🧠 Inspiration

The post-pandemic world has left many, especially younger generations, grappling with anxiety, burnout, and isolation. Traditional mental health support remains stigmatized, expensive, or hard to access. We envisioned a platform that uses cultural intelligence and AI to support emotional wellness in a private, engaging, and insightful way.

For the Qloo Hackathon, we adapted our mental wellness platform — originally built for another challenge — to integrate with Qloo’s Taste AI™. By combining LLMs with Qloo’s cultural graph, our solution recommends content that truly resonates with a person’s emotional state, from calming music to inspiring shows and uplifting books.


💡 What It Does

AuraCare (Qloo Edition) is a gamified, AI-powered mental wellness platform that helps users manage emotional well-being by connecting their moods with culturally aligned, mood-boosting recommendations.

Using a blend of Qloo’s Taste AI™ and LLM-driven chat, AuraCare offers:

  • 🎮 Mood-Based Assessment: Users rate their emotional state via a visual drag-and-drop interface
  • 🤖 LLM-Powered Chat Support: A ChatGPT/Gemini-based assistant offers empathetic conversation, reflections, and wellness prompts
  • 🎧 Taste-Aligned Suggestions: Qloo API recommends music, movies, books, and experiences personalized to the user’s emotional profile
  • 📊 Mood Tracker Dashboard: Users can monitor trends, behaviors, and recommended content over time
  • 🧘 Wellness Mini-Games: Reduce stress, improve focus, and build positive habits
  • 🪙 Reward Layer (Optional): Earn tokens/NFTs for daily engagement, streaks, and journaling

🛠️ How We Built It

🧠 AI + LLM

  • Gemini Pro (or ChatGPT) for empathetic mental health interaction
  • Natural language understanding for emotional tone detection
  • Context-aware chat memory to maintain session history

🎯 Qloo API Integration

  • Emotion-to-taste mapping: User input (e.g. “I’m stressed”) → “calming music” or “light comedies”
  • Semantic recommendations across multiple cultural domains (music, film, books, etc.)

💻 Frontend

  • React + TypeScript for modular structure
  • Tailwind CSS + Framer Motion for smooth UI/UX
  • Zustand for state management

📦 Backend / Other

  • Supabase for journaling & user metadata
  • NFT minting system (optional) on Base Sepolia
  • Ethers.js for blockchain interaction (optional)

🚧 Challenges We Ran Into

  • Building emotionally intelligent LLM prompts that feel human
  • Mapping vague emotional expressions to semantically relevant Qloo queries
  • Balancing simplicity, warmth, and privacy in UX
  • Ensuring cultural recommendations feel authentic and comforting
  • Harmonizing real-time chat, semantic search, and gamified wellness features

🏆 Accomplishments

  • Seamless integration of Qloo’s Taste AI™ for emotion-aware suggestions
  • Created a chatbot that understands emotional tone and responds empathetically
  • Developed a full wellness assessment UX with gamification
  • Positive tester feedback around interface, insightfulness, and personalization
  • First-of-its-kind emotion-to-culture recommender driven by both LLMs and Qloo

📚 What We Learned

  • The power of contextual, taste-driven personalization
  • How LLMs and cultural APIs can jointly craft deeper emotional intelligence
  • Designing for mental health requires extra compassion & clarity
  • How to translate emotions → behavior → taste → content
  • Gamification and self-reflection can reinforce daily mental wellness habits

🚀 What’s Next

🔄 In Progress

  • 🌐 Multi-language Support via Google Translate API
  • 📱 Mobile App with Offline Support
  • 🤝 Anonymous Communities and AI-moderated support groups
  • 📊 Deeper Analytics on mood-content interaction

📋 Future Roadmap

  • 🧑‍⚕️ Therapist Portal for professional partnerships
  • 🧬 Clinical validation with wellness experts
  • 🧾 CSR/Insurance integration for organizational wellness support
  • 🧠 Contribute anonymized data (opt-in) to mental health research
  • 🌍 Scale to support culturally adapted content for global audiences

Built With

Share this project:

Updates