gabrieltemtsen/afya-ai

Voice-first AI health triage for Africa — powered by Gemini 3.1 Flash Live # AfyaAI 🏥 > **Voice-first AI health triage for Africa** — built on Gemini 3.1 Flash Live AfyaAI lets anyone speak to an AI health assistant in their own language — English, Nigerian Pidgin, Yorùbá, Hausa, Igbo, Swahili, French and more. No reading, no typing. Just talk. ## What it does - 🎙️ **Real-time voice conversation** — hold the mic and speak, AI responds with voice - 🌍 **Multi-language** — auto-detects and responds in the patient's language - 📷 **Vision triage** — send a photo of a wound, rash, or medicine bottle → AI analyses it - 🚦 **Severity assessment** — Mild (home care) / Moderate (visit clinic) / Severe (hospital NOW) - ⚡ **Low latency** — powered by Gemini 3.1 Flash Live real-time audio model ## Stack - **Backend:** FastAPI + Gemini Live WebSocket proxy - **Frontend:** Vanilla JS, Web Audio API, mobile-first - **Deploy:** Railway (Docker) ## Quick Start ```bash # Clone git clone github.com cd afya-ai # Set up env cp .env.example .env # → Add your GEMINI_API_KEY # Install & run pip install -r backend/requirements.txt python backend/main.py # Open localhost ``` ## Deploy to Railway 1. Push to GitHub 2. New Railway project → Deploy from GitHub 3. Set env var: `GEMINI_API_KEY=your_key` 4. Done — Railway detects the Dockerfile automatically ## Disclaimer AfyaAI is an AI assistant and is NOT a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of a qualified healthcare provider. --- Built with ❤️ for Africa | Powered by Google Gemini 3.1 Flash Live