AgriMate Hub

An AI-powered agricultural assistance platform for Indian farmers, combining crop recommendation, plant-disease detection, market intelligence, and a produce marketplace in a single, accessible web application.

Project Overview

Indian farmers face fragmented access to agricultural knowledge and services: choosing the right crop for their soil, diagnosing crop diseases early, finding current mandi (market) prices, discovering government schemes, and reaching buyers. AgriMate Hub addresses this by unifying these services into one platform with a clean, multilingual-ready interface and a machine-learning backend.

The problem space includes:

• Crop selection based on soil nutrients (N, P, K), climate, and rainfall.
• Early disease detection from leaf images to prevent crop loss.
• Market transparency via live mandi price data and price alerts.
• Scheme discovery so farmers can access subsidies and government programs.
• Direct selling through a marketplace that connects sellers and buyers.

Solution

AgriMate Hub is a full-stack web application split into two cooperating components:

• Frontend — a React single-page application (SPA) that presents the farmer-facing experience: a landing page, a category-aware dashboard (Agriculture / Horticulture / Floriculture), and feature modules for crop recommendation, disease prediction, government schemes, mandi prices, marketplace selling, and profile management.
• Backend — a Django REST API that serves data and runs the ML models. It exposes endpoints for users, crops, disease predictions, schemes, mandi prices/alerts, and marketplace listings, and integrates with the data.gov.in Open Government Data API for real-time mandi prices.

Architecture

┌─────────────────────────────────┐         ┌──────────────────────────────────┐
│         Frontend (SPA)          │         │        Backend (Django)          │
│                                 │         │                                  │
│  React + Vite + TypeScript      │  HTTP   │  Django 5 + Django REST Framework│
│  Tailwind CSS + shadcn/ui       │ ──────► │  ML: scikit-learn / TensorFlow   │
│  React Router + TanStack Query  │  JSON   │  SQLite (dev) / PostgreSQL (prod)│
│  Recharts / framer-motion       │ ◄────── │  data.gov.in integration         │
│  (mock data, API-ready)         │         │  CORS-enabled REST API :8000     │
└─────────────────────────────────┘         └──────────────────────────────────┘

The frontend runs as a static SPA (typically on http://localhost:5173 or :8080) and communicates with the backend REST API on http://localhost:8000. Cross-Origin Resource Sharing (CORS) is configured on the backend to trust the frontend's dev origins. The backend automatically loads the trained ML models on startup and falls back to rule-based logic when a model or external API is unavailable.

Request flow:

1. The browser loads the React SPA and renders the landing/dashboard UI.
2. Feature modules call backend REST endpoints (e.g. POST /api/crops/recommendations/, POST /api/diseases/predictions/).
3. The backend runs the relevant ML model or queries data.gov.in, then returns structured JSON (with confidence scores) consumed by the frontend via TanStack Query.

Tech Stack

Frontend

• React 18 with TypeScript
• Vite (build tool / dev server) + @vitejs/plugin-react-swc
• Tailwind CSS + shadcn/ui (built on Radix UI primitives)
• React Router (client-side routing)
• TanStack Query (server-state / data fetching)
• framer-motion (animations), Recharts (charts), lucide-react (icons)
• react-hook-form + zod (forms & validation)
• Vitest + Testing Library (unit tests), ESLint (linting)

Backend

• Python 3.10+ with Django 5.0 and Django REST Framework
• django-cors-headers (CORS)
• scikit-learn (Random Forest crop recommendation) + joblib
• TensorFlow / Keras (CNN plant-disease detection) + h5py
• Pillow (image handling), pandas / NumPy / SciPy (data processing)
• SQLite (development DB; PostgreSQL recommended for production)
• Optional: FastAPI/uvicorn utilities, python-decouple (config), OpenAI fallback

Setup and Installation

Both components are independent and must be run together for the full experience.

1. Backend

cd backend
python -m venv venv
venv\Scripts\activate            # Windows  (use: source venv/bin/activate on Linux/macOS)
pip install -r requirements.txt
python manage.py migrate
python manage.py runserver       # http://localhost:8000

See backend/README.md for full details, ML model notes, and API endpoints.

2. Frontend

cd frontend
npm install                      # or: bun install
npm run dev                     # http://localhost:5173

See frontend/README.md for full details and scripts.

Tip: keep the backend running on :8000 while developing the frontend so the feature modules can call the REST API. Configure CORS_ALLOWED_ORIGINS in backend/agri_sahayak/settings.py to match the frontend's dev URL.

Project Structure

agrimate-hub/
├── README.md            # This file (root overview)
├── backend/             # Django REST API + ML models
└── frontend/            # React SPA (Vite + TypeScript)