Vocal / Accompaniment (2 stems)
Splits into a vocal file and an accompaniment file. Enough for backing tracks, cover recordings, and sing-along practice.
Local stem separation for Mac and Windows: split a song into vocals, accompaniment, drums, bass, and more
PureSpleeter takes a finished song and separates it into individual stems. You can split out just vocals and accompaniment, get four tracks with vocals, drums, bass, and other instruments, or go further and isolate guitar and piano for six stems. The separation model installs with the app and all processing happens on your computer, so it works offline too. Queue a whole folder for processing, or drive it from the command line in your own scripts. Paid download; everything is included after purchase.
Splits into a vocal file and an accompaniment file. Enough for backing tracks, cover recordings, and sing-along practice.
One track each for vocals, drums, bass, and other instruments. Good for transcribing, remixing, or practicing along without the drums or bass.
Adds separate guitar and piano tracks on top of the 4-stem split. This is an experimental mode; some songs separate less cleanly, and the app will tell you so.
Pick a folder and it automatically finds the WAV, MP3, FLAC, M4A, AAC, and OGG files inside and queues them. Each file's status is shown separately and you can cancel mid-run. When it finishes you get counts for succeeded, failed, and skipped, plus total time.
The model ships inside the installer and works as soon as the app is installed: no downloads, no network. No audio is uploaded during processing.
On Windows it tries NVIDIA CUDA first, then DirectML, and falls back to the CPU if neither is available. The Mac version computes on the CPU by default.
After setting options in the interface, you can copy them as a command line in one click. The CLI accepts input and output directories, the stem mode, and overwrite or skip for existing files, and returns fixed exit codes so scripts can check results.
| Mode | Stems produced | Good for |
|---|---|---|
| Vocal / Accompaniment | Vocals, accompaniment | Backing tracks, covers, vocal extraction |
| Standard 4-stem | Vocals, drums, bass, other | Transcribing, remixing, practicing without drums or bass |
| Professional 6-stem (experimental) | Vocals, drums, bass, guitar, piano, other | When you need an isolated guitar or piano track |
How well it separates depends on how the song was mixed. The more stems you ask for, the more they bleed into each other, and 6-stem mode shows this most. Start with 2 or 4 stems and listen first; use 6 only when you specifically need guitar or piano.
You can run PureSpleeter in a terminal, inside batch scripts or scheduled tasks. Common arguments:
--input-dir / --output-dir: input and output directories--stems 2|4|6: stem mode--overwrite / --skip-existing: overwrite or skip existing output--log: log levelThe easiest way is to set everything in the interface and click “Copy as CLI”. The Mac version runs in a sandbox, so the CLI can only reach directories you authorized in settings; the Windows version can access any path you give it.
The separation model installs with the app; processing never goes online and audio is never uploaded. PureSpleeter has no account system and no usage statistics. It only uses the network when you open a web link yourself, such as the feedback entry or pages for other products.
No. The separation model installs with the app, processing happens on your computer, and audio is never uploaded. The network is only used when you open a web link.
Input: WAV, MP3, FLAC, M4A, AAC, OGG. Output: WAV, one file per stem. For other formats (such as APE or DSD), convert them to FLAC or WAV with PureConvert Batcher first.
A paid download, buy once. After purchase everything is included, including folder batch processing and the command line. No in-app purchases and no subscription.
6-stem is an experimental mode, and some songs' guitar and piano parts don't come out clean. Switch to 4-stem instead; guitar and piano then land in the "Other" track.
Yes. Vocals and instruments overlap in frequency, so the model can't separate them perfectly. Songs with heavy reverb or vocals mixed far back tend to leave more residue.
Yes. Pick a folder and its WAV, MP3, FLAC, M4A, AAC, and OGG files are queued automatically. You can also batch from the command line and schedule it in scripts.
PureSpleeter is designed for offline local processing and does not depend on cloud transcoding. Read ourprivacy policyfor full data-handling details.
You can reach us via email feedback@efficient-lab.com or submit issues and suggestions through GitHub Issues.
Local music and sound-effect asset manager for Mac and Windows: AI tagging, batch organizing, safe writeback
Local voice denoising and de-reverb for Mac and Windows
Local audio restoration for Mac and Windows: repair recordings with hiss, noise, and missing frequency ranges
Batch-extract a whole folder of SACD ISOs into tagged DSF files