Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

Spleeter

github.com

  • Free plan
  • Linux
  • API
  • Self-hosted
7.0 Club score Visit github.com

What Spleeter is

Spleeter is Deezer’s open-source software for separating vocals and instruments from an audio file or rebuilding multitracks from a mixed recording. Its pre-trained models create two stems, four stems, or five stems: vocals and accompaniment; vocals, drums, bass, and other; or vocals, drums, bass, piano, and other. Developers with isolated-source datasets can train their own separation models. Written in Python and built with TensorFlow, Spleeter works from the command line or as a Python library. Installation is available through pip or Docker. Deezer describes uses such as creating instrumental tracks, restoring and remastering audio, making DJ mixes, and extracting voice tracks for subtitling. The library supports batch processing and WAV export, and offers API access. Spleeter Free costs nothing. Spleeter Pro is an enterprise API with contact-based pricing.

Who it suits

Spleeter suits developers and companies working with audio separation, especially those needing command-line, Python, Docker, or API access. It can also support music production, restoration, DJ, and subtitling workflows.

Good to know, and what to watch

  • Open-source MIT-licensed library
  • Supports two, four, and five stems
  • Python, command-line, pip, and Docker access
  • Batch processing with WAV export
  • API access available
  • No desktop app
  • Apple M1 TensorFlow issues are documented
  • Rights-owner authorization required for copyrighted material
  • Open-source forks are not directly supported by Deezer

Spleeter Pro

—

Spleeter Free

Free

Open source

The verdict

Spleeter provides configurable stem separation with developer-focused access and several deployment options. Its Apple M1 caveat and limited direct support matter when planning a production setup.

What the maker publishes

Deployment
The maker says Spleeter can be installed with pip or used with Docker. source
Developer access
Spleeter is written in Python, uses TensorFlow, and can be used from the command line or as a Python library. source
Enterprise offering
Spleeter Pro is described as an enterprise version available as an API, with contact required to learn more about integration. source
Integrations
The project README names Spleeter 4 Max for Ableton Live and lists professional audio products using its pre-trained models, including iZotope RX 8, Steinberg SpectraLayers 7, Acon Acoustica 7, VirtualDJ, and Algoriddim NeuralMix and djay Pro. source
Intended users
Deezer presents the free library for developers and the Pro offer for companies seeking integration into company tools. source
License and rights
The Spleeter code is MIT-licensed, and the README says users must obtain authorization from rights owners before processing copyrighted material. source
Limitations
The project README notes known Apple M1 issues related mostly to TensorFlow compatibility and provides a workaround. source
Performance
Deezer says the 4-stem separation can run 100 times faster than real time on a GPU. source
Purpose
Spleeter separates vocals and instruments from audio files and can recreate multitracks from a flat audio file. source
Security and compliance
The maker pages reviewed provide no Spleeter-specific security certification or compliance statement. source
Separation models
Pre-trained models support 2 stems (vocals and accompaniment), 4 stems (vocals, drums, bass, and other), and 5 stems (vocals, drums, bass, piano, and other). source
Support
The commercial offering advertises dedicated professional support, while Deezer says the open-source forks and websites built on Spleeter are not hosted, maintained, or directly supported by Deezer. source
Training
The library lets developers train source separation models when they have a dataset of isolated sources. source
Use cases
Deezer describes uses including generating instrumental tracks, audio restoration and remastering, DJ mixes, and extracting voice tracks to support subtitling. source

Checked against the maker's pages on 9 Oct 2026.