video1.0<iframe src="https://www.loom.com/embed/cbd8741834c549889f8ced4cc17693f4" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/cbd8741834c549889f8ced4cc17693f4-9a50da1778906e64.gif429.106Hi, I am Maxim Yeliseyev, the sole developer behind VibeBuddy AI, an AI music recommender. Instead of dropdowns or sliders, it uses natural language to extract your preferences via RAG over ChromaDB with 1,710 real Spotify songs across 114 genres stored as 8 dimensional feature vectors. It runs an 8 step agentic pipeline with guardrails and self critique, and outputs your top 5 tracks with scores and natural language explanations. In the demos, I tested chill folk, heavy metal, and a sad slow request to show guardrails and reflections. No direct viewer action was requested.