Game over sound effects

A game over sound is the reverse of a power-up: descending, in a minor key, and slower than you might expect.

Listen: 11 samples from the game over 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.

  • Sung defeat chant

    via audiocpp-ace-step

    dark orchestral-hybrid game over theme, slow dirge tempo, mournful choir vocals, deep strings and low brass, minor key, cinematic

  • Sad trombone

    via audiocpp-stable-audio-3-sfx

    comedic sad trombone wah wah wah, four descending muted brass notes, dry

  • Slow fade to black

    via audiocpp-stable-audio-3-sfx

    game over fade to black, warm sustained chord slowly detuning downward into silence, roomy

  • 8-bit death spiral

    via audiocpp-stable-audio-3-sfx

    8-bit death sound, fast descending square wave pitch spiral ending in a low noise thud, dry

  • Heartbeat stops

    via audiocpp-stable-audio-3-sfx

    two slow heartbeats then a long flat electronic monitor tone, clinical and close

  • Shatter fail

    via audiocpp-stable-audio-3-sfx

    failure sting, glass shattering into a low descending piano note and a dark reverb tail

  • Orchestral defeat

    via audiocpp-stable-audio-3-sfx

    orchestral defeat sting, low strings falling in a minor cadence under a struck timpani, hall reverb

  • Retry prompt

    via audiocpp-stable-audio-3-sfx

    retry prompt tone, neutral two-note synthetic chime, unresolved interval, dry

  • System power down

    via audiocpp-stable-audio-3-sfx

    machine powering down, whirring fan and drive slowing to a stop with a descending pitch, close mic

  • Crowd groan

    via audiocpp-stable-audio-3-sfx

    small crowd groaning in disappointment, murmured aww falling away, medium room

  • Tape ejects

    via audiocpp-stable-audio-3-sfx

    VHS tape mechanism whirring and ejecting with a mechanical clunk, close mic, dry

Prompts that work

These models respond to physical description. Name the material, the action and the space.

  • Classic game over

    chiptune game over jingle, six-note descending minor line, square wave, slow and final, 1.2 seconds

    "Slow" and "final" together stretch the note spacing near the end. A game over jingle has to sound like it is stopping as well as falling in pitch.

  • Death stinger

    player death stinger, short descending synth sweep with a low thud landing, 700 ms, dry and abrupt

    "Landing" on a low thud gives the sweep a definite endpoint.

  • Defeat fanfare

    orchestral-hybrid defeat fanfare, descending minor brass-synth line with a dissonant final chord, 1.5 seconds, light reverb

    "Dissonant final chord" separates a defeat fanfare from a merely sad one. Unresolved dissonance reads as loss.

  • Flatline

    flatline defeat tone, single sustained low sine wave note fading out over 1 second, no vibrato, dry

    "No vibrato" matters because any pitch wobble on a sustained tone sounds alive. A flatline has to stay flat.

  • Boss defeated

    boss defeated fanfare, dramatic descending-then-resolving synth line, 2 seconds, triumphant despite descent, moderate reverb

    "Resolving" after the descent is the key word. The enemy went down but the player won, so this one should end on a stable chord.

  • Sung defeat chant

    dark orchestral-hybrid game over theme, slow dirge tempo, mournful choir vocals, deep strings and low brass, minor key, cinematic

    This one uses ACE-Step, a music model. The prompt sets the genre and instrumentation, and a separate lyrics field ("[verse]\nGame over\nYou have fallen") carries the words a synthetic choir sings. A game over screen is one of the few places where a sung line sounds deliberate.

Making game over sounds that fit

Direction tells the model who lost

Descending pitch is the usual signal for game over, defeat and death. Ascending motion, even inside an otherwise falling line, reads as a win. For the player's loss (their own death, a timer running out), prompt for a clean descent that ends low. For an enemy's defeat, prompt for a descent that resolves upward or lands on a stable major chord.

Use pacing words for finality

Words like "slow", "final" and "landing" widen the spacing between the last notes and give the sound a definite stopping point, so it doesn't trail off. A game over sound that doesn't clearly end feels unresolved.

Decide if the moment is punitive or dramatic

A time's-up buzzer should be blunt and harsh: a flat, static pulse with no melodic content. A story-beat death wants a melodic descent and maybe dissonance, because it should land emotionally. Use different prompt shapes for the two, and decide which one the moment is first.

When to use a sung line

Every sample above except one comes from a sound-effect model: synthesized tones and swept pitches, no voice. ACE-Step is a lyrics-conditioned music model. It takes a genre and instrumentation prompt plus a separate block of lyrics, and sings them. On a game over screen, a jingle implies the loss and a choir states it. Use the choir for a story beat such as a final boss run or a permadeath. For the twentieth retry on a platformer level, a half-second stinger is the right size.

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.

Connect a coding agent
claude mcp add --transport http soundbox http://127.0.0.1:17494/mcp
Generate over HTTP
curl -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.
Why does one sample have a sung vocal instead of a jingle?
That one comes from ACE-Step, Soundbox's lyrics-conditioned vocal music model. The other samples on this page come from a sound-effect model. ACE-Step takes a genre and instrumentation prompt plus a separate lyrics block, and sings the words. Use it for a story-beat death or a permadeath, and keep short jingles for sounds the player hears fifty times a session.
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's game over sound is tuned to someone else's story beat, and in your ending it may clash with your music or feel too small. Generating lets you set the descent, the duration and whether it resolves, so a player death and a boss defeat in the same game can belong to one family and still sound different.