LynetteN/Green-Vision-Cameroon

An farmers disease guide # GreenVision Cameroon 🌱🇨🇲 GreenVision Cameroon is an AI-powered agritech application designed to assist farmers in Central Africa (specifically Cameroon) with early detection of crop diseases. It focuses on key staple crops like Cassava and Huckleberry (Njamarjama). ## 🚀 What the App Does The application serves as a digital extension officer, providing three core services: 1. **AI Disease Diagnosis (Visual)**: - Users can scan leaves using their camera or upload photos. - The app utilizes **Google Gemini 2.5 Flash** to analyze the image for specific symptoms (e.g., Mosaic patterns, leaf curling, chlorosis). - It provides a diagnosis with a confidence score and retrieves treatment protocols based on local agricultural research (simulated RAG). 2. **"Lili" - The Voice Assistant (Audio)**: - A conversational AI agent powered by **Gemini 2.5 Native Audio (Live API)**. - "Lili" speaks **Cameroon Pidgin English** and maintains a soft-spoken, sisterly persona. - Farmers can ask questions verbally and receive immediate spoken advice without typing. 3. **Local Context**: - Displays local weather conditions (mocked for Douala). - Provides treatment protocols specific to the region (e.g., referencing IITA/IRAD papers). ## 🛠️ Tech Stack - **Frontend**: React 19, TypeScript - **Styling**: Tailwind CSS - **AI/LLM**: Google Gemini API (@google/genai SDK) - *Vision*: `gemini-2.5-flash` for image analysis. - *Voice*: `gemini-2.5-flash-native-audio-preview` for real-time low-latency conversation. - **Icons**: Lucide React ## 🔮 Future Improvements & Roadmap To make this application production-ready for rural Cameroon, the following improvements are planned: ### 1. 📶 Offline Capabilities (Critical) Internet access in rural farms is often intermittent or non-existent. - **PWA Implementation**: Convert the app to a Progressive Web App (PWA) with service workers to cache UI assets. - **TensorFlow.js / MediaPipe**: Integrate lightweight client …