Free AI vocal remover for any song
Split a track into six stems (vocals, drums, bass, guitar, piano and other) with MIDI. Hear a free 60-second preview with no account, then sign in to split the whole song. 30 free minutes every month. Paid plans from $4.99 for 300 minutes, under 2¢ a minute.
How to remove vocals from a song
- 1
Upload your song
Drop an MP3, WAV, FLAC, M4A, OGG, AAC or AIFF file, up to 10 minutes and 120 MB.
- 2
Preview the stems free
Hear the first 60 seconds of all six stems in your browser. No account needed.
- 3
Split the whole track
Sign in to separate the full song and download each stem as a 320 kbps MP3 with its MIDI, or lossless WAV on a paid plan.
Make a karaoke version, an acapella or a backing track
Karaoke tracks
Mute the vocals and sing over the original backing, not a cover version.
Acapellas
Pull the lead vocal out on its own for remixes, mashups and edits of tracks you have the rights to use.
Practice and backing tracks
Remove the vocal to rehearse a part, or keep only the vocal to learn the melody by ear.
DJ edits
Build intros, outros and vocal-free breaks from a single stereo file.
Need the tempo or key instead? Try the BPM finder or the key finder.
Six stems, MIDI and more from one upload
Everything below comes from the same upload; the API runs the same engine.
Stem separation
Drop a track and hear a free 60-second preview of all six stems (vocals, drums, bass, guitar, piano and other) right in your browser, before you spend anything. Sign in to split the full song.
MIDI for every stem
A full split writes a MIDI file for each stem, plus one file with every stem, ready for your DAW.
Key and BPM
Every full split shows the tempo, with half and double time, and the key when the model is confident enough to name one.
Lyrics
Transcribe the vocals of a full split into timed lyrics, free.
Lead and backing vocals
Split the vocals again into the lead voice and the backing vocals. The button shows the minutes it will use before you press it.
Developer API
The same engine, live and key-gated, for automating separation in your own pipeline — a synchronous call for short clips, or an async job queue for anything longer.
Get an API keyPricing
Start free. Upgrade when you need more minutes.
Free
Free
30 minutes per month
Starter
$4.99/month
300 minutes per month
Pro
$12/month
1,000 minutes per month
Scale
$49/month
5,000 minutes per month
Same minutes on the web app and the API. Full pricing and comparison
Ready to hear it split?
Drop a track above for a free, no-account preview — it only takes a few seconds to hear the difference.
Jump back to the upload boxFrequently asked questions
How do I remove vocals from a song?
Upload the song at the top of this page and listen to the free 60-second preview; mute the vocals to hear the instrumental. Sign in to split the whole track and download the stems as MP3.
Can I make a karaoke version or extract an acapella?
Yes. Mute the vocals for a karaoke version built from the original backing track, or solo them for the acapella. If you publish anything you make, use tracks you have the rights to.
Which files can I upload?
MP3, WAV, FLAC, M4A, OGG, AAC and AIFF files, up to 10 minutes and 120 MB each.
What do I get with the free preview?
A 60-second preview of all six stems you can play right in your browser — no account, no card, no MIDI. Sign in only once you want the full track split and downloaded.
What are the six stems?
Vocals, drums, bass, guitar, piano and other. Other is everything that is not one of the first five, such as strings or synths. A full split also writes a MIDI file for every stem.
Is the Free plan actually free?
Yes — 30 minutes of separation a month, one job at a time, no card required. Sign up and it works today.
Can I buy a paid plan today?
Yes. Pick Starter, Pro or Scale on the pricing page and check out through Paddle, our payment provider. The Free tier stays free if you'd rather not upgrade.
Is there a developer API?
Yes. It's live today, gated by an API key you create at /developers/keys, with a synchronous endpoint for short clips and an async job queue for longer ones.
Who built the separation models?
Separation runs BS-RoFormer-SW (bs-roformer-infer, MIT; weights enerjazzer/BS-ROFO-SW-Fixed), with HT-Demucs 6s (Meta AI Research, demucs MIT) as the fallback. MIDI comes from YourMT3+ (mimbres/YourMT3) for pitched stems and ADTOF (CC BY-NC-SA) for drums. The footer credits every model we use.