---
title: "Audio Isolation - Extract sounds from audio/video"
method: POST
path: "/v1/ai/audio-isolation"
tags: ["audio-isolation"]
---

# Audio Isolation - Extract sounds from audio/video

`POST /v1/ai/audio-isolation`

## Request

- Base URL: `https://api.magnific.com`
- URL: `https://api.magnific.com/v1/ai/audio-isolation`
- Auth: API key in header `x-magnific-api-key`

Isolate and extract specific sounds from audio or video files using SAM Audio AI technology.
Describe the sound you want to isolate, and the API separates it from background noise.

**Use cases:**
- Extract speech from noisy recordings
- Isolate musical instruments from a mix
- Separate specific sound effects from video audio
- Remove background noise while preserving target sounds

**Input options:**
- Provide either an `audio` URL/base64 or a `video` URL/base64 (mutually exclusive)
- Supported audio formats: WAV, MP3, FLAC, OGG, M4A
- Supported video formats: MP4, MOV, WEBM, AVI
- For video input, use bounding box coordinates (x1, y1, x2, y2) to focus on a specific region

**Output:** WAV audio file containing the isolated sound

## Request body

- Request14
  - `description` string, required: Text description of the sound to isolate from the input audio or video. Be specific about the type of sound you want to extract.
  - `audio` string: Audio input - either a publicly accessible HTTPS URL or base64 encoded audio. Supported formats: WAV, MP3, FLAC, OGG, M4A. Mutually exclusive with `video`.
  - `video` string: Video input - either a publicly accessible HTTPS URL or base64 encoded video. Supported formats: MP4, MOV, WEBM, AVI. Mutually exclusive with `audio`.
  - `x1` integer: Bounding box x1 coordinate (left edge) in pixels for sound source localization. Only applicable for video input.
  - `y1` integer: Bounding box y1 coordinate (top edge) in pixels for sound source localization. Only applicable for video input.
  - `x2` integer: Bounding box x2 coordinate (right edge) in pixels for sound source localization. Only applicable for video input.
  - `y2` integer: Bounding box y2 coordinate (bottom edge) in pixels for sound source localization. Only applicable for video input.
  - `sample_fps` number: Frame sampling rate in FPS for video processing. Higher values provide more temporal precision but increase processing time.
  - `reranking_candidates` integer: Number of reranking candidates for quality vs latency trade-off. Higher values produce better quality but slower processing.
  - `predict_spans` boolean: Enable span prediction for better isolation of non-ambient sounds like speech or musical notes.
  - `webhook_url` string, uri: Optional callback URL that will receive asynchronous notifications whenever the task changes status. The payload sent to this URL is the same as the corresponding GET endpoint response, but without the data field.

## Response `200`

OK - The task exists and the status is returned

- InlineObject4
  - `data` TaskDetail1, required
    - `task_id` string, uuid, required: Task identifier
    - `status` 'CREATED' | 'IN_PROGRESS' | 'COMPLETED' | 'FAILED', required: Task status
    - `generated` string[], required

## Other responses

- `400`: Bad Request - The server could not understand the request due to invalid syntax.
- `401`: Unauthorized - The client must authenticate itself to get the requested response.
- `500`: Internal Server Error - The server has encountered a situation it doesn't know how to handle.
- `503`: Service Unavailable

---

[API](https://skmtc.dev/magnific/apis/magnific-api.md) · [All operations](https://skmtc.dev/magnific/apis/magnific-api/llms.txt) · [OpenAPI document](https://skmtc.dev/magnific/apis/magnific-api/revisions/60711fb082c7?raw)