Which method should you use?
If you have a stereo MP3 or WAV and want a result in the browser, start with the vocal remover. If you already work in Audacity and the lead vocal is dry and centered, center cancellation is free but often thin. If you mix for a living, a DAW or dedicated stem tool gives more control—and usually costs money. Removing the vocal and isolating it are the two files from that same split: keep the bed, or keep the singer. The choice is spelled out in isolate vs remove vocals.
| Method | What you need | Best for | Watch out for |
|---|---|---|---|
| Online AI tool | A short stereo file | Karaoke, practice, a first remix sketch | Bleed, 2-minute free cap, files expire in about an hour |
| Audacity | The free editor | Older mixes with a centered, dry lead | Also removes kick, snare, and bass in the center |
| Professional software | A DAW or stem plugin | Long files, batch jobs, a final mix | Install, learning curve, often a paid license |
1. Online AI tool
An online AI vocal remover estimates the singer and the backing track from one finished song. You do not need the original session. The same upload also feeds the stem splitter and the vocal isolator—two aligned files: instrumental and vocals.
When to use it. One-off karaoke, rehearsal, a quick sample audition, or checking whether a chorus is even worth editing in a DAW. No account, no installer.
- Export a clean sectionMP3, WAV, M4A, or AAC, up to 50 MB. Free jobs are limited to 2 minutes, so a verse or chorus is enough to judge quality.
- Upload and splitUse the vocal remover on this site. The model (Demucs htdemucs) writes an instrumental and a vocal stem on the same timeline.
- Preview both filesListen to choruses, consonants, cymbals, and reverb tails before you download. Keep the original mix as a reference.
- Download what you needFiles are scheduled to delete after about one hour. The first successful split today includes both downloads. A later free split can unlock both files with Credits.
2. Audacity (center cancellation)
Audacity does not detect a singer. It subtracts one stereo channel from the other. Centered sounds cancel; left/right differences remain. That can knock out a dry, mono lead vocal—and the kick, snare, and bass that share the center. Full walkthrough: how to remove vocals in Audacity.
When to use it. You already have Audacity, the vocal is obviously centered, and you can accept a thinner instrumental. Skip it on wide, wet, or heavily doubled vocals.
- Import a copyFile → Import → Audio. Do not overwrite your source file.
- Split stereo to monoOpen the track menu and choose Split Stereo to Mono so left and right are separate tracks.
- Invert one channelSelect only one track and apply Effect → Invert. Do not invert both.
- Play both tracks togetherEqual gain, both panned center. Shared center content cancels.
- Export if it is usableIf the vocal remains or the mix loses all low end, stop and use the online AI method instead.
3. Professional software
DAWs and dedicated stem tools (iZotope RX, SpectraLayers, local Demucs, paid online batch apps) let you process longer files, save presets, and export WAV. GarageBand can import an AI instrumental; that workflow is in the GarageBand guide. This path is “professional” because of control and format—not because it magically recovers the original studio tracks.
When to use it. Repeat jobs, full-length songs, a client mix, or when you already live in a DAW. Expect install time and, for the better processors, a license.
- Choose the tool you already ownA stem plugin inside Logic, Ableton, Reaper, or a standalone unmix app. Avoid installing five demos for one chorus.
- Import a high-quality authorized fileWAV or high-bitrate source beats a ripped YouTube MP3.
- Run vocal / instrumental (or multi-stem) separationPreview at realistic volume. Solo drums and bass; they often reveal artifacts first.
- Export and finish in the DAWTrim, fade, EQ. If you only needed a backing track, the online vocal remover is usually faster to try first.
Try the online method
Same upload box as the rest of the site. Drop a short song you have permission to process, then preview the instrumental and the vocals.