I had Claude build a private podcast station for me. It integrated with Gemini to create a script for the show, based on a topic of my choosing, each talking segment ends with a presentation of the next song, which is played via Spotify, and is selected to have some sort of tie-in with the previous discussion. A tts model generates audio files based on the script, and a playlist is generated to play local file audio segment, then Spotify track, then the next segment etc.
An AI made a program integrating with 2 other AI, it's AI all the way down! and the result is great! I'm learning so much by having my own private radio host speaking about topics that interest me.
Wow, I am very interested to see how this works