Practical overview
When it fits
Use Spleeter when you need programmatic or command-line separation of a mixed audio file into one of its documented stem arrangements.
Who it is for
- Developers adding audio source separation to a Python pipeline
- Audio users who want a command-line workflow for extracting vocals or instrument groups
- Researchers or developers training separation models with datasets of isolated sources
Documented uses
- Split a mix into vocals and accompaniment.
- Extract vocals, drums, bass and other parts, or use a configuration that also separates piano.
- Use separated audio files in a subsequent production, analysis or development workflow.
Requirements & limits
The source code is public under the MIT license. The README links to a Google Colab notebook for trying Spleeter without a local installation.
- Spleeter offers three documented pretrained stem groupings, so choose based on which parts you need.
- Local setup requires ffmpeg and libsndfile in addition to installing the Python package.
- The project uses TensorFlow, and the README flags known Apple M1 compatibility issues.
- Deezer distinguishes its library from third-party interfaces and services built on it, which it does not host or directly support.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Separate an example track into vocals and accompaniment
Install Spleeter and create two separated audio files from the documented example.
Before you start
- Python environment with pip
- ffmpeg and libsndfile
- An audio file; the README provides an example MP3
Steps
Install ffmpeg and libsndfile.
Install Spleeter with pip.
Download the documented example audio file.
Run the documented
spleeter separate -p spleeter:2stems -o output audio_example.mp3command.
Expected outcome
The example produces vocals.wav and accompaniment.wav in the output/audio_example folder.
Guide limitations
- The README no longer recommends Conda for installing Spleeter, although its example shows Conda for installing ffmpeg and libsndfile.
- Known Apple M1 issues are associated mainly with TensorFlow compatibility; the README links to a workaround.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 14 October 2026
- Original website title
- deezer/spleeter · Official homepage title · retained in its source language

