Home / Tools / Spleeter

Spleeter

Deezer stem splitter

Spleeter preview
github.com Open Source

Upvotes

0

Pricing

Open Source

Category

Audio Editing

In the database since

Oct 2026

Overview

About Spleeter

Spleeter is Deezer’s open-source music separation library. It splits a song into separate stems using deep learning. You can pull vocals out of a track or isolate drums, bass, and piano. The library ships with pretrained models, so you do not need to train anything. It runs from the command line or as a Python library. On a GPU it separates a track up to 100 times faster than real time. Developers use it in DJ apps, remixing tools, and research projects. The code is MIT-licensed and free to use.

Key Features

Three Pretrained Models

Choose 2-stem, 4-stem, or 5-stem separation. Vocals split from accompaniment, or drums, bass, and piano get their own tracks.

GPU-Speed Separation

On a graphics card it processes 4-stem audio up to 100 times faster than real time. A full song finishes in seconds.

Command Line Tool

Run separation with one command after installing. You point it at an audio file and get clean WAV stems back.

Python Library

Import Spleeter into your own code and build music apps. DJs, researchers, and developers all use this path.

Docker and Colab Support

Run it in a Docker container or try it in Google Colab. You can test separation without installing anything locally.

MIT Open Source

The code is MIT-licensed, so you can use it in free and paid projects. Deezer also sells a separate Spleeter Pro version.

FAQ

What does Spleeter separate?

Spleeter splits songs into stems. The 2-stem model separates vocals from accompaniment. The 4-stem model adds drums and bass. The 5-stem model adds piano too.

Do I need a GPU to run Spleeter?

No. Spleeter runs on a normal CPU, just slower. A GPU speeds it up to 100 times faster than real time for 4-stem separation.

How do I install Spleeter?

Install it with pip after setting up ffmpeg and libsndfile. You can also run it through Docker or try the ready-made Google Colab notebook.

Can I use Spleeter in a commercial app?

Yes. The code is MIT-licensed, so commercial use is allowed. For dedicated support and faster processing, Deezer offers a separate paid Spleeter Pro version.

Which apps use Spleeter?

iZotope uses it in RX 8’s Music Rebalance feature. Steinberg uses it in SpectralLayers. VirtualDJ and Algoriddim’s djay apps also build on it.

Conclusion

Spleeter gives you fast, free stem separation from Deezer’s research lab. Pretrained models, GPU speed, and an MIT license make it a solid base for music apps and research. It works from the command line or as a Python library, and tools like iZotope RX already build on it. Try the Colab demo, then install it locally for your tracks.

 

Newsletter

Tools like this change fast

New Audio Editing tools launch every week — the newsletter keeps you ahead of what's worth trying.

Keep exploring

Similar AI tools

Ranked by the job they do, their capabilities, and how closely they fit this workflow.