Inspiration
We were inspired by the vast genres of Taylor Swift's discography and wanted to find a way to allow people to listen to a specific song based on a blurb or word of how they're feeling.
What it does
The text box takes any amount of characters expressing their current emotions/thoughts and returns a song by Taylor Swift that would best complement their mood along with a description of why it's a good match as well as a Spotify link to the song.
How we built it
We used Streamlit to build a Taylor Swift-themed website that displays a text box that allows for user input. We used Google Gemini APIs to create responses to prompts that asked to generate the song name and description. We used Spotify API to provide the link to the song that Gemini outputted.
Challenges we ran into
We ran into a lot of issues including generating the right song link considering there were many APIs/different songs, fine-tuning the prompt to get the desired response, and getting the site itself to deploy.
Accomplishments that we're proud of
We were able to effectively utilize Gemini and Spotify APIs and prompt engineering to develop accurate results and also provide the song link. We were also able to develop a website perfect for any hardcore Swiftie or just a casual listener.
What we learned
We learned how to incorporate Gemini API into our code, how to use prompt engineering, use Spotify API to search for song urls, how to build a website using Streamlit, and how to work together as a team.
What's next for Taylored Tunes
We can expand it to all artists, and even consider the whole Spotify song database. We can also change the colors of the website based on the album of the song that the website outputs, and generate images based on the artist of the song.
Log in or sign up for Devpost to join the conversation.