The world's first amphibious, autonomous AI DJ.
This is a fun project that combines 3D visuals with generative AI audio. The concept is simple but technically kind of heavy. A frog named DJ Rawg stands at a turntable in a 3D underwater scene. When people in the connected stream's chat type a prompt (like "play some 80s synthwave" or "sad jazz for a rainy day"), the system actually generates that track from scratch and seamlessy mixes it into the set.
There are no playlists and no pre-recorded MP3s. Every single sound you hear is hallucinated by the AI in real-time.
The system is split into two main parts that talk to each other over WebSockets.
1. The Brain (Backend)
2. The Body (Frontend)
If you want to run this locally, you are going to need a decent GPU. Audio generation is heavy.
python dj_server.py
pnpm install
pnpm run dev
I'm working on adding an overlay system so the stream chat messages appear as bubbles inside the 3D scene. I also want to add more dance moves for Rawg depending on the genre (headbanging for metal, swaying for ambient).
MIT License. Feel free to fork this and make your own animal DJ.