Before buying, see what a Klakk platform license includes and when it is worth paying for. All 14 bundled packs are included in the website trial and the corresponding paid edition.
I’m Levin, Klakk’s developer. This guide was prepared with AI assistance. On October 12, 2026, we checked the website preview’s current mappings and analyzed its source clips offline; this was a numerical file check, without a listening session or a desktop-app latency test.
Which Klakk sound pack should I choose?
Start by previewing the 14 included packs. Compare Cherry MX Blue with Cherry MX Red, or Banana Split Stock with Banana Split Lubed, using the same phrase and output volume. Pick the sound you enjoy; there is no measured universal winner for writing, coding or focus.
This guide lists the packs shipped with the Klakk 1.4.2 website editions for Mac and Windows. Klakk adds audio when you type. It does not change the physical feel of your switches or reduce the noise your keyboard makes in the room.
The 14 included sound packs
The names below correspond to the packs bundled in the desktop apps. Some browser-menu labels are shorter: “NK Cream” is NovelKeys Cream, “Gateron Black” is Gateron Blacks — Revolt, and “Apex Pro” is Apex Pro TKL.
| Family | Included packs | What to compare |
|---|---|---|
| Cherry MX | Black (PBT), Blue (PBT), Brown (PBT), Red (PBT) | Compare the four recordings at the same volume. |
| Everglide | Crystal Purple, Oreo | Alternate the two samples while typing the same phrase. |
| Gateron | Blacks — Revolt, Browns — Revolt, Red — Revolt | Compare the three recordings on your usual headphones or speakers. |
| Banana Split | Stock, Lubed | Listen to the two versions side by side; these are separate included packs. |
| NovelKeys | Cream | Compare with one Cherry MX or Banana Split pack. |
| Other keyboards | Apex Pro TKL, Razer Blackwidow Elite | Compare these full-keyboard recordings with your favorite switch-named pack. |
These names identify sound recordings, not a promise that every physical keyboard using the same switch sounds identical. The keyboard, keycaps, desk and recording conditions also affect physical sound.
A useful comparison takes the same phrase
- Open the browser sound tester and choose a pack from the Sound Pack menu.
- Click the demo area, then type a short phrase. Include Space, Enter and Backspace to hear more than letter keys.
- Switch to a second pack and repeat the phrase without changing your output volume.
- Keep two favorites and try them in the desktop app during your own work.
Start at a low volume. Pack recordings have different levels, and your output device changes how they sound. A louder preview is not automatically a better pack.
How the current preview balances its recordings
Keep your output volume steady when switching packs, but treat that as a convenient listening comparison. It does not make every recording play at an identical digital level.
The current public preview script applies a base gain of 0.6 and a separate multiplier for each pack. For example, its combined multiplier is 0.6 for the Cherry MX packs, 0.33 for Apex Pro and Blackwidow Elite, and 2.1 for Gateron Red. The Web Audio standard defines this operation as multiplying the input samples. These settings alone do not establish equal perceived loudness.
A numerical snapshot of all 14 preview packs
We inspected the preview mappings for A, Space, Enter and Backspace: 56 mappings across 27 distinct audio files. Some packs use individual files; others select a short window from a longer recording. For the latter, we measured the selected window, not the full recording with unrelated keys and gaps.
The table shows only the A key for each pack, in the preview menu’s order. “Raw RMS” means the root mean square across all retained samples and channels, before the preview’s gain. The dB reference is a digital sample amplitude of 1.0. Quiet samples remain included. These are numerical clip measurements, not sound-pressure readings at your ears or a ranking of sound quality.
| Preview pack | A clip window / raw RMS |
|---|---|
| Cherry MX Black | 160.0 ms / -33.9 dB |
| Cherry MX Blue | 164.0 ms / -30.7 dB |
| Cherry MX Brown | 198.0 ms / -39.7 dB |
| Cherry MX Red | 158.0 ms / -33.7 dB |
| Crystal Purple | 129.0 ms / -41.0 dB |
| Everglide Oreo | 109.0 ms / -40.9 dB |
| Gateron Black | 13.6 ms / -32.1 dB |
| Gateron Brown | 92.1 ms / -39.1 dB |
| Gateron Red | 68.8 ms / -43.7 dB |
| Banana Split Lubed | 185.9 ms / -36.4 dB |
| Banana Split Stock | 198.1 ms / -37.1 dB |
| NK Cream | 184.7 ms / -45.4 dB |
| Apex Pro | 169.8 ms / -27.7 dB |
| Blackwidow Elite | 304.8 ms / -29.0 dB |
The A windows range from 13.6 to 304.8 milliseconds. That is the duration of the selected audio, including its retained tail; it is not the delay from pressing a key to hearing it. A short click and a longer clip with quiet samples are different measurement windows. Adding quiet samples can reduce RMS while leaving the original peak unchanged. Audacity’s Contrast documentation likewise distinguishes the selected region and includes silent or low-level audio in its RMS measurement.
For reproducibility, we decoded these website source files at their stored 44.1 or 48 kHz sample rate using libsndfile 1.2.2, kept their channels, and rounded timeline boundaries to the nearest sample frame. We applied no normalization, silence removal, frequency weighting or resampling. A separate calculation multiplied each isolated clip by the observed preview gain: the resulting A-key RMS values span approximately −44.2 to −30.5 dB relative to 1.0. That calculation excludes overlapping keystrokes, browser resampling, system volume and the output device. It is not a capture of the live preview.
Peak level and RMS also answer different questions. Peak checks the largest sample; RMS summarizes the chosen window. Audacity’s peak-normalization documentation describes equalizing peak amplitude. Programme loudness uses a separate measurement process, described in EBU R 128. We have not measured LUFS or true peak for these clips, and have not established which pack sounds loudest to a listener.
Use the numbers to check the comparison, then listen
Choose by whether you enjoy the sound at a comfortable level. If one pack is louder, lower its playback level during the desktop trial before deciding whether you prefer its tone. Compare the same phrase and output again; a raw RMS result cannot choose the sound you will enjoy.
A displayed percentage is a control setting for that application. It is not a universal loudness scale across apps or packs. One October 10, 2026 buyer comment asked why another app’s 100% and 30% settings sounded alike. Its maker reported changing that slider afterward; we did not test that claim. The comment describes one person’s experience with that product, not a Klakk defect or a measure of how common this issue is.
Use the browser to narrow the 14 choices, then confirm the balance in the Mac or Windows trial. The browser decodes audio for its own audio-context sample rate, as the Web Audio specification explains. This offline snapshot does not establish identical desktop playback, loudness or timing. For Windows’ separate app and system controls, see the keyboard-sound volume guide.
The browser preview runs inside the page. It cannot add typing sounds to your editor, browser tabs or other apps. Install Klakk for that.
What happens to the previous sound?
A clip’s tail is the audio after its initial impact. A new key press can arrive before that clip finishes. The recording’s duration tells you how much source audio is available; the app decides what happens when another sound starts.
One GX Mods review, posted on August 26, 2024 for version 6.3.1, described fast typing as “replays instead of continues.” That is one person’s listening preference. It does not diagnose the playback mechanism or establish the behavior of the current listed version, 6.5.1.
| Playback choice | What to listen for |
|---|---|
| Replace an earlier sound | A new trigger stops an earlier playback and starts again. |
| Allow sounds to overlap | A new playback begins while an earlier one can remain active. |
| Play a continuous typing recording | The recording continues independently of your next key press. |
These choices concern how sounds play together. Random samples or pitch variation concern which sound is chosen. A held key repeating text concerns another input behavior. Neither feature alone tells you whether earlier audio continues.
The preview and installed app have different playback paths
The current public preview script creates a separate audio source for each playback. It does not stop the previous source merely because another qualifying key starts a sound, so those sources can overlap. For a timeline pack, it plays the selected window rather than waiting through the whole source file.
Mozilla’s Web Audio documentation explains that a source node plays once, while its audio buffer can be reused by new nodes. The recording and the playback instance are different things. This source check is not a listening test or a guarantee about every browser.
The current Klakk website editions on Mac and Windows also support concurrent sounds, with finite playback capacity. Reusing that capacity can stop an earlier sound; it is not a promise that every clip always reaches its end at any typing speed. There is no user-facing tail-length or concurrent-playback-limit control. This describes the current implementation, not a measured maximum typing rate.
Buy the rhythm you want to keep hearing
Use isolated taps to hear the beginning and tail, then compare the same short phrase at your usual pace, including Space and Enter. Keep the output and comfortable volume unchanged, with only one typing-sound source enabled. Notice whether you enjoy the resulting texture rather than treating more overlap as automatically better.
Try that comparison in the installed edition before paying. The browser’s selected windows and desktop audio preparation are different paths, so this preview does not establish identical tails or timing in both. If you want sound that continues while your hands rest, compare a continuous recording or ambient player instead; Klakk does not include a background typing loop.
From preview to everyday typing
Download the Mac or Windows app. Klakk supports macOS 14 or later and Windows 10/11 on x64 or ARM64. The Windows installer chooses the matching architecture automatically.
On Mac, follow the installation and permission guide before checking your sound output. On Windows, follow the Windows setup guide. In the app, choose a sound pack and try typing in your usual editor.
Both desktop editions include the 14 packs during the 3-day free trial. Continuing after the trial costs US$6.99 once per platform. Mac and Windows are separate purchases.
Headphones, speakers and physical noise
Headphones keep Klakk’s generated audio in your ears. Speakers play it into the room. Neither choice makes the physical keyboard quieter, so a sound app cannot solve a loud switch or desk vibration problem.
For the most useful comparison, preview sounds on the output device you will actually use. If the sound feels delayed, compare with another output device before judging the pack. We do not claim a fixed latency across all computers or headphones.
Frequently asked questions
Are the packs different on Mac and Windows?
The Klakk 1.4.2 website editions for Mac and Windows include the same 14 named packs. The interfaces and installation steps differ.
Can I hear the sounds before downloading?
Yes. The keyboard sound tester lets you compare the included packs in the browser without installing the desktop app.
Do I pay separately for sound packs?
No. The 14 bundled packs are included in the trial and the one-time purchase for that platform.
Which pack is proven to improve productivity?
We have not established that any pack improves productivity. Choose by listening preference and whether you want to keep it enabled during your own work.
Does a pack change my keyboard’s switches?
No. It changes the audio Klakk plays, not the keyboard’s feel or physical noise.