Stella-Achar-Oiro/zeya

AI-powered WhatsApp chatbot for maternal health education in Kenya # Zeya - WhatsApp AI Antenatal Education Chatbot An AI-powered WhatsApp chatbot providing personalized antenatal education to pregnant women in Migori County, Kenya. Built for a research study comparing chatbot-based education against standard SMS reminders. ## Features - **AI-Powered Responses**: Uses Google Gemini to provide accurate maternal health information based on WHO and Kenya Ministry of Health guidelines - **Danger Sign Detection**: Automatically detects 8 categories of obstetric emergencies in both English and Swahili, triggering immediate emergency responses with local hospital contacts - **Multilingual Support**: Responds in English or Swahili based on user preference - **Gestational Age Tracking**: Provides trimester-specific advice based on pregnancy stage - **Admin Dashboard**: React-based dashboard for monitoring users, conversations, and danger sign alerts - **Data Export**: Export conversation and engagement data to CSV for research analysis ## Tech Stack **Backend** - Python 3.11 / FastAPI - PostgreSQL 15 (database) - Redis 7 (conversation history cache) - Google Gemini AI (gemini-2.0-flash-exp) - WhatsApp Cloud API **Frontend** - React 18 / Vite - TailwindCSS - Recharts (analytics) **Infrastructure** - Docker / Docker Compose - Nginx (frontend serving) ## Quick Start ### Prerequisites - Docker and Docker Compose - ngrok (for WhatsApp webhook) - WhatsApp Business API credentials - Google Gemini API key ### Installation 1. Clone the repository: ```bash git clone github.com cd zeya ``` 2. Create environment file: ```bash cp backend/.env.example backend/.env ``` 3. Edit `backend/.env` with your credentials: ``` GEMINI_API_KEY=your_gemini_api_key WHATSAPP_PHONE_NUMBER_ID=your_phone_number_id WHATSAPP_ACCESS_TOKEN=your_access_token WHATSAPP_VERIFY_TOKEN=your_verify_token ``` 4. Start the application: ```bash make up make migrate ``` 5. Access the services: - Backend API: localhost - …