May 8
Loading video player...
The "Playlist" is broken. I wanted a way to seed a single song and have a system intelligently mix related tracks forever—with professional crossfades and zero manual input. In this video, I walk through the full stack of my AI DJ project. I explain how I leveraged the Spotify and YouTube APIs, how I used an LLM for real-time track selection, and the technical hurdles of creating seamless transitions in a web browser. 🚀 Try the Live App (Free): https://spotify-ai-dj.onrender.com 🔧 The Tech Stack: Backend: Python / FastAPI Frontend: Vanilla JS / YouTube IFrame API Data: Spotify Web API & yt-dlp 🚧 THE ENGINEERING CHALLENGE Right now, public music APIs (Spotify/YouTube) are heavily restricted. We lose access to the Audio Analysis features (BPM, Key, Energy, Timbre) that make professional DJing possible. In this segment, I discuss how I’m using metadata-based LLM logic as a fallback and what the next steps are for "Black Box" audio analysis. 💼 Full technical breakdown on LinkedIn: https://www.linkedin.com/in/mananmaroo #AIDJ #Python #FastAPI #SystemDesign #SoftwareEngineering #SpotifyAPI #YouTubeAPI #WebDev #BuildInPublic #MachineLearning #MusicTech