Whoosh sound effects you generate yourself
A whoosh is an object moving through air, so the prompt is mostly physics: how fast it moved and how heavy it was. Soundbox runs open-weight audio models on your own GPU. You generate a whoosh at the length and weight your cut needs, and regenerate it when the edit changes, without searching a library for a clip that happens to fit.
Listen: 10 samples from the whoosh 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.
Reverse whoosh
via audiocpp-stable-audio-3-sfxreversed whoosh, air sucking backwards and accelerating into a hard silent cutoff, wide stereo
Heavy air movement
via audiocpp-stable-audio-3-sfxlarge heavy object displacing air as it swings past, thick low rush with a soft turbulent tail
Logo whoosh sting
via audiocpp-stable-audio-3-sfxbrand logo whoosh, clean airy sweep resolving into a short bright chime, polished and dry
Cape flourish
via audiocpp-stable-audio-3-sfxheavy fabric cape swept sharply through the air, cloth snapping with a broad low whoosh, close mic
Fire whoosh
via audiocpp-stable-audio-3-sfxflaming object flying past, roaring fire whoosh with a crackling doppler tail, outdoor and wide
Air burst
via audiocpp-stable-audio-3-sfxsudden compressed air burst released from a nozzle, sharp hiss with a quick decay, close and dry
Long slow whoosh
via audiocpp-stable-audio-3-sfxvery long slow cinematic whoosh, gradually building airy movement across eight seconds, wide stereo
Layered whoosh
via audiocpp-stable-audio-3-sfxlayered whoosh built from thin air, mid body and a low sub, all sweeping together, cinematic and wide
Bullet pass-by
via audiocpp-stable-audio-3-sfxbullet whipping past close to the listener, sharp cracking snap with a fast doppler zip, outdoor
Page flip whoosh
via audiocpp-stable-audio-3-sfxquick paper page flipping through air, light fluttering whoosh, close mic, completely dry
Prompts that work
These models respond to physical description. Name the material, the action and the space.
Fast swipe
fast thin whoosh, small light object passing close to the microphone, 200 ms, sharp attack, dry
"Close to the microphone" makes it read as fast. Proximity conveys speed better than a tempo word.
Heavy sweep
slow heavy whoosh, large object moving through open air, 1.5 seconds, low-frequency weight, subtle room tone
Name the frequency range ("low-frequency weight") to make the object sound heavy. Speed and mass need separate words in the prompt.
Doppler pass-by
realistic cinematic whoosh, deep object passing overhead, doppler pitch shift down as it passes, 2 seconds, roomy
"Doppler pitch shift down as it passes" names what a real pass-by does acoustically. Without it the whoosh swells and fades without moving.
UI transition
short UI swipe whoosh for screen transition, thin and clean, left-to-right pan implied, 250 ms, completely dry
Room tone on an interface whoosh clutters the mix. "Completely dry" keeps the sound out of the way of whatever plays under it.
Cartoon swish
cartoon whoosh swish, exaggerated pitch rise then fall, bright and airy, 400 ms, no low end
"No low end" stops a comedic swish from sounding cinematic. Cartoon sounds sit in the upper register.
Making whoosh sounds that fit
Set speed and mass separately
A whoosh is displaced air, and it changes with how fast the object moved and how big it was. Set both: speed as a duration ("200 ms" reads fast, "1.5 seconds" reads slow) and mass as a frequency description ("thin" and "light" for something small, "low-frequency weight" or "heavy" for something large). A prompt that sets only one drifts toward a generic mid-sized swoosh.
Use doppler for realism
A whoosh that only fades in and out sounds synthetic. When a real object passes a stationary listener, the pitch drops slightly as it passes and recedes. Writing "doppler pitch shift down as it passes" is the most reliable way to move a generation from cartoon swish to realistic pass-by.
Dry for cuts, roomy for reveals
A whoosh under a hard cut needs to be dry and short so it doesn't smear into the next shot. Say "dry" or "no reverb" and keep it under half a second. A whoosh for a slow reveal or a title card can have room tone and a longer tail. Decide which job the sound is doing before you write the prompt.
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 transition whooshes and slow for a whole edit's worth.
- 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 wiring up your UI transitions can generate the swipe sound in the same pass 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 someone else's whoosh, timed for someone else's edit. Generating gives you one at the duration your cut needs, with speed and weight matched to what moves on screen. It costs time and a GPU.