Stems ST-02: AI Audio Separator with Demucs v4
Stems ST-02 splits any song into vocals, drums, bass, and instrumental stems using Facebook's cutting-edge Demucs v4 model, complete with key and BPM detection.
Overview
Stems ST-02 is a streamlined audio separation tool built on top of Facebook's open-source Demucs v4 library, delivering a significant leap in quality over older solutions like Deezer's Spleeter. Rather than relying on outdated separation algorithms, Stems ST-02 taps into one of the most advanced source-separation models available today to cleanly extract vocals, drums, bass, and other instrumentation from a single audio file.
Beyond raw stem splitting, the tool also detects key and BPM automatically, saving DJs and producers the extra step of manually analyzing tracks before mixing or remixing. The entire workflow is designed to be frictionless: upload a song, let the model process it, and download individual stems in seconds. Whether you're crafting a remix, building an acapella, or dissecting a track to understand its arrangement, Stems ST-02 removes the technical barriers that once made high-quality audio separation a complex, specialist task.
Capabilities & Features
- Audio separation
- Vocal isolation
- Instrumental isolation
- Demucs
- Music production
- DJing
- Stem separation
- BPM detection
- Key detection
Core Features
- Four-way audio separation (vocals, drums, bass, other)
- Powered by Facebook's state-of-the-art Demucs v4 model
- Automatic key detection
- Automatic BPM detection
- Simple upload-and-download workflow
- Individual .wav file downloads for each stem
Use Cases
- Extracting clean acapellas for remixes or mashups
- Producing instrumental versions of tracks for karaoke or covers
- Prepping isolated tracks for DJ practice and live mixing
- Studying individual instrument parts to learn arrangement and production
- Quickly identifying a song's key and tempo before sampling or remixing
Best For
- DJs
- Music producers
- Remix artists
- Music students and learners
- Karaoke and cover creators
Pros
- •Uses Demucs v4, a notable upgrade over older models like Spleeter
- •Simple, intuitive upload-and-download interface requires no technical expertise
- •Built-in key and BPM detection saves extra analysis steps
- •Separates tracks into four distinct, useful stems
- •Outputs are standard .wav files compatible with most DAWs and audio software
Cons
- •No mention of batch processing for multiple songs at once
- •Separation quality still depends on the complexity and mix of the original recording
- •Limited to four stem categories, without finer sub-separation (e.g., individual drum elements)
- •Pricing details are not clearly disclosed, which may create uncertainty for prospective users
How to Use
1. Go to the Stems ST-02 website and upload your audio file.
2. Let the Demucs v4 model process the track, separating it into vocals, drums, bass, and other instrumental layers.
3. Review the automatically detected key and BPM information.
4. Download each isolated stem individually as a .wav file for use in your projects.
Frequently Asked Questions
Open in AI Studio
Ask our AI to evaluate if Stems ST-02 fits your specific workflow.
Pricing
Specific pricing details are not publicly listed, so prospective users will need to check the website directly for current plans or trial options.
Pricing data is provided as a summary. Visit the vendor website for full tier details.