ZizipoMalangeni/AgriSense-AI

AgriSense.AI is a free, mobile-first web application that lets any South African farmer — urban or rural — diagnose problems with their crops, livestock, soil and irrigation by uploading a single photo. Every answer is engineered to put water first, in line with the country's deepening water crisis. # AGRISENSE.AI – Project Documentation ### AI-powered diagnosis and water-first advice for South African agriculture 🌐 Live Website: AGRISENSE.AI **Version:** 1.1 · May 2026 **Developed by:** Zizipo Malangeni ## Overview AGRISENSE.AI is a mobile-first AI agriculture platform that helps South African farmers diagnose crop, livestock, soil, and irrigation problems using image analysis and AI-powered advisory tools. The system prioritizes water-saving recommendations to address South Africa’s growing water crisis. ## Core Features * AI crop, livestock, soil, and irrigation diagnosis * Water-first recommendations * AI agricultural advisor chat * One-tap sample image diagnosis * Mobile-first responsive interface * Markdown-based action plans * South Africa-focused agricultural guidance ## Tech Stack * **Frontend:** TanStack Start + React 19 + Vite 7 * **Styling:** Tailwind CSS v4 * **Backend:** Lovable Cloud + Supabase Edge Functions * **AI Model:** Gemini 2.5 Flash via Lovable AI Gateway * **Markdown Rendering:** react-markdown * **Notifications:** sonner ## Application Routes * `/` – Landing page and agriculture insights * `/diagnose` – Image upload and AI diagnosis * `/advisor` – AI-powered agriculture chat assistant * `/insights` – System architecture and technical documentation ## System Architecture AGRISENSE.AI uses a four-layer architecture: 1. **Client Layer** * React-based frontend with file-based routing * Mobile-first design for rural accessibility 2. **Edge Layer** * Supabase Edge Function acts as a secure proxy * Handles validation, CORS, and API protection 3. **AI Layer** * Gemini 2.5 Flash multimodal AI model * Supports image + text analysis 4. **Data Layer (Planned)** * Supabase Postgres + Storage * Diagnosis history and farm profiles ## Workflow 1. User uploads an image or selects a sample image 2. Image converts to Base64 format 3. Request sent to `agri-ai` edge function 4. AI generates diagnosis and water-saving recommendations …