Audial tool

AI Stem Splitter: isolate vocals, drums, bass and more

Studio-grade stem separation in the browser, in Python, over REST, or from your AI assistant.

Audial's stem splitter takes a full mix and returns clean, time-aligned stems: vocals, drums, bass and everything else. It runs on Audial's hosted GPU engine, so a three-minute song comes back in well under a minute with no setup on your machine.

The same engine is available everywhere you work. Drag a file into the landing page, call it from the Python SDK, post a file to the REST API, or ask Claude or Cursor to split a track through the Audial MCP server. Every result lands in your dashboard and can be pulled straight into a DAW.

What you get

  • Vocals, drums, bass and other stems from one upload
  • Tempo and key are detected automatically and tagged on every stem
  • Keeps stems sample-aligned so they drop back onto one grid
  • Works from the website, the Python SDK, the REST API and MCP
  • Accepts MP3, WAV, FLAC, OPUS, OGG, AAC, M4A and AIFF up to 100 MB

Hear it

Full mix
Vocals
Drums
Bass
Other

A demo mix split into four stems.

Use it from code

curl -X POST "https://api.audialmusic.ai/api/functions/run/stem-splitter" \
  -H "X-API-Key: your_api_key" -H "X-User-ID: your_user_id" \
  -F "userId=your_user_id" \
  -F "original=@/path/to/song.mp3"

Frequently asked questions

Vocals, drums, bass and an 'other' stem that holds everything else (guitars, keys, synths, pads). You can keep all four or just the ones you need.

Related