MoodFlow

MoodFlow: Revolutionary AI-Powered Dynamic Content Discovery

Inspiration Traditional recommendation systems use static categories like "sad → drama movies" instead of understanding nuanced human emotions. I was frustrated finding content that matched my specific emotional state—Netflix suggested generic categories when I needed something that truly understood my psychological needs. This sparked the idea: What if AI could dynamically generate completely unique search queries based on your exact emotional state?

What it does

MoodFlow revolutionizes content discovery by having OpenAI generate actual Qloo API queries in real-time based on your mood. Instead of predefined mappings, AI creates completely unique search parameters like "emotional-transition-narratives" or "healing-through-personal-strength" that have never been used before. Every mood gets a truly personalized content discovery experience across movies, music, books, podcasts, and travel destinations. How we built it

Frontend: Streamlit with custom CSS for elegant UI

AI Engine: OpenAI GPT-4 for psychological analysis and dynamic query generation Content API: Qloo API for fetching recommendations Core Innovation: AI generates creative Qloo API parameters in real-time instead of using predefined category mappings Architecture: User Mood → OpenAI Analysis → Dynamic Query Generation → Qloo API Execution → Personalized Results

Challenges we ran into

Breaking the predefined mapping paradigm - Most systems use static categories; we needed AI to generate completely new search parameters Making AI genuinely creative - Getting AI to invent new content discovery approaches rather than using existing patterns Balancing creativity with API constraints - Qloo API has specific formats, but we wanted unlimited creativity Real-time performance - Dynamic query generation needed to be fast and reliable Emotional accuracy vs technical precision - Making AI understand nuanced emotions while generating technically correct API calls