Voice Remover
Separate any song into vocals, drums, bass and a clean instrumental — powered by AI. Drop a track, get your stems in a few minutes.
AI separation runs on our servers. Your source file is discarded after processing; the vocals and instrumental are saved to your private My space.
How it works
- Sign in and drop your track. Voice Remover is part of AudioKit Premium — load an MP3, WAV, FLAC, M4A or OGG up to 60 MB.
- Click “Separate into stems”. Your file is sent securely to our servers, where our AI models (Mel-Band RoFormer for the vocals, Demucs for the rest) split it into four stems — vocals, drums, bass and other instruments — plus a ready-made instrumental (vocals removed). Usually in a few minutes.
- Listen, remix, download — or record over it. Switch stems in one player, open the Mix mode (volume sliders, karaoke and a cappella presets, custom WAV mix rendered in your browser), or send the instrumental straight to Vocal Studio and record your own vocals over it.
Features
- AI separation. Vocals by Mel-Band RoFormer, drums, bass and other instruments by Demucs v4 (htdemucs), two widely used open-source source-separation models.
- Sample-accurate WAV option. Download the stems as 24-bit WAV, realigned in your browser on your original file — the instrumental is then exactly your track minus the vocals.
- Four stems plus an instrumental. Vocals, drums, bass and other instruments, plus a full instrumental, ready for remixing, sampling, karaoke or practice.
- Secure processing. Encrypted upload, processing on our servers, source file discarded after the job — stems archived in your private “My space”.
- Smart upload. Large lossless files are compressed to MP3 320 kbps in your browser before upload — faster and more reliable.
FAQ
Is Voice Remover free?
No — it's part of AudioKit Premium (€9.90/month or €99/year), because the AI separation runs on dedicated servers. AudioKit Premium includes 100 AI credits per month, shared across all AI tools. Each separation costs 3 credits. Need more? Credit packs are available from your space (100 credits for €5.99, 250 for €11.99) — pack credits never expire. AudioKit's browser-based tools remain free.
What happens to my audio file?
Your file is sent to our servers over an encrypted connection, processed by the separation model, then discarded — it isn't kept. The resulting stems are saved to your private My space library so you can find them again later. Large lossless files are first compressed to MP3 320 kbps in your browser so the upload stays fast and reliable.
Which formats are supported?
MP3, WAV, FLAC, M4A and OGG, up to 60 MB per file. All stems come back ready to play in the browser and to download. The download links expire after a few hours, but vocals and instrumental stay available in My space.
How good is the separation quality?
AudioKit uses Mel-Band RoFormer for the vocals and Demucs for drums, bass and other instruments, two widely used open-source separation models. On most modern mixes you get a clean a cappella and instrumental. Dense arrangements, heavy reverb or vocals buried in the mix can leave slight artifacts — every separation model still struggles with these cases.
Can I use the instrumental for a remix or karaoke?
Technically yes — that's the main use case, along with isolating vocals for sampling or practice. Legally, separating a track doesn't grant you rights to it: for private practice or karaoke at home you're fine, but publishing a remix of copyrighted material still requires permission from the rights holders.
Can I make a karaoke or custom version?
Yes — after separation, the Mix mode gives you a volume slider per stem and one-tap presets (karaoke, guide vocals at 20%, no drums, a cappella). Your custom mix is rendered locally in your browser and downloaded as a WAV — it uses no extra credits, since the separation has already been paid. And one click sends the instrumental to Vocal Studio, where you can record your own vocals over it.