Is the vocal remover really free?
Yes. The free tier lets you separate one song per day at standard quality, up to 8 minutes long. You upload, you download both stems, you keep them. No watermark, no signup, no email required. We pay for the GPU compute through ads on the result page.
What does the Pro tier add?
Pro is $4 / month or $29 / year. You get unlimited daily separations, songs up to 30 minutes long, lossless 24-bit FLAC output via the BS-Roformer model (the highest-SDR vocal-isolation model publicly available), priority queue, and no ads.
How do I remove vocals from a song?
Drop your audio file onto the upload area, wait a few seconds, and download the two stems: a vocals-only acapella and an instrumental with the vocals removed. No signup, no software to install.
What file formats are supported?
Input: MP3, WAV, FLAC, M4A, AAC, OGG, Opus. Output: same format you uploaded — drop a FLAC, get FLAC stems; drop an MP3, get MP3 stems. The Pro tier always outputs lossless 24-bit FLAC regardless of input.
What are the file size and length limits?
Maximum file size is 100 MB. Free tier songs are limited to 8 minutes; Pro accepts up to 30 minutes per song. Songs longer than these limits will be rejected up-front before upload.
Why is the free tier 'standard' and Pro 'lossless'?
The free tier uses Mel-Band Roformer, a fast model with very good quality (~11.4 dB SDR). Pro uses BS-Roformer Viperx-1297, which has the highest separation quality publicly available (~12.9 dB SDR) and is roughly 6× more compute-intensive. Pro covers that compute cost.
Can I extract an acapella from any song?
Yes — the vocal stem is an isolated acapella. Use it for remixes, mash-ups, transcription, vocal practice, or any other purpose the audio's licence allows. We don't add any restrictions, but we also can't grant rights we don't hold.
Can I make a karaoke or instrumental backing track?
Yes. The instrumental stem is the song with vocals removed — ideal for karaoke, sing-along, or recording your own cover version.
Is uploading copyrighted music legal?
You are responsible for your uploads. Use the tool only on audio you own, audio you have permission to process, audio in the public domain, or audio under a permissive licence (e.g. Creative Commons). Do not upload audio you don't have rights to. We process the file, return the stems, and delete everything within one hour — see our DMCA policy for takedown procedure.
Are my uploads private?
Yes. Uploads and stems live for at most one hour and are then automatically deleted. We don't fingerprint songs, don't keep copies, and don't share anything with third parties. Download links are single-use signed URLs that expire in one hour.
How long does it take?
A typical 4-minute song processes in 30–60 seconds at standard quality (free tier) or 45–90 seconds at lossless (Pro). The first job after a quiet period adds ~30 seconds for cold-start. Pro users get priority routing to keep this short.
Why isn't the vocal track 100 % clean?
AI vocal isolation works extremely well on cleanly-produced music but isn't perfect — heavy reverb, vocal distortion, harmonized backing vocals overlapping with the lead, and dense bus compression all bleed slightly into the wrong stem. Modern pop, hip-hop, and electronic music typically produce the cleanest separation.
Why did my upload fail?
Common reasons: the file is over 100 MB, the song is over 8 minutes (free tier), the format isn't in our supported list, or you've hit the daily limit. Pro removes the daily limit and raises the song length to 30 minutes.
How do I cancel my subscription?
Sign in, click your account menu, then Manage Subscription. You'll land on Stripe's hosted billing portal where you can cancel any time. Cancellations take effect at the end of your current billing period — you keep Pro access until the period ends.
Do you offer an API?
Not yet. If you need batch processing or programmatic access, email hello@losslessvocalremover.com with your use case — we're considering an API tier for the second half of 2026.