Retro game sound effects, made on your machine
Arcade and early-console audio has a specific grain. This page shows how to describe it to a model.
Listen: 10 samples from the retro game 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.
Pause toggle
via audiocpp-stable-audio-3-sfxretro arcade pause sound, single descending two-tone chirp, pulse wave, dry
Extra life
via audiocpp-stable-audio-3-sfxretro arcade extra life jingle, rapid ascending square wave run of six notes, bright, dry
Insert coin
via audiocpp-stable-audio-3-sfxarcade cabinet coin insert, mechanical coin drop into metal slot followed by a bright electronic accept chirp
High score entry
via audiocpp-stable-audio-3-sfxretro arcade high score entry sound, cheerful looping four-note pulse wave motif, bright major key, dry
Enemy spawn
via audiocpp-stable-audio-3-sfxretro arcade enemy spawn alert, short descending noise sweep with a square wave tail, dry
Shield hit
via audiocpp-stable-audio-3-sfxretro arcade shield deflect, short metallic square wave ping with a fast pitch bend upward, dry
Warp zone
via audiocpp-stable-audio-3-sfxretro arcade warp zone, long rising pitch sweep of stacked pulse waves resolving into a bright sustained tone, dry
Attract mode loop
via audiocpp-stable-audio-3-sfxretro arcade attract mode background loop, bouncy square wave melody over a simple pulse bass, upbeat major key, 140 BPM, seamless loop
Continue countdown
via audiocpp-stable-audio-3-sfxretro arcade continue countdown, single urgent low square wave beep repeated at one second intervals, dry
Tilt alarm
via audiocpp-stable-audio-3-sfxpinball machine tilt alarm, harsh buzzing electromechanical klaxon, short bursts, dry
Prompts that work
These models respond to physical description. Name the material, the action and the space.
Menu blip
8-bit menu navigation blip, single square wave note, 80 ms, mid pitch, completely dry
"Mid pitch" keeps the blip from reading as a coin (high) or an error (low). Cursor sounds sit in the middle of the register.
Confirm chime
retro menu confirm chime, two-note square wave rise, low to high fifth, 150 ms, dry
Name the interval as a fifth. A fifth gives the open, resolved quality a confirm sound needs, distinct from the tighter climb of a coin pickup.
Level-start fanfare
chiptune level-start fanfare, four-note ascending square wave arpeggio, bright and punchy, 700 ms
"Punchy" keeps each note's attack sharp so the arpeggio doesn't blur into a pad. The sharp attack is what makes a fanfare sound energetic.
Error buzz
retro error buzz, low pulse wave held flat for 200 ms, slight detune, harsh and dry
"Held flat" matters most. A static pitch reads as a rejection; any pitch movement starts to sound like a musical note.
Boss alert
8-bit boss-appears alert, descending pulse wave siren over 600 ms, wide duty cycle, ominous but dry
"Wide duty cycle" gives the pulse wave a thicker, more aggressive timbre than the thin lead used for lasers. That timbre signals threat.
Making retro game sounds that fit
Prompt for the sound's function
Most retro-game sounds are functional signals: confirm, deny, advance, warn. Prompt for the function. Say whether the sound should read as a success or a rejection, and let pitch direction carry that meaning (rising for good, static or falling for bad). Keep melodic detail for the few sounds that are musical, like fanfares and jingles.
Keep the waveform palette consistent across a set
In a convincing retro game, every sound comes from the same two or three oscillators. Use square wave for melodic UI cues, pulse wave with a stated duty cycle for lasers and alerts, and filtered noise for hits and errors. Keep that palette across every prompt in the set so the game sounds like it has one sound chip.
State duration to control how much a sound interrupts
UI feedback that runs too long feels like it is blocking the player. Menu blips and confirms want 80 to 200 ms. Fanfares and boss alerts can run to 600 to 800 ms because they mark a real event. Put the number in the prompt. The model reads a word like "short" differently depending on the rest of 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 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 retro asset pack gives you one designer's whole UI set: their confirm, their error, their fanfare. It rarely matches your game's coin or laser sounds, which came from somewhere else. Generating lets you build every UI cue from the same prompt family as your gameplay sounds, so a menu confirm and a coin pickup share a waveform and a register.