Inspiration

Farming today is more complicated than ever. Many industries are quickly transitioning to digital, but a lot of farmers still don’t have access to the right information when they need it. Changing market prices, unpredictable weather, and unexpected crop diseases create uncertainty that can greatly affect livelihoods.

We decided to create Krishi Saathi, which means Farming Companion, to close the gap between modern technology and farming. Our goal is to give farmers data-driven insights, AI support, and community help, making farming not just sustainable but also profitable.

What it does

Krishi Saathi is an all-in-one Agricultural Intelligence Platform that serves as a personal agronomist for farmers. It includes a detailed Web Dashboard and the Unnati mobile app. Key features include:

  • Smart Market Insights: Shows live Mandi prices (minimum, maximum, and modal) from markets across India, helping farmers choose the best time and place to sell.
  • AI Crop Recommendation: Looks at region, soil type, and season to suggest the most suitable and profitable crops along with a confidence score.
  • Disease Doctor: Uses AI-based computer vision to spot crop diseases, like leaf blight, from images and provides immediate treatment plans.
  • Predictive Analytics: Offers 7-day price predictions to ease uncertainty when selling produce.
  • Krishi Mitra Chatbot: A 24/7 multilingual AI assistant that responds to agricultural questions in local languages, such as Hindi and English.
  • Community & Commerce: Features a marketplace for seeds and fertilizers and a forum for peer-to-peer farmer support.

How we built it

We took a full-stack approach to ensure accessibility and scalability across devices:

  • Frontend: A responsive Web Dashboard for detailed analytics and the Unnati Mobile App for on-the-go access, built using modern frameworks like React or Flutter.
  • AI & Machine Learning:

    • Computer Vision: Trained a CNN-based model to analyze leaf images and identify crop diseases.
    • Predictive Models: Applied regression-based algorithms on historical agricultural data for crop recommendations and price predictions.
    • NLP: Used Natural Language Processing to support the multilingual Krishi Mitra chatbot.
  • Data Integration: Collected real-time data from weather APIs and government Mandi databases to ensure accuracy and timeliness.

Challenges we ran into

  • Data Fragmentation: Agricultural data in India is often fragmented and inconsistent in format, making it hard to integrate.
  • Multilingual Complexity: Training the chatbot to comprehend regional languages and agriculture-specific terms required significant adjustments.
  • Image Variability: Ensuring reliable disease detection despite poor lighting, blurry images, or inconsistent backgrounds was a challenge.

Accomplishments that we're proud of

Seamless Multilingual Experience: Instant switching between English and Hindi, enhancing accessibility for farmers.
Disease Detection Feature: Successfully providing instant, actionable recommendations through the AI-powered “Doctor” module.
Holistic Platform: Creating not just a tool but an ecosystem that combines advice, analytics, commerce, and community support.

What we learned

We discovered that accessibility is just as important as accuracy. Designing for farmers meant making complex data simple and actionable. We also gained a better understanding of how factors like soil quality, weather patterns, and market trends are interconnected and influence agricultural outcomes.

What's next for Krishi Saathi

IoT Integration: Connecting soil moisture sensors and drones for hyper-local farm intelligence.
Offline Support: Allowing essential features such as disease detection to function without internet access.
Supply Chain Integration: Linking farmers directly with logistics providers to transport produce to the most profitable Mandis.

Built With

Share this project:

Updates