Hi HN, I just launched FeelStr, a lightweight completely vibe-coded app that recommends movies and shows based on your emotional state—not genres, ratings, or clickbait thumbnails. You simply type in how you’re feeling, like: “Post-breakup haze” “Numb and need something gentle” “Hopeful but not cheesy” FeelStr processes the input using a custom emotional graph and NLP pipeline to match your vibe to streaming content. The goal is to make discovery emotional, not algorithmic. How it works: Uses LLMs to classify freeform mood input into nuanced emotional vectors Matches those vectors to content tagged with human-curated emotional tone (e.g., comforting, chaotic, bittersweet) Data comes from TMDb + custom overlay for tone, tempo, and theme Currently suggests movies and shows available on major streaming platforms It’s free, fast, and completely anonymous (no login needed). Why I built it: Scrolling through streaming platforms when you’re tired, overwhelmed, or just not in the mood to decide is a terrible UX. I wanted a tool that lets you lead with emotion and gets out of the way. This was also an excuse to explore vibe coding and build something intuition-first. Would love your thoughts on: UX clarity: is it obvious what to do? Mood detection accuracy What you'd want next (TV filters, mood shifting, social features?) Try it out at https://feelstr.com Appreciate your feedback!