/api/vocal-removerRemoves vocals from an audio or video file and returns a lead vocal, backing vocal, and clean instrumental. Requires an API key — create one from your dashboard.
Content-Type: application/json · Authorization: Bearer <api-key>
| Field | Type | Description | |
|---|---|---|---|
fileUrl | string | required | Publicly accessible URL of the audio or video file to process — or a YouTube video URL to download and process directly. |
outputFormat | string | optional | Audio format for the returned files — mp3 wav flac. Defaults to mp3. |
Remove vocals from a song (default mp3 output)
curl -X POST https://karadeo.com/api/vocal-remover \
-H "Authorization: Bearer kd_..." \
-H "Content-Type: application/json" \
-d '{ "fileUrl": "https://example.com/song.mp3" }'Remove vocals → lossless FLAC output
curl -X POST https://karadeo.com/api/vocal-remover \
-H "Authorization: Bearer kd_..." \
-H "Content-Type: application/json" \
-d '{ "fileUrl": "https://example.com/song.mp3", "outputFormat": "flac" }'{
"leadVocalUrl": "https://.../lead-vocal.mp3",
"backingVocalUrl": "https://.../backing-vocal.mp3",
"instrumentalUrl": "https://.../instrumental.mp3"
}Files are exported in the requested outputFormat. If the lead/backing split step is unavailable, backingVocalUrl will be null and leadVocalUrl will be the full unsplit vocal. If separation falls back to a 2-stem engine, all files are returned as MP3 regardless of outputFormat.
All error responses return application/json with an error string field.
| Status | Condition | Body |
|---|---|---|
| 401 | Missing or invalid API key | { "error": "Authentication required" } |
| 403 | Usage limit exceeded | { "error": "Usage limit exceeded", "usage": {...} } |
| 402 | File duration exceeds remaining credits | { "error": "...", "usage": { tier, usedMinutes, limitMinutes, remainingMinutes, mediaDurationMinutes } } |
| 400 | Missing fileUrl, or can't determine duration | { "error": "..." } |
| 500 | Separation error or unexpected failure | { "error": "..." } |
Each request consumes credits from your monthly quota. Cost is calculated from the file duration, rounded up to the nearest minute, at 2× the standard rate — vocal-remover runs a second AI pass to split lead vs. backing vocal (e.g. a 3-minute song costs 6 minutes of quota).
| Tier | Monthly limit | Reset |
|---|---|---|
| free | 10 min | 1st of each calendar month (UTC) |
| subscribed (monthly) | 500 min | Billing cycle start date |
| subscribed (annual) | 6,000 min | Billing cycle start date |