AgriSaathi

Let's stand with the farmers!

Comment

Inspiration

Farmers often struggle to access timely information, expert advice, and market insights due to fragmented agricultural resources. I wanted to create a single, smart digital assistant that could simplify farming decisions — from crop management and weather forecasting to financial planning and government schemes — all in one place. Our goal was to use AI and data-driven tools to make agriculture more efficient, accessible, and profitable for farmers and all of this with no language barrier.

What it does

AgriSaathi is a farmer-assistance web application designed to make agricultural information more accessible, reliable, and easy to understand. It helps farmers:

📊 Check real-time mandi (market) prices of crops from government datasets (data.gov.in), filtered by commodity, market, and state.

🌐 Translate agricultural data into local languages using Google Gemini API, helping overcome language barriers.

🗣️ Listen to information through Text-to-Speech, enabling accessibility for farmers who prefer audio.

💬 Summarize complex information, allowing quick understanding of reports or advisories.

AgriSaathi aims to empower farmers by bringing all essential tools — price awareness, communication, and language accessibility — under one intuitive platform.

How we built it

Frontend:

Simple and responsive UI built with HTML, CSS, and JavaScript (jQuery).

User inputs for commodity, market, and state to fetch mandi data dynamically.

Backend:

Built with Node.js and Express.js.

Routes handle external API calls (like data.gov.in for mandi prices and Google Gemini APIs for summarization, translation, and TTS).

Data fetched using Axios, filtered locally for accuracy, and served in clean JSON.

APIs & Services Used:

data.gov.in Agmarknet API for real-time agricultural market data.

Google Gemini APIs for AI-powered text generation, translation, and audio synthesis (TTS).

.env configuration for API key security.

Challenges we ran into

⚙️ Handling circular JSON structures during data parsing from the government API.

🧩 Integrating multiple Google APIs (TTS, Translate, Summarizer) while keeping the backend lightweight and RESTful.

🔒 Managing API key security and environment variables in .env files.

🧠 Making the app work for farmers with limited technical knowledge, which meant simplifying the UI and adding multilingual + audio features.

Accomplishments that we're proud of

Successfully built a working prototype of AgriSaathi that integrates real government data with AI-based translation and speech synthesis.

Designed a system where farmers can query market prices easily and hear them aloud in their own language.

Built the entire backend architecture using Node.js, Express, and Google APIs in just a short development cycle.

Overcame multiple data-format and filtering issues while working with large, inconsistent government datasets.

What we learned

Deep understanding of RESTful API integration and error handling in Express.js.

Efficient use of Axios for real-world API consumption.

The importance of user accessibility and localization in real-world applications for diverse audiences.

Secure handling of environment variables and API secrets using .env.

How to build modular and maintainable backend routes for data fetching and processing.

What's next for AgriSaathi

🚀 Building a mobile-friendly version using React or Flutter to ensure farmers can easily access AgriSaathi from any device.

🧠 Adding AI-powered crop and market suggestions by integrating classical machine learning and statistical models to help farmers understand price fluctuations, demand patterns, and profitable crop decisions.

💬 Expanding translation coverage to include more regional Indian languages, with offline support for areas with limited connectivity.

🛰️ Integrating real-time weather and soil data APIs to provide localized insights for better crop and irrigation management.

📈 Deploying AgriSaathi publicly so farmers and agricultural institutions can actively use and contribute to improving the platform.

Built With

Share this project:

Updates

Submission history