-
-
Enter your crop, soil, and city — get live weather-based advice on exactly when to water next.
-
Compare today's crop prices across nearby markets instantly, so you know where to sell for the best return
-
Snap a photo of any leaf — Gemini AI instantly flags disease or pests and suggests what to do next.
Inspiration
Agriculture supports the livelihoods of millions of farmers, yet many still lack timely, affordable access to expert advice. A farmer might notice a disease spreading across their crops but have no quick way to diagnose it, or grow a healthy harvest only to sell it at a low price simply because they didn't know a nearby market was paying more. We wanted to build something that puts a piece of that expertise directly into a farmer's hands — using nothing more than a phone camera and an internet connection.
What it does
AgriSense is a web app that helps farmers make smarter, faster decisions about their crops:
- Crop health check — Farmers upload a photo of a leaf or plant, and our AI model identifies likely diseases or pest damage and suggests remedies.
- Smart irrigation advisor — Based on soil type, crop type, and live local weather data, the app recommends an optimal watering schedule.
- Market price dashboard — Farmers can compare prices across nearby markets so they know where to sell for the best return.
- Simple, accessible UI — Designed to be usable by farmers with limited tech experience, with plans for multilingual support.
How we built it
- Frontend: React with Tailwind CSS for a clean, responsive interface
- Backend: Node.js/Express (or Flask) handling API requests and data flow
- AI/ML: Gemini API for image-based crop disease detection
- Weather data: OpenWeatherMap API for real-time local conditions
- Market data: Sample/mock mandi price dataset to demonstrate the price comparison feature
- Deployment: Vercel/Render for a live demo link
Challenges we ran into
- Finding reliable, free real-time market price data within the hackathon timeframe
- Tuning the AI model to give useful, farmer-friendly explanations rather than technical jargon
- Balancing feature scope against the limited build time
Accomplishments that we're proud of
- Got an end-to-end working prototype — photo upload to diagnosis to actionable advice — in a single hack day
- Successfully integrated the Gemini API for real image analysis
- Built a UI simple enough that someone with no tech background could use it
What we learned
- How to quickly integrate and prompt-engineer a vision-based AI API for a practical use case
- The real-world complexity behind agricultural data (market prices, weather variability) and how to design around incomplete data
- How to prioritize ruthlessly under a hard time limit
What's next for AgriSense
- Partnering with real agricultural data providers for live mandi price feeds
- Adding multilingual and voice-based interaction for low-literacy users
- Expanding disease detection to more crop types and integrating real IoT soil sensors for automated irrigation
Built With
- css3
- express.js
- gemini-api
- html5
- javascript
- node.js
- openweathermap-api
- react
- tailwindcss
