I want to build an agent that lets me enter a youtube video, and get instrumental/vocal returned, along with the BPM of the track.
It should be quite a simple agent.
First, user enters a YouTube URL (variable1), and then the mp3 file is extracted.
Next we take this MP3 file and then extract the instrumental and vocal from it.
This then returns the MP3 of the instrumental and vocal for the file - and that is our final output, with no additional processing needing (although we could potentially add a final output to return the URLs of each file in JSON).
.