VibeDJ Hugging Face 1
VibeDJ Hugging Face with mood, playlist, songs, and links
VibeDJ Hugging Face with download options

🎧 VibeDJ: A Playlist Generator That Feels Your Mood

💡 Inspiration

Music is therapy. But choosing the right songs when we're in our feels? Emotionally exhausting. We wanted to build something that listens to our emotions and instantly responds with music that truly understands them.

VibeDJ was inspired by:

The struggle of finding the right song for every mood.
The idea that AI can be a vibe-curator, not just a fact-spewer.
A love for both sarcasm and serious jams — and the desire to turn moods into music instantly.

🎶 What It Does

VibeDJ is a mood-to-music playlist generator powered by:

Gemini (Google Generative AI) for creative playlist curation
Spotify Web API for real song and track links
Streamlit for a fast, fun user interface

You type your mood — happy, heartbroken, chaotic, gym-raging, sleepy, anything — and it responds with:

A unique playlist title and emoji
5–8 song recommendations (via Spotify)
Options to download your playlist in .json, .csv, or .txt formats

🛠️ How We Built It

We built VibeDJ using:

streamlit: frontend UI and deployment
google-generativeai: to prompt Gemini to generate human-like, themed playlists
spotipy: to search songs and retrieve Spotify URLs
pandas: for converting playlists to downloadable formats

Workflow:

The user enters their mood via a text input.
A few-shot prompt is sent to Gemini to generate a themed playlist (e.g., sad breakup → Adele, Lorde, etc.).
Each generated song is searched via Spotify API to get:

Title
Artist
Direct Spotify link
1. Users see the playlist, songs, and download them in their preferred format.

🧱 Challenges We Ran Into

Parsing Gemini’s output into usable song names required careful pattern matching.
Not every track had a Spotify preview URL — we had to skip or substitute gracefully.
Streamlit's file handling is stateless — we moved to dynamic download buttons using st.download_button.
Spotify’s redirect URI requirement added friction (though resolved with a dummy HTTPS URI).

🏆 Accomplishments That We're Proud Of

Creating vibe-responsive playlists that genuinely feel emotionally accurate
Seamless integration of multiple APIs in a lightweight app
Giving users actual playable previews in the browser with clean formatting
Making it fun and human, not just functional

📚 What We Learned

Gemini performs much better with few-shot prompting and styled tone control
Spotipy’s Client Credentials Flow is perfect for music search without login
Streamlit is incredibly fast to prototype and fun to design expressive apps
UI/UX matters — even little things like emoji and tone made the bot feel alive 🎭

Also, math is hard, but music is math. Even if we didn’t write any equations, we felt the rhythm like $y = \sin(x)$.