🎵

Spleeter (Deezer)

Name: Spleeter (Deezer)
Author: IndiAI Tools Editorial Team

Instant stem extraction for remixing — AI Music & Audio tool

Free ⭐⭐⭐⭐⭐ 4.5/5 🎵 AI Music & Audio 🕒 Updated April 21, 2026

IA Reviewed by the IndiAI Tools editorial team How we review →

Visit Spleeter (Deezer) ↗ Official website

Quick Verdict

Spleeter (Deezer) is an open-source audio source-separation toolkit that splits tracks into 2, 4, or 5 stems via pretrained TensorFlow models. It’s ideal for producers, DJs, and researchers who need local, scriptable stem extraction without vendor lock-in. Deezer distributes model weights, a CLI, Python API and a Docker image; there are no paid plans from Deezer — the project is free under an MIT-style license.

Spleeter (Deezer) is an open-source audio source separation tool that extracts individual stems (vocals, drums, bass, accompaniment) from mixed audio. It uses pretrained TensorFlow models provided by Deezer Research to deliver reproducible separation for music production, remixing, and music information retrieval tasks. The key differentiator is a lightweight, scriptable workflow: a command-line interface, Python API and an official Docker image that enable batch processing and integration into DAW tooling. Spleeter is freely available under a permissive license, making it accessible to hobbyists and professionals in the AI Music & Audio space.

About Spleeter (Deezer)

Spleeter is an open-source source-separation toolkit published by Deezer Research in 2019. Designed as a research-to-practice release, Spleeter provides pretrained TensorFlow models and tools to split audio into discrete stems for downstream audio work. Deezer positioned Spleeter to accelerate research in music information retrieval while giving creators a practical local toolset. Because the project is available under a permissive open-source license and distributed on GitHub, users can audit models, run them locally, and incorporate the models into batch audio pipelines without vendor-hosted subscriptions.

Spleeter ships with several concrete features. It includes pretrained TensorFlow models for 2-, 4- and 5-stem separation (vocals, drums, bass, and accompaniment variants). A command-line interface allows commands like: spleeter separate -i input.mp3 -p spleeter:4stems -o output_dir to process single files or entire directories. The Python API exposes Separator objects and save_to_file methods for scripted workflows and integration into notebooks or DAWs. Deezer also publishes an official Docker image (deezer/spleeter) so teams can containerize separation on servers or CI. The tool accepts common formats (MP3, WAV) and supports batch processing via CLI or scripted loops.

Pricing-wise, Deezer provides Spleeter as a free, open-source project with no official paid tiers. There is no hosted subscription from Deezer; the software and pretrained models are available at no charge under an MIT-style license. Organizations that need managed, scalable hosting must either self-host Spleeter on their cloud infrastructure or purchase hosted separation services from third-party vendors who may wrap Spleeter in a paid product. Enterprise costs therefore depend on self-hosting infrastructure or third-party vendor pricing — Deezer itself does not sell a pro plan.

Spleeter is used by remixers, DJs, audio engineers, music technologists, and MIR researchers. Example workflows include: a DJ/Remix Producer using Spleeter to extract vocal stems for a new bootleg remix, and a Music Researcher using batch separation to generate isolated instrument corpora for training models. Podcast editors use it to remove musical beds, while archivists separate voices from field recordings for analysis. Compared to Demucs, Spleeter favors a lightweight TensorFlow-based toolchain and an MIT license, while Demucs often targets state-of-the-art perceptual quality via different model architectures.

What makes Spleeter (Deezer) different

Three capabilities that set Spleeter (Deezer) apart from its nearest competitors.

✨ Deezer Research distributes official pretrained 2/4/5-stem TensorFlow models under a permissive MIT-style license.
✨ The project includes an official Docker image 'deezer/spleeter' so teams can containerize separation reproducibly.
✨ Spleeter emphasizes scriptable CLI and Python API integration, enabling directory-level batch jobs and CI usage.

Is Spleeter (Deezer) right for you?

✅ Best for

Remix producers who need quick vocal and instrumental stems for new mixes
Music researchers who need reproducible, scriptable stem extraction at scale
Podcast editors who want to isolate or remove musical beds from episodes
Archivists needing batch separation of legacy recordings for analysis

❌ Skip it if

Skip if you require a cloud-hosted, SLA-backed separation API from Deezer.
Skip if you need highest-perceptual-quality separation; newer models like Demucs may outperform.

✅ Pros

Free, open-source distribution under a permissive license—no Deezer subscription required
Multiple pretrained models (2/4/5 stems) covering common remix and analysis needs
CLI, Python API and Docker image enable reproducible, scriptable batch workflows

❌ Cons

Separation artifacts remain on complex mixes; stems may need manual cleanup
No official Deezer-hosted paid service or SLA — users must self-host or use third parties

Spleeter (Deezer) Pricing Plans

Current tiers and what you get at each price point. Verified against the vendor's pricing page.

Plan	Price	What you get	Best for
Free	Free	Full feature set, run locally; performance depends on your hardware.	Developers, hobbyists, researchers needing local separation
Enterprise (Self-hosted)	Custom	Scale depends on your infra; no vendor SLA from Deezer.	Organizations needing on-premise legal control and scalable processing

Best Use Cases

DJ/Remix Producer using it to extract vocal stems for 20 remixes per week
Music Researcher using it to generate training corpora of isolated instruments from 10,000 tracks
Podcast Editor using it to remove background music from 50 episode files monthly

Integrations

GitHub (source code and releases) PyPI (pip install spleeter) Docker Hub (deezer/spleeter image)

How to Use Spleeter (Deezer)

1
Install via pip or Docker

Install the package: run pip install spleeter or pull the Docker image docker pull deezer/spleeter. Success looks like the spleeter command appearing in your terminal or the Docker image listed in docker images.
2
Run a simple CLI separation

Execute a separation: spleeter separate -i song.mp3 -p spleeter:4stems -o output_dir. The command creates an output_dir with per-stem WAV files; verify by checking output_dir/song/vocals.wav.
3
Use the Python API for scripting

In Python, from spleeter.separator import Separator; Separator('spleeter:2stems').separate_to_file('song.mp3','out_dir'). Confirm success by loading out_dir/song/vocals.wav in your DAW or player.
4
Batch process directories or containerize

For many files, run spleeter separate -i ./input_folder -p spleeter:5stems -o ./outputs or start the Docker container with mounted volumes. Success looks like one subfolder per input file with separated stems.

Spleeter (Deezer) vs Alternatives

Bottom line

Choose Spleeter (Deezer) over Demucs if you prefer a lightweight, MIT-licensed TensorFlow tool with simple CLI and Docker integration.

Head-to-head comparisons between Spleeter (Deezer) and top alternatives:

Compare

Spleeter (Deezer) vs Avatarify

Read comparison →

Frequently Asked Questions

How much does Spleeter (Deezer) cost?+

Spleeter is free and open-source (MIT license). It is distributed at no charge by Deezer Research and there are no official Deezer-paid plans. Costs arise only from the compute and storage you provide when self-hosting or from third-party hosted providers that wrap Spleeter in a paid service. Enterprise-grade SLAs or managed hosting will require separate vendor pricing.

Is there a free version of Spleeter (Deezer)?+

Yes — Spleeter is fully free and open-source software. Deezer publishes the code, pretrained TensorFlow models, and example usage on GitHub under a permissive license. You can pip install spleeter or pull the deezer/spleeter Docker image and run unlimited separations locally; your only operating costs are CPU/GPU, storage, and any cloud VM charges if you self-host.

How does Spleeter (Deezer) compare to Demucs?+

Spleeter emphasizes pretrained TensorFlow models and a scriptable CLI, while Demucs focuses on different architectures and often better perceptual quality. Choose Spleeter for reproducible, lightweight TensorFlow toolchains, Docker and Python API integration. Choose Demucs if you prioritize state-of-the-art perceptual separation quality—Demucs models may reduce artifacts on complex mixes but are differently licensed and heavier to run.

What is Spleeter (Deezer) best used for?+

Spleeter is best for extracting 2, 4 or 5 stems for remixing, sampling, and research datasets. It’s widely used by DJs extracting vocals for bootlegs, producers creating acapellas, and researchers generating instrument corpora. Because it’s scriptable, Spleeter is suited for batch processing thousands of tracks for analysis or preparing stems for DAW-based editing and further cleaning.

How do I get started with Spleeter (Deezer)?+

Install and run the CLI to begin: pip install spleeter or docker pull deezer/spleeter, then run spleeter separate -i song.mp3 -p spleeter:4stems -o out. Check the output folder for per-stem WAV files. If you prefer scripting, import Separator from spleeter.separator in Python and call separate_to_file to integrate into notebooks or pipelines.