Closest to what you described is http://aiva.ai http://aiva.ai (commercial product). You could experiment with OpenAI Jukebox but there’s a steep learning curv

Closest to what you described is http://aiva.ai http://aiva.ai (commercial product). You could experiment with OpenAI Jukebox but there’s a steep learning curve and you would need a decent GPU. Other than that I don’t think there are good open source models to do conditional music generation. I might be wrong though.