audio-stem-splitter - SKILL.md Agent Skill

name: audio-stem-splitter id: OC-0105 version: 1.0.0 description: "Split audio into stems using Demucs via Replicate – separate vocals, bass, drums, other" env: - REPLICATE_API_TOKEN commands: - split - split-four-stems - list-models

Audio Stem Splitter

Separate audio tracks into individual stems (vocals, bass, drums, other) using Demucs on Replicate.

Prerequisites

A valid REPLICATE_API_TOKEN environment variable.

Commands

Command	Description
`split`	Split audio into selected stems
`split-four-stems`	Split audio into all four stems at once
`list-models`	List available stem separation models

Usage

export REPLICATE_API_TOKEN="r8_..."
python3 scripts/audio_stem_splitter.py split --input song.mp3 --stems vocals --output-dir ./stems
python3 scripts/audio_stem_splitter.py split --input song.mp3 --stems all --output-dir ./stems
python3 scripts/audio_stem_splitter.py split-four-stems --input song.mp3 --output-dir ./stems
python3 scripts/audio_stem_splitter.py list-models