AgriSense – AI-Powered Smart Farming Platform

Think of us as your farm’s Jarvis — diagnosing plants, predicting yields, and even checking the weather before the clouds do.

Comment 1

🌾 AgriSense – AI-Powered Smart Farming Platform

🚜 Inspiration

Farmers in India face unpredictable weather, fluctuating market prices, and limited access to expert guidance. We wanted to build something that gives farmers AI-backed insights directly on their phone or browser — no middleman, no jargon — just actionable recommendations.

🌱 What It Does

AgriSense is a full-stack smart farming platform that empowers farmers to:

  • Get AI-powered crop recommendations tailored to soil and climate
  • Predict crop yields using machine learning models
  • Perform real-time weather and soil analysis
  • Upload plant photos for disease detection and treatment advice
  • Access market price data from Indian agricultural markets
  • Interact with a multi-language AI chatbot for instant support
  • Track farm assets and sensors via a farmer dashboard

🛠️ How We Built It

  • Frontend: React.js (Vite + Tailwind), multi-language support, camera/image upload, dashboard visualizations
  • Backend: Node.js/Express.js with MongoDB, JWT authentication, and integrations to weather and Agmarknet APIs
  • AI/ML Layer: Python (FastAPI + TensorFlow + HuggingFace Transformers + OpenAI API) for crop prediction and plant disease classification
  • APIs & Data: WeatherAPI, Agmarknet, Unsplash, Nominatim for geolocation
  • Deployment/DevOps: Docker containers, PM2 process management, Nginx reverse proxy

Auth & Farmer Management

  • POST /farmer/signup – Register new farmer
  • POST /farmer/login – Farmer login
  • GET /farmer/profile – Fetch farmer profile
  • PUT /farmer/update – Update farmer data
  • GET /admin/farmers – View all farmers (Admin)

Dashboard & Data

  • GET /farmer/dashboard – Farmer dashboard with weather, soil, and asset data
  • GET /farmer/crop-prices – Real-time crop price data via Agmarknet API

AI Model Endpoints (Python Service)

  • POST /farmer/recommend-crop – Calls {PYTHON_API_URL}/m1/crop-recommendation
  • POST /farmer/predict-yield – Calls {PYTHON_API_URL}/m1/yield
  • POST /farmer/voice-query – Calls {PYTHON_API_URL}/m1/voice-query
  • POST /farmer/detect-disease – Upload image, calls {PYTHON_API_URL}/m2/plant-disease
  • POST /farmer/chat – HuggingFace GPT-OSS model
  • POST /farmer/finance-advice – GPT-OSS + Unsplash images

Assets Management

  • GET /farmer/assets/:farmerId – Get all assets for a farmer
  • POST /farmer/assets – Create/update farmer assets
  • GET /farmer/assets/:farmerId/stats – Fetch asset statistics
  • GET /admin/assets – Admin get all farmer assets

Devices Management (IoT Layer)

  • GET /farmer/devices – List all devices of a farmer
  • POST /farmer/devices – Add a new device
  • PUT /farmer/devices/:deviceId – Update a device
  • DELETE /farmer/devices/:deviceId – Delete a device
  • PATCH /farmer/devices/:deviceId/status – Update device status
  • GET /farmer/devices/analytics – Device analytics
  • POST /devices/:deviceId/data – Simulate incoming device data
  • GET /admin/devices – Admin view all devices

Python Model Endpoints (Service Side)

  • POST /m1/crop-recommendation – Internal crop recommendation logic
  • POST /m1/yield – Internal yield prediction logic
  • POST /m1/voice-query – Internal handle voice queries
  • POST /m2/plant-disease – Internal plant disease detection

⚡ Architecture Diagram

Architecture

💡 Key Features

  • AI Recommendations – crops, yield, soil, disease detection
  • Smart Dashboard – analytics, weather, soil health, market prices
  • Asset Management – sensors, equipment, and drones
  • AI Chatbot with Voice Support – English, Hindi, and Kannada
  • Financial Advice – government schemes and microloans suggestions

⚔️ Challenges We Ran Into

  • Integrating multiple APIs (weather, Agmarknet, HuggingFace) with rate limits
  • Running AI inference quickly enough on low-power servers
  • Multi-language support without bloating the frontend bundle
  • Ensuring secure auth and location data privacy

🏆 Accomplishments That We’re Proud Of

  • Built a fully working microservices platform in the hackathon timeframe
  • Successfully deployed an AI model for plant disease detection via web upload
  • Integrated real-time crop price data for Indian markets
  • Designed a farmer-friendly UI with multilingual support and voice input

📈 What We Learned

  • Designing AI pipelines that run on both cloud and edge devices
  • Balancing usability with deep analytics for non-technical users
  • Leveraging modern dev tools (Vite, Tailwind, PM2, HuggingFace) for fast iteration

🚀 What’s Next for AgriSense

  • Mobile app version with offline support
  • IoT integration for automatic soil sensor readings
  • Expanding to more Indian languages and regional dialects
  • Blockchain-based supply chain traceability for crops

👥 Team

Member Role GitHub
Ravindra Full Stack Dev & AI Integration ravindraogg
Nitesh Panati Project Lead & Backend Architecture PanatiNitesh
Pooja CG Frontend Dev & UI/UX Design Pooja-CG

🧑‍💻 Technologies Used

  • Frontend: React.js, Tailwind CSS, Vite, Lucide React
  • Backend: Node.js, Express.js, MongoDB, JWT
  • AI/ML: Python, FastAPI, HuggingFace Transformers, OpenAI API
  • External APIs: WeatherAPI, Agmarknet, Unsplash, Nominatim
  • DevOps: Docker, PM2, Nginx

Built With

  • gpt-oss
  • huggingface
  • openai
  • react
Share this project:

Updates