(Note: Use our website to hear the songs without YouTube compression)
Make a sick Y Combinator Power Metal Anthem:
Or lament the "Type 2 fun" we founders love so much:
Or hear from Elvis about the time he pivoted:
Or discover Ella Fitzgerald’s content policy:
Ryan graduated in December with CS and History degrees from Cornell, though was also nearing completion of an early masters program focusing on ML. He also spent a year and a half researching music AI on the side in collaboration with the HarmonAI (by Stability) and open-source researchers.
Hayden was a Junior at Cornell until December. He was previously published at NeurIPS, the author of one of the first open-source transformer summarization libraries, and support unicorn Ada's youngest ML engineer intern at eighteen.
Records, radio, digital, streaming, TikTok. The music industry is well acquainted with transformation. However, participation has remained exclusive throughout. The software and sound engineering skills necessary to produce music at the quality mainstream listeners expect remain out of the hands of millions of would-be musicians. Accordingly, humanity is missing out on countless incredible songs and musical styles that could have been.
Starting today, everyone is a professional musician. Make your first song right now at https://sonauto.ai!
Sonauto is training a music foundation model to allow you and musicians of any skill level to produce radio-quality songs with no prior knowledge. We're calling it Melodia.
Our first release is a text-to-music service reminiscent of Midjourney, along with a sharing platform. You can enter a description of the musical and lyrical qualities of the music you're imagining and receive a fully-produced song in seconds, or keep editing for something even more special. In addition, you can share, comment on, and remix the creations of others.