Press the key on sight

Notes fall down four lanes toward a green hit line. Press the lane key (A, S, D, or G) the moment its note arrives. This is the keyboard counterpart to the click targets: instead of clicking where a match appears, you detect a note at the line and fire a keypress. Pressing a key from a detection is the Pro keyboard actuator, and you get a full free preview in Observe (dry run) before anything touches a real key.

keyboard press on sight falling notes Pro · free preview 15 min

Key Cascade is built for exactly this. Notes are pure flat blue blocks with no text and no gradient, so the cleanest way to lock onto one is its colour: match the blue itself, not a picture of it. A picture would have to carry the note's rounded corners and glow, and those shift a hair each frame as the note falls; the flat fill never does. The green hit line then gives you an obvious place to fire.

Open the target first. It opens in a new tab. Point QHuman at one lane and read its score as you go.

Open Key Cascade

Four lanes, four keys: A S D G from left to right (F is the focus key, so the fourth lane is G, not F). A note that reaches the green line and gets its key pressed in time is a hit and clears; one that slides past unpressed is a miss. Space starts the chart, and pauses it with the notes frozen in place so one holds still while you pick its colour and draw the region. R resets every counter and restarts from the top. The chart is the same 44 notes every run, 11 per lane, and it is over in about 27 seconds.

Build the detection

  1. Pause the chart to freeze a note Start Key Cascade with Space, then press Space again the moment a note is partway down lane A on the left. The notes stop where they are, with no dimming and no overlay over the board.
  2. Build two steps: a watcher and a presser In a New Detection named key cascade, build two steps, not one: name the first Lane A (it only watches), the second Press A (it only presses). Why two? A step re-runs its Do list every time it looks, so a keypress living in the watch step would fire on every look, note on the line or not, and spam the key. Keep them apart: one step waits for the note, the other taps the key.
  3. Wire the watcher to fire on the line Double-click Lane A and leave its Do list empty: a watcher never acts, it only waits. Route it to fire on the line: Go to step: Press A, Move on: When a target is found / absent. Set Find: to Pixel colour: a falling note is a flat block of one colour, which a colour match reads far more reliably than a picture of a moving, glowing note. Pick colour... off the blue note face, or type its hex #4e98c6 (the target's header lists it); leave Move on when the target on is found on screen. Nudge Minimum area up to a few hundred pixels so a lone speck of blue cannot trip it, only a whole note face. Now the region, which is the whole lab: tick Limit detection to a screen region, Select region on screen..., and drag a shallow band across lane A alone, centred on the green line. Every note is the same blue, so the region is the only thing that says now: a full-lane region fires the instant a note appears at the top, four seconds early, which the board counts as a Wrong / early press.
    The green band around the line is the real hit window: 55 pixels either side. A note falls at a fixed 230 pixels a second, so it is inside that window for a touch under half a second, which is all the time your press has to land. Match the region to the band and the timing takes care of itself.
  4. Give the presser its keypress Double-click Press A and add Add advanced ▾ → Press a key.... Next to Key: tap A (Esc cancels), and leave Hold key for: at 0.00 s, a quick tap. Then loop back to the watcher: Go to step: Lane A, Move on: Always (every time).
    Hold key for is for games that need a key held down; a rhythm note wants 0.00. OK stays refused until a key is actually captured.
  5. Preview it with Try it (dry run) Press Space to unfreeze the chart, then hit Try it (dry run). It runs the whole flow live without sending a keystroke: Lane A boxes each note as it reaches the green line, hands off to Press A, and logs the key it would tap, note after note. Watch the timing line up for free before any of this touches a real key.
  6. Go live and fire the real key (optional, Pro) The lab is done at step 5; this is the payoff. Sending a live keypress from a detection is Pro. Focus the Key Cascade tab, press R to restart the chart, set the flow running, and watch lane A clear itself while you sit on your hands. To bail out, throw the mouse into a screen corner: the fail-safe is on by default, under Settings → Playback if you ever turned it off.
Checkpoint

Try it (dry run) finds each note as it reaches the green line and logs the key it would press, note after note. If that lands on the line every pass, the timing is right and the lab is done, no Pro needed.

If you go live, read the right counter. Hits should reach 11, because lane A has exactly 11 notes in the chart, and Wrong / early should stay 0, because you only ever pressed with a note on the line. Misses will climb to 33 and Accuracy will settle at 25%, and that is a pass, not a failure: the counters score the whole board, and you are playing one lane of four. Your result is 11 and 0.

Make it yours

Take the other three lanes and make Accuracy mean something. Add a watcher-and-presser pair for each: same colour, same key action, only the region and the key change (S, D, G, left to right), each pair looping back on itself. With all four running, Misses finally has nowhere to go: a clean run is 44 hits, 0 misses, 0 wrong, Accuracy 100%, and a Best combo of 44. The chart is identical every run, so that number is a real score you can beat.

If it went wrong

  • Wrong / early counting up This is the one that means you got it wrong, and it is nearly always the region: it reaches above the hit window, so the key fires while the note is still on its way down. Redraw it onto the green band.
  • Hits stuck below 11 Notes are reaching the line before the press does. The window is only about half a second wide, so anything slowing the loop down eats it: drop any wait you added to the step, and keep the Key Cascade tab in the foreground, since a background tab throttles the whole chart.
  • Misses at 33 and Accuracy at 25% Not a fault. That is one lane of four played perfectly, because the counters score the whole board. Check Hits reads 11 and Wrong / early reads 0. Automate all four lanes to move the rest.
  • Nothing presses when live The Key Cascade tab is not focused, so the keystroke is going somewhere else. Click the target, press R to restart the chart, and set the flow running without touching another window.
  • The note is never found The picked colour or the tolerance is off, so the blue is not being recognised. Pause with Space, re-pick the colour straight off a frozen note face, and nudge Colour tolerance up a few points. If instead it fires on nothing, your Minimum area is too low: raise it so only a solid patch of blue counts.

Stuck?

Grab the finished flow. Download it and open QHuman's Import shared file... to load it. It runs as-is: the watch step matches the note's blue by colour, not a screen-grabbed snippet, so nothing is tied to the machine it was built on and there is nothing to recapture. Its band region is drawn for a 1920x1080 lane A at 100% zoom, so redraw it over your own lane if your window sits elsewhere. Then focus Key Cascade, press R, and preview it in Observe (dry run) for free, or run it live on Pro.

Download the solution (.qhz)

Keep going

Want the full detail behind pressing and holding keys from a detection? Read the click and press keys guide. Ready for the next lab? Next up is Read the number, stop at N, where you read a status word and a count off the screen and act on what they say. Pressing keys live from a detection is Pro; when you want it, see pricing.

Taking this to a real game

These practice targets are ordinary web pages, so every macro here just works. A real game adds three things to check first:

  • Direct input is off by defaultQHuman sends input the normal way, which reaches almost every app. Full-screen and anti-cheat-aware games often only listen to the lower-level path, so if a macro does nothing inside the game, turn on direct input (Pro).
  • Elevated windows ignore ordinary clicksA game or launcher running as administrator will not accept synthetic input from a non-elevated app. Run QHuman as administrator too, or the clicks and keypresses land nowhere.
  • Full-screen vs windowed, and play fairBorderless-windowed is the easiest to detect and click against; exclusive full-screen can hide the pixels QHuman reads, so windowed or borderless is the safe start. Automating an online game can break its rules or trip anti-cheat, up to a ban, so only automate games where you own the risk.