Vocal & Instrument Stem Separation

Use Suno’s official get‑stem API to split tracks created on our platform into clean vocal, accompaniment, or per‑instrument stems with state‑of‑the‑art source‑separation AI.

POST /api/v1/vocal-removal/generate

Separate Vocals from Music

cURL

curl --request POST \
  --url https://api.kie.ai/api/v1/vocal-removal/generate \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "taskId": "5c79****be8e",
  "audioId": "e231****-****-****-****-****8cadc7dc",
  "callBackUrl": "https://api.example.com/callback",
  "type": "separate_vocal"
}
'

200

{
  "code": 200,
  "msg": "success",
  "data": {
    "taskId": "5c79****be8e"
  }
}

Usage Guide

• Separate a platform‑generated mix into vocal, instrumental, and individual instrument components.
• Two processing modes are available:
  • separate_vocal — 2‑stem split
  • split_stem   — up to 12‑stem split
• Ideal for karaoke creation, remixes, sample extraction, or detailed post‑production.
• Best results on professionally mixed AI tracks with clear vocal and instrumental layers.
• Billing notice: Each call consumes credits; re‑calling the same track is charged again (no server‑side caching).
• Pricing: Check current per‑call credit costs at https://kie.ai/pricing.

Separation Mode Details

Mode (type)	Stems Returned	Typical Use	Credit Cost
separate_vocal (default)	2 stems – Vocals + Instrumental	Quick vocal removal, karaoke, basic remixes	10 Credits
split_stem	Up to 12 stems – Vocals, Backing Vocals, Drums, Bass, Guitar, Keyboard, Strings, Brass, Woodwinds, Percussion, Synth, FX/Other	Advanced mixing, remixing, sound design	50 Credits

Parameter Reference

Name	Type	Description
taskId	string	ID of the original music‑generation task
audioId	string	Which audio variation to process when multiple versions exist
type	string	Required. Separation mode: separate_vocal or split_stem

Developer Notes

• All returned audio‑file URLs remain accessible for 14 days.
• Separation quality depends on the complexity and mixing of the original track.
• separate_vocal returns 2 stems — vocals + instrumental.
• split_stem returns up to 12 independent stems — vocals, backing vocals, drums, bass, guitar, keyboard, strings, brass, woodwinds, percussion, synth, FX/other.
• Billing: Every request is charged. Re‑submitting the same track triggers a new credit deduction (no server‑side caching).

Authorizations

Authorization · string · header · required

All APIs require authentication via Bearer Token.

Get API Key:

1. Visit API Key Management Page to get your API Key

Usage: Add to request header: Authorization: Bearer YOUR_API_KEY

Note:

• Keep your API Key secure and do not share it with others
• If you suspect your API Key has been compromised, reset it immediately in the management page

Body

application/json

taskId · string · required

Unique identifier of the music generation task. This should be a taskId returned from either the "Generate Music" or "Extend Music" endpoints.

Example:

"5c79****be8e"

audioId · string · required

Unique identifier of the specific audio track to process for vocal separation. This ID is returned in the callback data after music generation completes.

Example:

"e231****-****-****-****-****8cadc7dc"

callBackUrl · string<uri> · required

The URL to receive vocal separation task completion updates. Required for all vocal separation requests.

• System will POST task status and results to this URL when vocal separation completes
• Callback content varies based on the type parameter: separate_vocal returns vocals and accompaniment, split_stem returns multiple instrument tracks
• Your callback endpoint should accept POST requests with JSON payload containing separated audio file links
• For detailed callback format and implementation guide, see Vocal Separation Callbacks
• Alternatively, use the Get Vocal Separation Details endpoint to poll task status
• To ensure callback security, see Webhook Verification Guide for signature verification implementation

Example:

"https://api.example.com/callback"

type · enum<string> · default: separate_vocal

Separation type with the following options:

• separate_vocal: Separate vocals and accompaniment, generating vocal and instrumental tracks
• split_stem: Separate various instrument sounds, generating vocals, backing vocals, drums, bass, guitar, keyboard, strings, brass, woodwinds, percussion, synthesizer, effects, and other tracks

Available options: separate_vocal, split_stem

Example:

"separate_vocal"

Response

200

application/json

Request successful

code · enum<integer>

Response status code

• 200: Success - Request has been processed successfully
• 400: Format Error - The parameter is not in a valid JSON format
• 401: Unauthorized - Authentication credentials are missing or invalid
• 402: Insufficient Credits - Account does not have enough credits to perform the operation
• 404: Not Found - The requested resource or endpoint does not exist
• 409: Conflict - WAV record already exists
• 422: Validation Error - The request parameters failed validation checks
• 429: Rate Limited - Request limit has been exceeded for this resource
• 455: Service Unavailable - System is currently undergoing maintenance
• 500: Server Error - An unexpected error occurred while processing the request

Available options: 200, 400, 401, 402, 404, 409, 422, 429, 455, 500

msg · string

Error message when code != 200

Example:

"success"

data · object

Show child attributes