Moises
Splits songs into stems and adds practice tools such as tempo and key change, chord detection and a click.
- Free plan or free version
- Works in the browser
- Mobile app
Stem separation takes a finished stereo song and splits it into its parts: vocals, drums, bass and everything else, and with some tools guitar, piano or synths as separate files. For a working band this turns any recording into something you can practice with, rehearse against or take apart to understand.
The everyday uses are simple. Mute the bass on a song you need to learn and play along. Make a backing track for a rehearsal when your drummer can't come. Pull the vocal out of a cover to check the harmony line by ear. Rescue a part from an old live recording of your own band where the multitrack session was lost years ago.
Quality depends more on the source than on the tool. A clean studio mix with clear space between instruments separates well. A loud, dense master or a phone recording from the back of a club leaves more bleed, so you hear a ghost of the cymbals in the vocal stem or a smear of guitar in the bass. Try two or three tools on the same difficult song before you pick one; most let you preview a short section.
What a featured listing costs ยท from $99 a month
Splits songs into stems and adds practice tools such as tempo and key change, chord detection and a click.
Free web tool that splits songs into stems and detects key and tempo, with remix and DJ tools.
Free open-source desktop app for vocal and stem separation using several AI models.
Open-source separation model from Meta research that splits music into drums, bass, vocals and other.
Deezer's open-source separation library that splits songs into 2, 4 or 5 stems.
Browser-based stem splitter for vocals, drums, bass, guitar, piano, synths and more, with desktop apps.
Stem separation built for labels, publishers and sync teams, also available to independent artists.
Desktop editor that separates audio into notes and layers you can edit, retune or replace.
There are two kinds of tool here. Hosted services and apps such as Moises, LALAL.AI and Fadr do the work in the cloud and add practice features around it. Open-source projects such as Demucs and Spleeter, and desktop apps built on models like them, run on your own computer, which suits batch jobs and recordings you don't want to upload.
Rights matter as soon as a stem leaves your practice room. Separating a commercial song gives you a copy of someone else's recording; you can learn from it, but using that stem in a remix, a sample or a video you publish needs permission from whoever owns the recording and the song. Your own recordings are yours to split and reuse however you like.
A practical workflow for a band: keep one folder of practice stems made from other people's songs, clearly labeled, and a separate folder of stems from your own recordings that you are free to reuse. Name files with the song, the part and the tool you used, so you can redo a separation later with a better model and compare the results side by side.
It changes from song to song. Run the same difficult track through two or three tools and compare the vocal and bass stems on headphones. Dense, heavily compressed mixes are hard for every tool.
Yes. Demucs, Spleeter and the Ultimate Vocal Remover desktop app run on your own computer. A machine with a recent graphics card processes songs much faster than one without.
Only with permission from whoever owns the original recording and the song. Separating the audio yourself does not change who owns it. Stems from your own recordings are yours to use.
Yes. It can pull a usable part out of a stereo rehearsal or live recording, for example to remix an old gig, fix a level, or share a single part with a new member who needs to learn it.