Water sound effects
Name the container and the amount of water. A drip in a cave and a drip in a sink are different sounds.
Listen: 10 samples from the water 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.
Ocean waves
via audiocpp-stable-audio-3-sfxocean waves rolling in and breaking on a sandy shore, receding foam hiss, wide open beach, seamless loop
Waterfall
via audiocpp-stable-audio-3-sfxlarge waterfall crashing into a plunge pool, dense white noise roar with spray, close outdoor
Submerging
via audiocpp-stable-audio-3-sfxhead submerging underwater, sudden muffling with bubbles rushing past the ears, enclosed and low
Boiling pot
via audiocpp-stable-audio-3-sfxwater boiling vigorously in a metal pot, rolling bubbles with lid rattling, close kitchen mic
Toilet flush
via audiocpp-stable-audio-3-sfxtoilet flushing, sudden rush of water swirling down with the cistern refilling afterwards, tiled room
Shower
via audiocpp-stable-audio-3-sfxshower running, water striking tiles and a shower tray, enclosed bathroom reverb, seamless loop
Swimming
via audiocpp-stable-audio-3-sfxperson swimming front crawl, rhythmic arm strokes splashing with breathing between, pool reverb
Fountain
via audiocpp-stable-audio-3-sfxornamental stone fountain in a plaza, layered falling water into a basin, gentle and continuous
Ice in a glass
via audiocpp-stable-audio-3-sfxice cubes dropped into a glass then liquid poured over them, cracking and fizzing, close mic
Rowing
via audiocpp-stable-audio-3-sfxwooden oars dipping and pulling through calm lake water, rhythmic splashes with rowlock creaks
Prompts that work
These models respond to physical description. Name the material, the action and the space.
Cave drip
single water drip falling into a shallow cave pool, sharp plink, long reverberant tail, close mic
"Long reverberant tail" is what the cave adds to the drip. Without it the model renders a drip in a neutral space.
Tap drip
single water drip from a kitchen tap into a metal sink, tight dry plink, no reverb
"Metal sink" and "no reverb" put the same drip on a small, hard, close surface. The vessel is what separates it from the cave version.
Pouring
water poured from a bottle into a small glass, steady stream, rising pitch as glass fills, 2 seconds, dry
"Rising pitch as glass fills" names a real effect: the air cavity shrinks as the liquid rises. It separates a convincing pour from a static stream loop.
Big splash
large splash of a person jumping into a pool, heavy impact then bubbling settle, 2 seconds, outdoor
"Heavy impact then bubbling settle" names the two phases of a splash. The single word "splash" tends to give a shorter, less physical sound.
Stream
small rocky stream flowing over stones, continuous bubbling trickle, 3 seconds, outdoor close mic
"Over stones" turns flat flowing water into a bubbling trickle. As with wind, water needs something to disturb it.
Making water sounds that fit
Name the vessel
A drip in a stone cave, a drip in a metal sink and a drip in a plastic bucket have different pitches, decay times and reverb, from the same physical event. Always tell the model what the water is in or falling into: cave pool, sink, glass, puddle, pool, bucket. The container matters more than any adjective about the water.
Pick the event
A drip, a stream, a pour and a splash are different physical events with different durations and shapes. A drip is a single sharp transient; a stream is continuous and textured; a splash has an impact phase and then a settling phase. Name the event you need instead of a "big" or "small" "water sound".
Give impacts a settle phase
Anything landing in water (a jump, a dropped object, a thrown bucket) has a loud impact followed by several hundred milliseconds of bubbling and dripping as the water settles. Name both phases ("heavy impact then bubbling settle"). Otherwise the model often cuts the sound off right after the transient and the splash sounds clipped.
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 a fixed drip, pour or splash recorded in someone else's kitchen or pool, and you adjust your scene to fit the clip. Generating lets you specify the vessel, the amount of water and the phases you need (a drip in a cave or a sink, a small pour or a bucket dump), and regenerate if the first take doesn't match. It costs time and a GPU.