets you pull the bad switch and drop in a new one, while soldered boards need a soldering iron.
A membrane keyboard presses a rubber dome onto a conductive trace. Dust and crumbs make a dome collapse unevenly, and a damaged membrane under it can bridge or break a key for good.
A spill leaves sticky residue that bridges contacts and fills the gap under the keycap. Disconnect the board at once, because sugar leaves a film that keeps causing trouble long after it dries, and an affected key shows up in the tester as lingering or repeating.
This key test takes about a minute. Press the suspect key a set number of times, count the events on the screen, and then repeat on two or three similar keys. A keyboard test only means something when you compare, so the neighbors act as your control. If the computer behaves the same in another program, the fault is probably hardware rather than software.
Neighboring keys share rows, traces and a section of membrane. If only the suspect key misbehaves, the fault sits in that one switch. If several adjacent keys misbehave, suspect a spill zone or a damaged membrane section.
You tap E exactly 20 times at a normal pace, then W and R as controls. The tester logs 27 keydown events for E but 20 each for W and R. Extra events are the logged events minus your presses:
$$\text{extra events} = 27 - 20 = 7$$
That is 7 phantom events, or 35% more than you pressed, while the neighbors sit at 0%. The E switch is chattering, and the neighbors prove the rest of the board is fine.
If single keys pass but combinations fail, you are looking at ghosting rather than a stuck switch. Hold four or more keys together to check for N-key rollover and anti-ghosting. Broader keyboard diagnostics also help when you suspect firmware.
A used 87-key board I plan to resell has a spacebar that "sometimes double-types". Before listing it, I need one answer: is it a stuck, repeating keyboard key, or is the fault in my software? I open the tester, click into the page and tap Space 36 times at roughly one tap per second, then do the same on B and N as controls. Tapping that slowly matters, because Windows only starts auto-repeat after about half a second of holding, so a repeat can't explain an extra event.
The tester logs 41 keydown events for Space and 36 each for B and N. Extra events are 41 − 36 = 5, which is 5 ÷ 36 = 13.9% more than I pressed. The controls match my taps, so the board and browser are fine. This spacebar is the problem.
I then change one thing at a time and rerun the same 36 taps:
| Round | Logged events | Extra events | Share extra |
|---|---|---|---|
| Before cleaning | 41 | 5 | 13.9% |
| After compressed air | 38 | 2 | 5.6% |
| After swab and isopropyl alcohol | 36 | 0 | 0% |
The air blast removed most of the problem, and the alcohol cleaned the residue it could not reach. My judgement call: any nonzero count at a one-second pace is a failure, so I only trust the third round once it holds on a second run of 36 taps. It does.
The edge case is the one to watch. The tester cannot see your finger lift, so an extra event from a pressed-and-held key looks identical to one from a chattering switch. If I had tapped fast, or rested a thumb on the bar, the 5 extra events could have been ordinary repeats. Remapping or accessibility software can also fake a sticking key, so check those settings first. The decision for this board: list it as tested and keep a spare hot-swap switch on hand, so a relapse costs a few cents instead of a return.
Troubleshooting should start with cheap, reversible steps, and cleaning solves most jammed keys before any repair is needed. Work from the least invasive step to the most.
Turn the keyboard upside down and try gently tapping the back to shake loose dust. Next, use a compressed air cleaner around the keycap. For sticky residue, dab a cotton swab with isopropyl alcohol once the board is unplugged and fully dry, then run the test again.
If whole groups of keys fail, inspect the keyboard connection first. Try a second USB port, then a known-good board in the same port, and finally reinstall the keyboard driver. A port, cable or driver problem is a software or wiring issue, not a stuck switch.
If cleaning fails and the damage is a soaked membrane, a replacement keyboard costs less than your time. A wireless keyboard is a fair upgrade, and routine maintenance with the same test every few months catches wear early.
Press the suspect key slowly and repeatedly in the live tester, then watch whether it lights up every time, fires more than you pressed, or stays lit after release. Enter the counts here to see the difference.
Yes, it can count repeated keydown events. Holding a key triggers normal auto-repeat, so tap slowly and look for events you did not cause.
Yes. If neighbors log exactly your presses and only one key adds extras, the fault is in that switch. If several adjacent keys misbehave, suspect a spill zone or a membrane problem.
No. A stuck key still sends signals, too many or never-ending ones, while a dead key sends none. Both look unresponsive while typing, but the repairs differ.
Often yes. Lift the keycap, clean out dust and crumbs with compressed air, and dab sticky residue with a cotton swab and isopropyl alcohol on an unplugged, dry board.
The tester logged fewer events than you pressed, so some presses never registered. That points to an intermittent or failing key rather than a repeating one.
Holding a key triggers auto-repeat after a short delay, which adds events that are not a fault. A slow tapping pace keeps repeats out of your count.