Drum separation
for your product.

Your API account

Quick start

# Upload and separate
# algorithm: auto | natural | puncher (puncher = paid only)
# input_type: full_mix | drum_stem (Direct mode for already-isolated drums)
# quality:    standard | studio (studio = paid only)
curl -X POST https://drumsplit.io/api/v1/separate \
  -H "X-API-Key: ds_live_your_key" \
  -F "file=@track.mp3" \
  -F "output_mode=full" \
  -F "algorithm=auto" \
  -F "input_type=full_mix" \
  -F "quality=standard"

# Poll status
curl https://drumsplit.io/api/v1/status/{job_id} \
  -H "X-API-Key: ds_live_your_key"

# Download stems
curl -O https://drumsplit.io/api/v1/download/{job_id} \
  -H "X-API-Key: ds_live_your_key"

# pip install httpx
import httpx, time

API = "https://drumsplit.io/api/v1"
KEY = "ds_live_your_key"
headers = {"X-API-Key": KEY}

# Upload — see /openapi.yaml for all options
with open("track.mp3", "rb") as f:
    r = httpx.post(f"{API}/separate",
        headers=headers,
        files={"file": f},
        data={
            "output_mode": "full",
            "algorithm": "auto",      # auto | natural | puncher
            "input_type": "full_mix",  # full_mix | drum_stem (Direct mode)
            "quality": "standard",     # standard | studio
        })
job_id = r.json()["job_id"]

# Poll until complete
while True:
    s = httpx.get(f"{API}/status/{job_id}",
        headers=headers).json()
    if s["status"] == "complete": break
    time.sleep(2)

# Download stems ZIP
r = httpx.get(f"{API}/download/{job_id}",
    headers=headers)
open("stems.zip", "wb").write(r.content)

const API = "https://drumsplit.io/api/v1";
const KEY = "ds_live_your_key";

// Upload — see /openapi.yaml for all options
const form = new FormData();
form.append("file", audioFile);
form.append("output_mode", "full");
form.append("algorithm", "auto");      // auto | natural | puncher
form.append("input_type", "full_mix"); // full_mix | drum_stem (Direct mode)
form.append("quality", "standard");    // standard | studio

const { job_id } = await fetch(
  `${API}/separate`,
  { method: "POST",
    headers: { "X-API-Key": KEY },
    body: form }
).then(r => r.json());

// Poll
let status;
do {
  await new Promise(r => setTimeout(r, 2000));
  status = await fetch(
    `${API}/status/${job_id}`,
    { headers: { "X-API-Key": KEY } }
  ).then(r => r.json());
} while (status.status !== "complete");

// Download
const zip = await fetch(
  `${API}/download/${job_id}`,
  { headers: { "X-API-Key": KEY } }
).then(r => r.blob());

# Install the MCP server
pip install drumsplit-mcp

# Add to Claude Code
claude mcp add drumsplit -- python -m drumsplit_mcp

# Add to Gemini CLI
gemini mcp add drumsplit -- python -m drumsplit_mcp

# Set your API key
export DRUMSPLIT_API_KEY=ds_live_your_key

# Now any AI agent can just say:
# "Separate the drums from track.mp3"
# "Extract kick and snare from this song"
# "How many DrumSplit splits do I have?"

Pricing

Full Mix splits include 5 drum stems + drum bus + music-minus-drums track. Direct mode accepts an already-isolated drum stem and skips the music stem. Splits never expire. No subscription. Rate limits: Standard 30/min · Volume 120/min.

Endpoints

Auth: X-API-Key header on all endpoints except /health and /api/keys. Full spec at /openapi.yaml