Lossless Vocal Remover is a browser-based tool that splits any song into a clean acapella and a clean instrumental. We started it because existing free vocal removers are slow, paywalled, or output lossy MP3 — fine for casual karaoke, useless for music producers, DJs, and anyone working with the stems further.
How it works
Drop a song. Our AI vocal-isolation model runs on dedicated cloud GPUs and produces two stems: a vocals-only acapella and an instrumental with the vocals removed. A typical 3-minute song processes in under a minute and the files are deleted within an hour.
Free vs Pro
The free tier runs Mel-Band Roformer at standard quality and outputs in your input format (drop a FLAC, get FLAC; drop an MP3, get MP3). It is supported by ads on the result page. Pro ($4/month) runs the higher-SDR BS-Roformer model, always outputs lossless 24-bit FLAC, removes the daily limit, allows songs up to 30 minutes, prioritises your jobs, and removes the ads.
Why lossless matters (for Pro)
Most free vocal removers compress output to 128 or 320 kbps MP3 — which loses high-frequency detail and adds artefacts every re-encode. Pro outputs 24-bit FLAC because that is what professionals working further with the stems actually need: DJ sets, remixes, mash-ups, and cover recordings all start from full-quality source audio.
Privacy
Your uploads and the resulting stems live for at most one hour before being deleted automatically. We don't fingerprint songs, we don't keep audio copies, and we don't share anything with third parties. See the privacy policy for details.
Contact
General: hello@losslessvocalremover.com — Press: press@losslessvocalremover.com — Takedowns: dmca@losslessvocalremover.com.