Skip to main content
ModelRunner
elevenlabs avatar

ElevenLabs Audio Isolation API

elevenlabs/audio-isolation

Strip background noise and music from a recording to isolate clean, studio-quality speech, returning the isolated voice as an MP3.

remove background
0.001667 per second of output video
OpenAPI

Model Input

Input

URL of the audio file to isolate voice from (mp3, ogg, wav, m4a, aac). Must be at least 4.6 seconds long.

You need to be logged in to run this model and view results.
Log in

Model Output

Output

Loading
Generated in 4.372 seconds
Logs (1 lines)

Model Example Requests

Examples

ElevenLabs Audio Isolation API

ElevenLabs Audio Isolation is a audio-to-audio AI model by elevenlabs. On ModelRunner it runs through a REST API or via MCP from any AI assistant, at $0.001667 per second of audio.

POST https://queue.modelrunner.run/elevenlabs/audio-isolation

cURL

# Submit a request to the queue. Input fields go at the top level of the
# body. The optional reserved "metadata" object holds your own flat string
# tags — stored on the request, never sent to the model; filter later with
# GET https://queue.modelrunner.run/requests?metadata=<url-encoded JSON>.
curl -X POST https://queue.modelrunner.run/elevenlabs/audio-isolation \
  -H "Authorization: Key $MRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "audio_url": "https://media.modelrunner.ai/Hw2F9odfnP4VgOvA-harvard-clean-speech.wav",
    "metadata": {
      "project": "my-project"
    }
  }'
# → { "request_id": "...", "status_url": "...", "response_url": "..." }

# Poll status_url until "COMPLETED", then fetch the result
curl "https://queue.modelrunner.run/elevenlabs/audio-isolation/requests/$REQUEST_ID/status" \
  -H "Authorization: Key $MRUN_API_KEY"
curl "https://queue.modelrunner.run/elevenlabs/audio-isolation/requests/$REQUEST_ID" \
  -H "Authorization: Key $MRUN_API_KEY"

JavaScript

import { modelrunner } from "@modelrunner/client";

const result = await modelrunner.subscribe("elevenlabs/audio-isolation", {
  input: {
    "audio_url": "https://media.modelrunner.ai/Hw2F9odfnP4VgOvA-harvard-clean-speech.wav"
  },
});
console.log(result);

Python

import os
import requests

headers = {"Authorization": f"Key {os.environ['MRUN_API_KEY']}"}

submitted = requests.post(
    "https://queue.modelrunner.run/elevenlabs/audio-isolation",
    headers=headers,
    json={
      "audio_url": "https://media.modelrunner.ai/Hw2F9odfnP4VgOvA-harvard-clean-speech.wav"
    },
).json()

# Poll submitted["status_url"] until "COMPLETED", then:
result = requests.get(submitted["response_url"], headers=headers).json()

Input parameters

NameTypeRequiredDescription
audio_urlstring (uri)yesURL of the audio file to isolate voice from (mp3, ogg, wav, m4a, aac). Must be at least 4.6 seconds long.

Machine-readable: OpenAPI schema · llms.txt

Use ElevenLabs Audio Isolation from Claude & Cursor (MCP)

Point Claude Code, Claude Desktop, Cursor, or any MCP client at the ModelRunner MCP server and ElevenLabs Audio Isolation becomes a tool your assistant can call directly — it authorizes via OAuth (no API key in config) and runs this model with the run_model tool using the endpoint elevenlabs/audio-isolation.

MCP client config (Claude Desktop, Cursor)

{
  "mcpServers": {
    "modelrunner": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://mcp.modelrunner.run/mcp"]
    }
  }
}

Claude Code

claude mcp add --transport http modelrunner https://mcp.modelrunner.run/mcp

Then ask your assistant, for example: “Run elevenlabs/audio-isolation on ModelRunner to generate audio”. MCP setup guide.

Model Details

Model Details

ElevenLabs Audio Isolation removes background noise, music, and ambience from a recording and returns the isolated voice as a hosted MP3. Pass a URL to an audio file (mp3, ogg, wav, m4a, aac) and it returns studio-quality speech with the same duration as the source. Its strength is salvaging usable dialogue from noisy field recordings, podcasts, interviews, and screen captures without a manual de-noise pass.

## Best for - Cleaning up noisy interview, podcast, or voice-memo recordings into clear speech - Removing background music or ambience so only the spoken voice remains - Restoring usable dialogue from low-quality or far-field recordings - Pre-processing field recordings before transcription or voice cloning

## Choose another model when - You want a written transcript of the speech rather than cleaned audio — use a speech-to-text model - You want to generate brand-new speech from text — use a text-to-speech model - Your clip is shorter than 4.6 seconds — the model rejects media below that minimum duration

## Tips - Provide the recording as `audio_url`; the output is an MP3 regardless of the input format - The input must be at least 4.6 seconds long, or the request is rejected - Output duration matches the input, so trim long files first if you only need a section

## Limitations - Heavily clipped or extremely low-volume voices may stay partly buried under residual noise - Overlapping speakers are isolated together, not separated into individual tracks

To run via the ModelRunner JavaScript client: ```js import { modelrunner } from "@modelrunner/client";

const result = await modelrunner.subscribe("elevenlabs/audio-isolation", { input: { audio_url: "https://media.modelrunner.ai/Hw2F9odfnP4VgOvA-harvard-clean-speech.wav", }, }); ```