ElevenLabs Audio Isolation (Voice Isolator)
Remove background noise, music, and ambience from recordings and keep only speech. Built on ElevenLabs Voice Isolator for cleaner dialogue and transcripts.
## Overview
Audio Isolation by ElevenLabs separates speech from an audio mix using a trained deep learning model. You upload an audio or video file, and it outputs a new track that keeps the voice while removing noise, music, and ambience. The result is easier to edit, transcribe, dub, or feed into downstream speech workflows. It's a practical fix for real-world recordings from phones, field mics, and screen captures.
## What you can build
- Clean podcast and interview audio before publishing
- Prep dialogue for speech-to-text so transcripts have fewer errors
- Pull usable speech from noisy event footage or street recordings
- Strip room tone, hiss, and background music from voice notes
- Create cleaner voice samples for post-production and review
## Inputs
- One source file to clean up, as an audio file or a video file with speech in it.
## Outputs
- One isolated-voice audio track with the speech kept and non-speech content removed.
## Recommended settings
- Use the cleanest source you have. Avoid clipping, heavy distortion, or extreme compression.
- Trim long leading silence. It helps you review results faster and spot issues sooner.
## Limitations
- The tool is not designed for isolating vocals from music tracks, even if it sometimes works.
- Crowds, overlapping speakers, and strong echo can leave artifacts or reduce separation quality.
- Poor inputs reduce quality. Heavy distortion, low-bitrate encodes, or noisy screen recordings can degrade results.
## Safety & compliance
- Only upload audio you have rights to use and process.
- Don't use isolated speech to impersonate someone, commit fraud, or mislead others.
- Don't submit content involving minors in ways that create identity or voice harm.
- Follow ElevenLabs safety rules and prohibited use restrictions for deception, civic interference, harassment, and privacy violations.
API quick start
Run audio isolation with a single API call.
{
"inputAudio": "https://your-cdn.com/input.mp3",
"inputVideo": "https://your-cdn.com/input.mp4"
}curl -X POST "https://api.wiro.ai/v1/Run/elevenlabs/audio-isolation" \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_WIRO_API_KEY" \
--data-binary @- <<'JSON'
{
"inputAudio": "https://your-cdn.com/input.mp3",
"inputVideo": "https://your-cdn.com/input.mp4"
}
JSONOther models you might like
Explore similar audio & speech models.



