Agrosarthi

Inspiration India’s agriculture sector is powered by millions of small and marginal farmers, many of whom lack timely access to expert advice, market data, and modern tools. We were inspired to create AgroSarthi after witnessing how challenges like unpredictable weather, pests, low yield, and poor market connectivity affect rural livelihoods. We wanted to build a platform that bridges this gap using the power of AI, regional language support, and real-time data.

What We Learned Throughout this project, we learned:

How to apply AI/ML to real-world problems like crop disease detection and yield estimation.

The importance of user-friendly design for rural users, especially in regional languages.

How to integrate multiple data sources (weather, market prices, soil info) into a unified platform.

The need for offline-first and voice-based interfaces to truly reach underserved communities.

How We Built It

Frontend: Built using React/Next.js for a responsive, multilingual interface.

Backend: Developed with Node.js and Express, connected to a MongoDB database.

AI Integration: We used custom ML models for crop recommendations and yield predictions, along with Gemini API for a conversational farming assistant.

APIs & Tools: Integrated real-time weather and market price APIs, and designed calculators for irrigation, fertilizer, and profit.

Deployment: The platform is hosted using Vercel and Render, ensuring easy access across devices.

Challenges We Faced

Ensuring multilingual input/output support with regional accuracy.

Building AI models that work with limited and noisy datasets.

Designing a simple UX suitable for farmers unfamiliar with technology.

Making the system usable in low-connectivity or offline environments.

Integrating real-time data from diverse and unreliable sources.

Despite these challenges, we created a robust MVP that can scale with future additions like IoT sensor integration, drone-based analysis, and a digital marketplace.

Built With

• gemini
• javascript
• nextjs
• shadcn