Vocal & Instrumental Separation API

Separate vocals and instrumental from any song programmatically — the same 3-model studio-grade engine (SDR 12.97 dB) behind the web app, over a simple REST API. A pay-as-you-go alternative to the LALAL.AI, Moises and Spleeter APIs: no monthly fee, no SDK, just one HTTP call.

Quickstart

Three steps:

1. Get a key — sign in on your Account page and click Generate API key.
2. Add credits — the API is pay-per-call; top up credits on the same page ($0.50 per song).
3. Call the API — submit a song URL, then poll until it's done and download the stems.

Authentication

Send your key in the X-API-Key header on every request. Keep it secret — treat it like a password. You can rotate or revoke it any time on your Account page.

X-API-Key: vsk_your_api_key_here

Base URL

https://aivoiceseparator.com

POST/api/v1/separate

Submit a song URL (YouTube, SoundCloud, TikTok, Bandcamp, Vimeo, or a direct audio link). Runs asynchronously. Charges your credit up front — 1 credit per song, or more for clips over 7 min — and refunds it if the job fails.

Request body (JSON)

# Submit a song
curl -X POST https://aivoiceseparator.com/api/v1/separate \
  -H "X-API-Key: vsk_your_api_key_here" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://youtu.be/dQw4w9WgXcQ", "output_format": "mp3"}'

Response 200

{
  "job_id": "a1b2c3d4e5f6...",
  "status_url": "https://aivoiceseparator.com/api/v1/jobs/a1b2c3d4e5f6...",
  "cost_credits": 1,
  "credit_balance": 950
}

GET/api/v1/jobs/{job_id}

Poll the job. state moves queued → downloading → processing → done (a 5-min song is typically ready in ~1 min). When done, stems holds public download URLs, valid until the job auto-deletes at 24 h.

curl https://aivoiceseparator.com/api/v1/jobs/a1b2c3d4e5f6... \
  -H "X-API-Key: vsk_your_api_key_here"

Response 200 (done)

{
  "job_id": "a1b2c3d4e5f6...",
  "state": "done",
  "progress": 100,
  "output_format": "mp3",
  "stems": [
    { "name": "vocals",       "url": "https://aivoiceseparator.com/api/download/a1b2.../vocals" },
    { "name": "instrumental", "url": "https://aivoiceseparator.com/api/download/a1b2.../instrumental" }
  ]
}

End-to-end example

#!/usr/bin/env bash
KEY="vsk_your_api_key_here"
BASE="https://aivoiceseparator.com"

# 1) submit
JOB=$(curl -s -X POST $BASE/api/v1/separate \
  -H "X-API-Key: $KEY" -H "Content-Type: application/json" \
  -d '{"url":"https://youtu.be/dQw4w9WgXcQ"}' | jq -r .job_id)

# 2) poll until done
while true; do
  STATE=$(curl -s $BASE/api/v1/jobs/$JOB -H "X-API-Key: $KEY" | jq -r .state)
  echo "state: $STATE"; [ "$STATE" = "done" ] && break
  [ "$STATE" = "error" ] && exit 1; sleep 4
done

# 3) download the instrumental
URL=$(curl -s $BASE/api/v1/jobs/$JOB -H "X-API-Key: $KEY" \
  | jq -r '.stems[] | select(.name=="instrumental") | .url')
curl -sL "$URL" -o instrumental.mp3

Pricing & limits

• $0.50 per song (1 credit). Clips over 7 min cost more: 10 min = 2, 16 min = 3 credits. Max 20 min/clip.
• Buy credits on your Account page (card or PromptPay). Credits never expire; failed jobs are refunded automatically.
• Rate limit: 30 requests per 10 minutes per key.

Errors

FAQ

Is there an API to separate vocals from a song?

Yes — this is a REST API that separates the vocals and instrumental from any song URL programmatically. You POST a link, poll the job, and download two stems. It runs the same studio-grade 3-model engine (SDR 12.97 dB) as the web app, so results match what you hear on the site.

How much does the vocal separation API cost?

Pay-as-you-go: $0.50 per song (one credit), no monthly fee. Clips over 7 min cost proportionally more (10 min = 2, 16 min = 3 credits). Credits never expire, and any job that fails is refunded automatically.

Is this a LALAL.AI, Moises or Spleeter API alternative?

Yes. It's a pay-as-you-go alternative to the LALAL.AI, Moises and Spleeter APIs — no subscription, no SDK, a single HTTP endpoint, and studio-grade quality. Good for karaoke apps, remix/DJ tools, sample-flipping, transcription pipelines, and resellers who separate stems in bulk.

What sources and formats are supported?

Submit a YouTube, SoundCloud, TikTok, Bandcamp or Vimeo link, or a direct audio URL. Output is mp3 (default), wav or flac. Optional lyrics transcription (Whisper) can be turned on per request.

How do I get an API key?

Sign in on your Account page, click Generate API key, and add credits. Send the key in the X-API-Key header on every request.