Transition sound effects
A transition sound should be as long as the transition. Generation lets you set that length.
Listen: 10 samples from the transition 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.
Tape stop
via audiocpp-stable-audio-3-sfxtape stop effect, sound pitching and slowing rapidly down to a halt, analogue warble, dry
Rewind
via audiocpp-stable-audio-3-sfxfast rewind effect, audio scrubbing backwards at high speed with tape flutter, close and dry
Shutter cut
via audiocpp-stable-audio-3-sfxcamera shutter snapping with a mechanical mirror slap, used as a hard scene cut, close mic, dry
Water drop
via audiocpp-stable-audio-3-sfxsingle water drop falling into still water with a clean plink and a widening ripple, close mic
Reverse into hit
via audiocpp-stable-audio-3-sfxreversed swoosh building backwards then landing on a clean percussive hit, wide stereo, cinematic
Radio tune
via audiocpp-stable-audio-3-sfxradio dial sweeping across stations, fragments of static and tones passing by, close and dry
Elastic snap
via audiocpp-stable-audio-3-sfxelastic snap transition, rubbery tensioning stretch releasing into a sharp bright snap, dry
Granular dissolve
via audiocpp-stable-audio-3-sfxgranular dissolve transition, sound shattering into fine particles that scatter and fade, wide stereo
Doppler pass
via audiocpp-stable-audio-3-sfxobject passing the listener at speed with a strong doppler shift used as a scene cut, outdoor, wide
Harp gliss
via audiocpp-stable-audio-3-sfxharp glissando sweeping upward across the full range, warm and bright, concert hall reverb
Prompts that work
These models respond to physical description. Name the material, the action and the space.
Scene swoosh
cinematic scene transition swoosh, smooth rising sweep into a soft impact, 1 second, subtle room air
"Rising sweep into a soft impact" gives the model a two-part shape, build then land, so the transition resolves.
Riser into hit
tension riser building into a sharp cinematic hit, 2 seconds, low rumble underneath, dry hit at the end
"Low rumble underneath" gives the riser cinematic weight. Without low end it sounds like a cheap synth sweep.
Glitch cut
short digital glitch transition, stuttering bitcrushed texture, fast and choppy, under 500 ms, dry
"Stuttering" and "choppy" describe rhythm as well as texture. A glitch transition needs irregular timing to sound broken.
UI slide
UI slide panel transition, thin whoosh with a soft landing thud, 300 ms, completely dry
Most prompts leave out the landing thud. Without an endpoint the sound fades, and a panel sliding into place needs an audible stop.
Downlifter
cinematic downlifter, falling pitch sweep with fading tail, 1.5 seconds, roomy, ends in silence
"Ends in silence" stops the sweep resolving into a hit, which makes the downlifter suitable for a somber or quiet cut.
Making transition sounds that fit
Describe the transition's shape
Most transition sounds take one of three shapes: build and land (riser into a hit), sweep and fade (a downlifter or a soft whoosh that recedes), or a single sharp cut (a glitch stutter or a page turn). Decide which shape your edit needs and describe it as a sequence, such as "rising sweep into a soft impact" or "falling pitch sweep with fading tail". The model follows a described arc more reliably than a mood word.
Match the transition's weight to the cut's weight
A cut between two casual UI screens wants something thin and quick: under half a second, dry, no low end. A cut between two trailer scenes wants low-frequency weight and a couple of seconds to build. A heavy cinematic riser on a UI panel slide sounds oversized, and a thin UI whoosh under a trailer cut disappears. Name the duration and the frequency weight together so they stay proportional.
Silence is a valid ending
Not every transition needs to land on a hit. Downlifters and fades that end in silence let the next shot establish itself without a stinger over it. For that, say so ("ends in silence", "fading tail", "no hit"). The model tends to add an impact at the end, especially for cinematic prompts.
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 trailer riser built for someone else's two-second build, which you then stretch or cut to fit your edit. Generating gives you one at your duration, with the weight and shape matched to the cut it covers, whether that is a UI panel sliding or a scene change in a trailer. It costs time and a GPU.