True two-stem output
You get the vocal and the instrumental as separate downloads — no need to run the job twice or settle for whichever side the tool favors.
one song in, two stems out
Separate any mixed track into an acapella and an instrumental you can actually use — preview both stems before you commit to a download.
The model identifies vocal content across the frequency spectrum rather than just carving out the middle of the stereo image, which is why center-channel tricks fail on modern mixes and this doesn't. Reverb tails and doubled harmonies are the hard cases — expect the cleanest results where the production kept the voice distinct.
Both stems come back at full length and full quality. The acapella suits remixes, covers, and study; the instrumental becomes a karaoke track, a rehearsal bed, or the base layer for a new topline.

Every vocal remover promises isolation; the difference is what's left behind. This one returns two files from a single upload — the voice on its own and the instrumental with the hole closed up — and lets you audition both in the browser before anything is downloaded.
Source quality decides most of the outcome. A lossless or high-bitrate file with the vocal sitting clearly in the mix separates cleanly; heavy master-bus compression and thick reverb are what leave ghosts. Feed it the best copy of the song you have and judge the stems in the preview player, not on faith.
You get the vocal and the instrumental as separate downloads — no need to run the job twice or settle for whichever side the tool favors.
Solo either stem in the browser and listen for bleed at the moments that matter: chorus entries, ad-libs, the last reverb tail.
Processing runs in the cloud from any browser. Upload, separate, audition, download — a coffee-length round trip for a full song.

The tool is simple by design; the craft is in what you feed it and how you check the result.
Use the highest-quality file you have — lossless if possible. Re-encoded, low-bitrate copies smear exactly the detail separation depends on.
Separation runs automatically; there are no knobs to misconfigure. A typical track processes in well under the time it takes to play.
Solo the acapella at the busiest chorus and the instrumental at the quietest verse. Those two spots reveal any bleed or artifacts fastest.
Strip the original vocal to rehearse over the real arrangement, or solo it to study phrasing, breaths, and runs up close.
Client supplied a song with vocals where narration needs to sit? The instrumental stem drops in without re-licensing a new track.
Pull the acapella, re-harmonize underneath it, or drop the instrumental into a set — the raw material for flips and edits.
A voice drenched in hall reverb leaves traces in both stems — that's physics, not a bug. Drier mixes separate cleaner; plan around it.
Aggressive separation can soften snare hits that shared frequencies with the voice. Listen for dulled drums before building on the stem.
A/B each stem against the source at the same volume. If the instrumental feels hollow in the choruses, the mix may need the vocal's energy replaced, not just removed.
On most modern productions, remarkably close. Where the voice and instruments were printed with shared reverb or heavy bus compression, some texture change is unavoidable — the preview exists so you can judge before downloading.
Yes, though results track the production style. Pop, hip-hop, and electronic mixes with upfront vocals separate cleanest; dense rock walls and live recordings are harder because the voice is physically entangled with the room.
Two. The isolated vocal and the full instrumental are both produced from a single upload, and you can download either or both.
Separation for private practice and study is generally uncontroversial; publishing or monetizing stems from someone else's recording needs permission from the rights holders. For material you own or made here, you're free to use the stems as you like.
Upload the song, solo both stems, and hear exactly what you're getting before you download a thing.
Remove the vocals