About Health Nest
Health Nest is a personalized AI-powered healthcare assistant I built to provide instant medical insights, symptom analysis, and nearby provider suggestions — all from one interface.
Inspiration
I was inspired to build Health Nest after seeing how difficult it can be to get quick, reliable health advice or locate a relevant doctor when you’re unwell. I wanted to create something that feels like a trusted digital health companion.
How I Built It
The project combines several modern tools and AI integrations: • Frontend: React + Tailwind CSS + Framer Motion
• Backend: Python Flask
' • LLM: OpenAI GPT-3.5 Turbo
• Vector Store: FAISS (used for RAG-based document retrieval)
• Data Sources: WHO, Mayo Clinic, NIH — curated PDFs, CSVs, and JSONL files
• API: Google Maps Places API (used to locate doctors nearby based on illness and ZIP)
Core Features • 💬 AI Chat Assistant – Ask questions and get reliable answers from an AI backed by real medical data
• 📍 Nearby Doctor Lookup – Type a symptom + ZIP and get the right specialist near you
• 📚 Document-grounded RAG – AI pulls answers from verified healthcare PDFs using FAISS
• 💊 Symptom-to-Specialist Detection – Detects illnesses with fuzzy logic and matches to medical specialty
• 📞 Call & Contact Features – Includes support number, email contact, and animated UI
What I Learned • How to implement Retrieval-Augmented Generation (RAG) using FAISS and OpenAI
• Practical use of LangChain, prompt templates, and semantic search
• How to build a polished, responsive UI with Tailwind CSS and Framer Motion
• How to structure and clean health-related Q&A data from real PDFs
Challenges I Faced • Pinecone DNS issues led me to pivot to FAISS, which turned out more stable
• Designing AI responses to feel helpful without guessing
• Creating a UI that feels as approachable as it is functional
• Balancing cost (API usage), performance, and complexity
Current Status
Health Nest is fully functional and ready for real-world use. It combines the power of GPT, medical data, and real-time location logic — all inside a clean, modern UI.
Log in or sign up for Devpost to join the conversation.