Proposing a novel approach to music recommendation that takes the audio of the user as input, converts it to text and performs sentiment analysis, tf-idf, and normalization. We then use content-based and collaborative filtering techniques to recommend songs