Notification sound effects
A notification has to be noticed each time it plays without becoming annoying over months of use. Few sounds manage both.
Listen: 10 samples from the notification collection
Every clip shows the exact prompt that produced it. The render script reads its prompts from the same file as this page. Copy one, change a word and generate your own.
Bright ping
via audiocpp-stable-audio-3-sfxsingle bright notification ping, clean bell tone with a short shimmering decay, dry
Marimba triple
via audiocpp-stable-audio-3-sfxthree-note marimba notification, warm wooden mallet tones ascending, close and dry
Phone vibrate
via audiocpp-stable-audio-3-sfxphone vibrating on a wooden desk, buzzing rattle in two short bursts, close mic
App doorbell
via audiocpp-stable-audio-3-sfxdigital doorbell notification, friendly descending two-tone chime with a soft synthetic pad, clean
Negative alert
via audiocpp-stable-audio-3-sfxnegative UI alert, two descending detuned tones with a short buzz, unmistakably a failure, dry
Calendar reminder
via audiocpp-stable-audio-3-sfxgentle calendar reminder, soft glassy bell struck twice with a warm tail, unobtrusive, dry
Sale notification
via audiocpp-stable-audio-3-sfxcash register sale notification, bright cha-ching bell with a mechanical drawer thunk, close mic
Escalating alarm
via audiocpp-stable-audio-3-sfxescalating notification alarm, repeating tone rising in pitch and urgency over four repeats, dry
Chat pop
via audiocpp-stable-audio-3-sfxchat message pop, short water-drop style bloop with a clean bright top, dry
System startup
via audiocpp-stable-audio-3-sfxcomputer system startup chime, warm expanding synthetic chord resolving upward, spacious and clean
Prompts that work
These models respond to physical description. Name the material, the action and the space.
Soft ping
gentle notification ping, single clean sine tone, short natural decay, no reverb, under 500 ms
"Natural decay" keeps a single tone from sounding like a UI click. A notification can have a small tail; an interface tick should not.
Urgent alert
urgent notification alert, three fast repeated tones, sharp attack, slightly harsh, no reverb
The repeat count sets urgency. One tone is informational; three fast repeats signal that something needs attention now.
Email chime
email received chime, single warm bell-like tone, medium decay, subtle room air, under 700 ms
"Warm" and "bell-like" move the timbre from the thin sine of a generic ping toward a typical mail client chime, without naming a real app.
Achievement
achievement unlock notification, short ascending three-note major arpeggio, bright and rewarding, no reverb
"Major arpeggio" and "rewarding" separate a celebratory sound from a neutral confirm chime. The note count and the mode both matter.
Low battery
low battery warning tone, two low descending notes, flat and cautionary, dry, no reverb
Descending pitch in a low register reads as caution, so a battery warning doesn't sound like an error.
Making notification sounds that fit
Set urgency with repetition and register
A single soft tone reads as informational. The same timbre repeated three times fast reads as urgent, and an octave lower it reads as a warning. Decide where the notification sits (information, gentle nudge, or something the user must act on) and set it with the repeat count and pitch register. The model interprets words like "urgent" less consistently than these musical parameters.
Give it a small amount of tail
UI clicks should be as short as possible, but notification sounds benefit from a short natural decay of 300 to 700 ms. The tail makes a tone sound like a chime instead of an accidental blip, and gives it enough presence to be noticed over background noise without becoming music.
Keep the whole set in one register family
An app with five unrelated notification sounds sounds like five different apps. Generate your success, warning and alert tones from the same base timbre description (same waveform character, same rough pitch center) and vary only direction, note count and dissonance between them. A consistent set matters more than any single sound.
Why generate locally
The models run on your machine, which changes the cost, the privacy and what you can automate.
Offline after download
Weights are fetched once. After that, generation, the library and the API all work with the network off. Everything lives in one SQLite file and one folder of WAVs under ~/.soundbox, and nothing is uploaded.
Your agent can use it
Soundbox runs an MCP server in the same process, with tools for generating and listing sounds. Generated files come back as absolute paths, so an agent can copy them into a project.
Licences stated up front
Every model's terms are shown before you download it, and non-commercial weights need an explicit confirmation. Soundbox is MIT. The licence of the model that made a sound governs what you may do with it.
claude mcp add --transport http soundbox http://127.0.0.1:17494/mcpcurl -X POST http://127.0.0.1:17494/api/generate \
-H 'Content-Type: application/json' \
-d '{"model":"audiocpp-stable-audio-3-sfx","prompt":"8-bit coin pickup blip","duration_s":1}'Frequently asked questions
- Can I use these sounds in a commercial video or game?
- It depends on the model that made them. Soundbox shows each model's licence before you download its weights. The Stable Audio 3 models use the Stability AI Community License: commercial use is free below USD $1M annual revenue, and above that Stability requires an enterprise licence. ACE-Step 1.5 is MIT. MiniMax's models allow commercial use under a revenue cap and require an attribution notice. Some voice and stem-separation models are non-commercial only. Soundbox itself is MIT, but the audio is governed by the licence of the model that made it. Read that licence before you ship; this summary is not legal advice.
- Does this run offline?
- Yes, once the weights are downloaded. Soundbox runs the models locally and stores everything in one SQLite file and one folder of WAVs under ~/.soundbox. There is no account, API key or upload step. Only the initial model download needs the network.
- What hardware do I need?
- An NVIDIA GPU with 8 GB or more. CPU-only is supported, but each generation then takes minutes. That is fine for a few short clips and slow for a large batch.
- Can my coding agent generate these for me?
- Yes. While Soundbox is running it serves MCP at http://127.0.0.1:17494/mcp, with tools for generating and listing sounds. An agent can generate a sound as part of the task it is already doing and gets back an absolute file path to use in your project.
- How is this different from downloading a sample pack?
- A pack gives you unrelated pings, dings and buzzes that were not designed to sit together. Generating gives you a success chime, a warning tone and an urgent alert from one prompt family with a consistent timbre and register, so they sound like one notification system. It costs time and a GPU.