-
-
Agri assist Dashboard
-
Multi Language support
-
Smart Irrigation
-
Disease Diagnosis
-
weather prediction
-
Multilingual Chatbot with voice interaction
-
Gesture support for illeterate
-
Government schemes
-
Live market insights for farmers
Inspiration
Farmers in rural India still rely heavily on guesswork because they don’t have access to real-time soil insights, early disease prediction, market updates, or simple digital tools they can actually use. Many use low-end phones and face literacy barriers. We wanted to build something that feels like a digital farming companion — one that speaks their language, understands their field conditions, and guides them every day. AgriAssist was inspired by the need to give farmers AI-powered, easy, and hyper-local support.
What it does
AgriAssist provides live soil sensor readings, 5-day NDVI-based disease prediction, smart AI crop suggestions, and camera-based disease diagnosis. It creates an automatically updating farming checklist, sends weather alerts, gives market insights, shows nearby fertilizer shops, and highlights relevant government schemes. With multilingual voice support, gesture navigation, and a farmer community space, it becomes a complete farming assistant for low-literacy users.
How we built it
IoT sensors capturing soil moisture, pH, and NPK
AI/ML models including XGBoost for crop recommendations, LSTM for weather trends, CNN for disease diagnosis, and NDVI for early prediction
Backend built using FastAPI and serverless Firebase functions
Frontend built with Flutter (voice and gesture modes) and a Next.js dashboard
Databases on Firebase and Supabase
Integrated weather, government, and satellite NDVI APIs
Challenges we ran into
Poor and unstable network connectivity in rural regions
Making AI models lightweight enough for limited hardware
Designing an intuitive UI for uneducated farmers
Ensuring accurate early disease prediction with NDVI
Maintaining reliable real-time sensor integration
Delivering clear, multilingual recommendations
Accomplishments we’re proud of
Created a fully functional AI-powered farming assistant
Achieved NDVI-based early disease prediction up to 5 days in advance
Built gesture-based navigation for low-literacy users
Developed a dynamic, auto-updating farming checklist
Seamlessly integrated sensor data with actionable recommendations
Implemented a complete voice-driven assistance system
Connected farmers to markets, schemes, and shops in one place
What we learned
How to build AI systems that actually work in rural field conditions
Why designing for low literacy and low-end devices is crucial
The power of combining sensors, satellite layers, and ML models
How multilingual voice interfaces significantly increase adoption
The importance of continuous validation for disease prediction models
What’s next for AgriAssist
Automated irrigation and nutrient delivery
A village-level predictive agriculture intelligence dashboard
Partnerships with government departments, FPOs, and NGOs
Crop insurance integration, logistics booking, and marketplace features
More accurate hyperlocal disease models using multiple satellite layers
Deploying AgriAssist across more villages for real-world testing
Built With
- css3
- html5
- javascript
- next.js
- python
- react.js
- typescript
