Our AI-powered platform unlocks the undiscovered 85% of artists on streaming services, revolutionizing how consumers discover emerging artists. Our proprietary semantic embedding technology offers hyper-personalized recommendations and advanced natural language search features.

We spent months building a machine learning model that takes the audio file of a song and maps its musical and extra-musical components to a semantic embedding space. It is able to capture the emotional feel, musical components such as instruments and type of vocalist, and so many more abstract features of the song; anything someone can use to describe a song, the model captures. We then combine this with the lyrical meaning of the song to get a complete understanding of the piece. This approach requires no metadata, just the song itself. On top of this semantic space, we invented personalization techniques to allow each user to have their own unique navigation through this space such that each song has an adjusted meaning for each user.

The first version is a mobile app where users can chat with natural language to find songs. A user can search for songs specific as “sad folk-rock songs with a 90s vibe and a male vocalist that are good to run to,” as abstract as “jellyfish vibes,” or anything in between. The app will show previews of songs for each search that a user can like or dislike: automatically creating a playlist of the liked songs on whichever streaming service they use. Our next development steps will unfold as we learn how users experience and share music in our initial build.

Not only will we expand the music industry to capture these 85% of artists, but through personalization data and first mover advantage, we plan to be the only ones able to navigate this expanded space.

© Egg Intelligence Inc. 2024