NeuralSound
NeuralSound is an AI music source separation platform. You give it a song or a recording, and it separates the audio into parts so you can keep the one you need.
The vocal remover takes the vocals out of a song and leaves the music, which is handy for karaoke and practice:
https://neuralsound.org/vocal-remover
The instrumental remover does the opposite and leaves the vocals on their own:
https://neuralsound.org/instrumental-remover
The background music remover takes the music out from behind a recording:
https://neuralsound.org/background-music-remover
NeuralSound also has a vocal isolator and stem separation for working with each part of a track separately.