DeviceBenchFree space bar press counter
Spacebar counter that throws away a held key
Every press of the space bar registers here as exactly one press, however long you lean on the key: auto-repeat arrives carrying a flag that marks it, and those events are discarded and tallied separately where you can see how many there were. Click the bar on screen to start listening, then run 5, 10, 30 or 60 seconds, or leave the clock out of it entirely. What the page reads is the keydown the operating system passed on, not the switch itself, so the useful reading is not the total but the strip of gaps beneath it: an even run of taps with one interval at twice the height is a press your space bar failed to make.
- 100% free
- No signup
- Timed or untimed
- Held key ignored
- Counts releases too
Nothing is being read yet — the page scrolls on space as normal.
Releases should track presses within one, because the last one is still down while you read this. A run that ends with fewer releases than presses lost a key-up somewhere — alt-tabbing mid-press does it, and so does a keyboard that reconnects during the run.
How to test a space bar that feels like it is missing presses
Three passes, and the second one is the pass that finds the fault most people blame on the switch.
Click the bar on screen, then use the real one
Capture is off until you ask for it, because a page that grabs the space bar on load takes scrolling away from everyone who only came to read. Once the on-screen bar is armed it turns cyan, dips amber under each press, and the space key belongs to the test until you press Escape or Stop.
Hit the left end, then the middle, then the right end
Do about twenty presses at each end and twenty in the middle rather than a uniform run down the center. The switch under a space bar sits in the middle and the ends are held level by a wire stabilizer, so a bar that counts perfectly in the center and misses at one end has a stabilizer that has slipped its clip or dried out — a fault you can feel as grittiness and fix, rather than a dead switch you cannot.
Read the gap chart, not the total
The total only tells you what arrived. The strip underneath plots the time between each accepted press and the next, and an even run of taps produces an even row of bars; a single bar at roughly twice the height of its neighbors is one press that never made it. The typical gap and the count of suspicious ones are printed beside the chart once six presses are in.
Technical specifications
| Key watched | KeyboardEvent.code equal to Space — the physical key, so it counts the same on QWERTY, AZERTY and Dvorak |
|---|---|
| Auto-repeat | Excluded from the tally and shown as its own figure, so you can see how many were discarded |
| Run lengths | Four fixed windows — 5, 10, 30 and 60 seconds — plus an untimed run that ends only when you end it |
| Dropped-press flag | Any gap at least 1.8× the run's own median, applied once six presses have been recorded |
| Releases | keyup counted separately from keydown; the two should sit within one of each other |
| Stray keys | Presses on any other key are counted and left out of the total; Tab, Escape and the reload keys are not treated as strays |
| Key capture | Active only between Start and Stop, canceled by Escape, and never applied to the reload or full-screen keys |
| Gaps recorded | Up to 4,000 intervals per run, with the last 140 drawn on the strip |
Frequently asked questions
Does holding the space bar down run the count up?
No, and the page shows you what it threw away. A held key produces auto-repeat events that carry a flag marking them as repeats, and every one of those is excluded from the tally and added to a separate counter instead. That matters because the repeat rate is fast enough to fake a very good score: Windows will send up to about 30 a second after a delay of between 250 ms and a second, and macOS at its fastest setting sends one roughly every 30 ms.
How can I tell whether my space bar is really dropping presses?
Compare the gaps, not the total, and do it at speed — a failing space bar usually works fine when you press it deliberately. Tap as evenly as you can for thirty seconds and look at the interval strip: a healthy bar gives a flat row, and a contact that failed leaves one bar about twice as tall because your finger produced two presses in the time the page recorded one. The page flags anything at least 1.8 times the run's own median once it has six presses to compare against.
Why does the page stop scrolling while the test is running?
Because the space bar scrolls, and a test that let it do both would move the page out from under you on every press. While capture is on, the space key is taken from the rest of the page and given to the counter; Escape ends that, so does the Stop button, and the reload and full-screen keys are never intercepted so you always have a way out.
Why is my space bar rate lower than my click rate?
Because the key has further to travel and has to come back up further before it can fire again. A full-travel mechanical key drops about 4 mm, actuates around 2 mm down, and does not reset until it has risen back past roughly 1.9 mm, so your thumb has to move the cap through nearly two millimeters each way for every press. A mouse button's micro-switch moves a fraction of that, and it is being driven by an index finger rather than a thumb.
Presses and releases do not match. Is that a fault?
A difference of one is normal, because the last key you pressed is often still down while you read the result. A larger gap means key-up events went missing, and the usual cause is not the keyboard: switching windows mid-press, a screenshot tool taking focus, or a wireless keyboard reconnecting all leave the browser without the release. If it happens repeatedly on a wired board with nothing else running, that is worth noting alongside the press count.
It counts on my phone, but the numbers look wrong.
On-screen keyboards mostly do not report which key you touched. Android's send one anonymous event per character for nearly everything, so a page has no way to tell a space from a letter, and this one puts up a warning rather than counting guesses. The space bar is one of the few keys some mobile keyboards do identify, which is why a partial count can appear — treat any mobile figure as unusable and plug in a physical keyboard.
What is the untimed mode for?
For counting a real number of presses rather than scoring a run — reps, rows, repetitions of something you are doing away from the screen. It starts on your first press, has no deadline, and only stops when you end the run, and it still filters held keys and still charts the gaps, so it doubles as a long soak test for a key you suspect but cannot catch in ten seconds.
About the space bar, and why it fails before the other keys
The space bar wears out first because it is used most and built differently. Roughly one character in six of ordinary English prose is a space, so on a keyboard where every other key shares the remaining traffic, one key takes a share no other key comes close to. It is also the widest key on the board — 6.25 units on a standard ANSI layout, near enough 119 mm — while the switch that registers it sits under the middle. Everything to either side of that switch is held level by a wire stabilizer clipped into the plate, and the stabilizer, not the switch, is what usually goes.
That distinction is worth the thirty seconds it takes to test, because the two faults have completely different fixes. A bar that misses presses wherever you hit it has a switch problem, and on most keyboards that means a replacement board. A bar that counts perfectly in the middle and drops presses when you strike an end has a stabilizer that has come loose, lost its grease or bent — the cap tilts instead of traveling straight down, the far side of the stroke never reaches the switch, and the fix is a clip and a dab of lubricant. Neither shows up in normal typing, because your thumb naturally lands near the middle; both show up here if you deliberately work the ends. The rest of the board gets the same treatment on the keyboard test, which lights each key as it registers.
One number on this page needs its limits stated plainly. The gap flag compares each interval against the run’s own median, which means it can only see a press that was missing from an otherwise even rhythm — it cannot tell a failed contact from a hesitation, a sneeze or a moment of thinking, and it will find nothing at all in a run where you were pressing irregularly to begin with. Tap evenly, tap for long enough that a stumble is a small fraction of the run, and the flagged gaps mean something. If what you actually want to know is whether the key costs you time rather than presses, the typing speed test shows which keys are slowing real words down, and the CPS test is the mouse-button equivalent of the run you just did.
Why nothing here is written down
Every number on this page is worked out by JavaScript running in the tab you are reading it in. Nothing you type, paste or open is uploaded, logged or kept, which is also why the tools carry on working after you disconnect from the network.
This page keeps no record at all, not even a best run: each press bumps a number, adds a gap to the strip and is then gone, which is why reloading loses the lot. Nothing is assembled into text — the counter never learns what you typed, only that the space key went down and came back up.