Spice

User authentication site with PropelAuth

Inspiration

We were inspired by a few team members' personal desires to diversify our music tastes. Art and music can be a great way to hear about different people's life experiences or unique perspectives, and discover aspects of other cultures from around the world that you might not otherwise be familiar with.

What it does

Spice is a music recommendation webapp that evaluates your music listening history and suggests diverse artists (POC, LGBTQ+, international, etc.) in the genres you already know and love.

How we built it

We used the free last.fm developer API to get data on a user's all-time top artists. Then, we get a response using Fabric, OLlama running Llama 3.1, and a custom pattern that we generated. This response includes a list of diverse artists and bands which are similar to the top artists/genres that a user listens to, as well as descriptions of these artists. Finally, we put together a front-end for the webapp which ties together these details in a way that cleanly presents this output to the user.

Challenges we ran into

Parsing the JSON data as part of the backend took a lot of troubleshooting. It also took a lot of trial and error to get our AI components returning the data that we wanted.

What we learned

We learned a lot about working with APIs and LLM models, which are skills that we can carry over onto other projects and further develop throughout our careers.