Analysis updated 2026-07-03 · repo last pushed 2026-06-30
Isolate the guitar track from a song to learn and practice a solo.
Pull out just the vocals to use in a remix or mashup.
Mute the original vocals to create a custom backing track for singing.
Split a full song into separate instrument tracks for studying the arrangement.
| stemdeckapp/stemdeck | aattaran/deepclaude | cloudflare/security-audit-skill | |
|---|---|---|---|
| Stars | 1,836 | 2,180 | 2,252 |
| Language | JavaScript | JavaScript | JavaScript |
| Last pushed | 2026-06-30 | 2026-05-16 | 2026-07-03 |
| Maintenance | Active | Maintained | Active |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 2/5 | 3/5 | 2/5 |
| Audience | general | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
On first launch the app downloads background AI tools it needs to run, but after that it works completely offline with no internet required.
Stemdeck is a free desktop app that splits any song into its individual parts, vocals, drums, bass, guitar, piano, and everything else. Instead of paying a monthly subscription to a cloud service and uploading your music to someone else's server, you just drag an audio file into the app and it gives you isolated tracks you can remix, study, or practice along with. You can also paste a YouTube URL to pull audio directly. Under the hood, the app uses a machine learning model called Demucs (built by Meta AI) to identify and separate the different instruments in a recording. Once the separation is finished, the app opens a mixing interface that looks like the kind you'd find in professional music production software. You can mute specific instruments, solo just the vocals, loop a specific section, and even export a custom mix of only the stems you selected. It also analyzes the song to tell you its tempo (BPM), musical key, and loudness. This is built for musicians, producers, and hobbyists who want to pull songs apart for practice or creative projects. A guitar player could isolate the guitar track to learn a solo, a producer could pull out just the vocals to use in a remix, or a singer could mute the original vocals to create a backing track. The big advantage is privacy, since the processing happens entirely on your own computer, your audio files never leave your machine. The project is very upfront about what it is and isn't. It positions itself as a free, local alternative to paid services like Moises or LALAL.AI. The tradeoff is that it doesn't have the polished mobile apps, cloud processing speed, or extra features (like chord detection or pitch shifting) that those paid services offer. It processes one song at a time, and the speed depends entirely on your computer's hardware, it's fast if you have a dedicated graphics card, but slower if you're running on a standard processor. What's notable is that the project refuses to accept any money, sponsorship, or funding. It's purely a free, open-source tool built for the community. The first time you launch it, the app downloads the necessary background tools it needs to run, but after that, it works completely offline.
Free desktop app that splits any song into separate instrument tracks (vocals, drums, bass, etc.) using AI. Process music entirely on your own computer for remixing, practice, or creative projects.
Mainly JavaScript. The stack also includes JavaScript, Demucs, Meta AI.
Active — commit in last 30 days (last push 2026-06-30).
The explanation does not specify a license, but the project is described as free, open-source, and refuses any money or sponsorship.
Setup difficulty is rated moderate, with roughly 5min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Verify against the repo before relying on details.