Audio & SpeechActive
ElevenLabs Sound Effects v3 (SFX v3)
Generate high-fidelity sound effects from a text description with ElevenLabs Sound Effects v3. Built for games, video edits, Foley, and ambient loops.
Text to Speech
Model ID
sound effects v3
Provider
elevenlabs
Updated
1793360561
wiro playground—elevenlabs/sound effects v3
Updated 1793360561
Overview
Sound Effects v3 (SFX v3) is ElevenLabs’ text-to-sound model for generating short, high-quality sound effect clips from written descriptions. It turns a single prompt into a ready-to-use audio file, which helps you prototype scenes fast and avoid digging through stock libraries. SFX v3 is designed to follow multi-part prompts more reliably than earlier versions, and it handles prompts with musical or vocal elements better.
What you can build
- Game audio packs like UI clicks, weapon shots, footsteps, and environmental ambiences
- Video sound design for trailers, social clips, and motion graphics
- Foley-style one-shots like glass breaks, doors, fabric movement, and object impacts
- Seamless background loops like rain beds, room tone, wind, engines, and crowds
- Rapid iteration boards for sound direction before a final mix
Inputs
- A written description of the sound you want, in plain language or audio terms. Keep it specific about action, setting, atmosphere, and performance. Very long prompts may be rejected.
- A target clip length, selected in whole seconds. The supported range is 1 to 30 seconds.
- An MP3 quality preset that sets sample rate and bitrate. Options include 22.05 kHz at 32 kbps, or 44.1 kHz at 32, 64, 96, or 128 kbps.
- An optional looping toggle. Turn it on when you need a seamless loop with no obvious start or end.
- An optional prompt-adherence control from 0 to 1. Higher values track the description more literally with less variation.
Outputs
- One generated sound effect as an MP3 audio file.
- The file uses the sample rate and bitrate from the selected MP3 preset.
- The audio content matches the described scene as a single clip, which can include multiple events if your prompt describes a sequence.
Limitations
- Maximum clip length is 30 seconds per generation.
- The model may not follow requests for an exact count of repetitions (example: a sound “4 times”).
- Looping is designed to blend the end into the start, but some prompts can still produce a noticeable seam.
- Vague prompts can produce generic audio. Overloaded prompts can cause missing or merged events.
- Noisy, inconsistent, or poorly structured descriptions increase the risk of artifacts like clipping, harsh highs, or unintended background sounds.
- SFX v3 availability can be limited by account access and region in ElevenLabs-hosted environments.
Safety & compliance
- Follow ElevenLabs’ Prohibited Use Policy. Don’t generate content for harm, illegal activity, or deception.
- Don’t use outputs to mislead people about real events or imply real-world danger where none exists.
- Sound Effects output is meant to be used as part of a larger production. Don’t resell or distribute generated clips as a standalone sound library.
- Treat prompts as potentially logged content in standard configurations. Use extra care with sensitive or regulated data.
Example prompts
Great starting points for sound effects v3.
A heavy wooden door creaks slowly open, then slams shut with a deep echo in an empty stone hallwayAudio & Speech
Busy city street at rush hour, cars honking, distant siren, footsteps and muffled chatterAudio & Speech
Sci-fi laser gun charging up with a rising electronic whine, then firing three sharp zapsAudio & Speech
Crackling campfire at night with popping embers, crickets chirping and a soft breeze through the treesAudio & Speech
API quick start
Run sound effects v3 with a single API call.
POST https://api.wiro.ai/v1/Run/elevenlabs/sound-effects-v3
{
"prompt": "A heavy wooden door creaks slowly open, t…",
"duration": 5,
"outputFormat": "mp3_44100_128",
"loop": "false"
}curl
curl -X POST "https://api.wiro.ai/v1/Run/elevenlabs/sound-effects-v3" \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_WIRO_API_KEY" \
--data-binary @- <<'JSON'
{
"prompt": "A heavy wooden door creaks slowly open, t…",
"duration": 5,
"outputFormat": "mp3_44100_128",
"loop": "false"
}
JSONOther models you might like
Explore similar audio & speech models.



