Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
Intelligent, accessible, and transparent medical triage powered by AI
Democratizing healthcare access in Vietnam and emerging markets through AI-powered triage
Comprehensive project documentation is available in the /docs folder:
- Problem Statement & Solution - Healthcare challenges in Vietnam and how Medagen addresses them
- Patient Intake System - Interactive body map, image upload, multi-step wizard
- AI Triage Engine - ReAct Agent with transparent reasoning (Gemini + LangChain)
- Computer Vision System - 3-tier medical image analysis (derm/eye/wound)
- RAG Knowledge System - Vector search, Vietnam medical guidelines
- Real-Time Chat & WebSocket - Streaming ReAct flow, session management
- UI Components - Design system, responsive mobile-first UI
# Install dependencies
npm install
# Configure environment
cp .env.example .env
# Edit .env with your API keys
# Run development server
npm run devAccess: http://localhost:7860
API Docs: http://localhost:7860/docs
cd frontend
# Install dependencies
pnpm install
# Run development server
pnpm devAccess: http://localhost:3000
- ✅ AI-Powered Triage: Gemini 2.5 Flash + ReAct framework for transparent reasoning
- ✅ 3-Tier Computer Vision: Specialized models for dermatology, ophthalmology, wound care
- ✅ RAG Knowledge System: Vietnam Ministry of Health guidelines + WHO protocols
- ✅ Real-Time Streaming: WebSocket-based ReAct flow visualization
- ✅ Visual Symptom Input: Interactive body map + image upload
- ✅ Multi-Language: Vietnamese + English support
- ✅ Session Memory: Context-aware conversations
Backend: Fastify, LangChain, Gemini 2.5 Flash, Supabase (PostgreSQL + pgvector)
Frontend: Next.js 16, TypeScript, Tailwind CSS, shadcn/ui, Framer Motion
Private and proprietary.
Built with ❤️ for accessible healthcare
For detailed information, see the documentation.