What the tool does

How it works

Audio separation is done on our own server with an AI source-separation model: you upload a file, a worker processes it, and you get the separated tracks back. There is nothing to install, and no account is needed — a signed session cookie keeps your own task list separate from other visitors. If you register later, the tasks you created as a guest are merged into your account.

Processing is serial and shared, so a busy queue can take a few minutes. Files longer than 20 minutes are rejected to keep the queue usable for everyone.

Honest limitations

Copyright

The tool is free, but the rights to the music stay with its owners. Only process audio you are allowed to use, and get permission before publishing or monetising a separation.

Contact

Questions, bug reports or takedown requests: please use the contact address listed on our site profile.

Separate for free