CoffeeGuardian AI

CoffeeGuardian AI – Empowering mountain farmers with instant, AI-powered coffee plant health diagnosis and sustainable farming insights.

Comment

Inspiration

Coffee farmers in mountainous regions often struggle with plant diseases, unpredictable harvests, and limited access to expert advice. We wanted to use AI to bring real-time, accessible diagnostics and sustainable practices directly to their fingertips — empowering even small farmers to improve yield and reduce losses.

What it does

CoffeeGuardian AI allows farmers to:

  • Diagnose coffee leaf diseases instantly by uploading or snapping a photo.
  • Predict optimal harvest times based on current weather and plant conditions.
  • Get sustainable farming recommendations.
  • Access a smart assistant to answer questions about cultivation, pests, and best practices.

How we built it

  • Frontend: React-based web app with a beautiful, mobile-friendly UI.
  • Prototype logic: Simulated disease detection based on image clarity and randomized outputs for realistic flows.
  • Data: Compiled a database of 100+ coffee diseases with symptoms and treatments.
  • AI integration (planned): Designed the API contract for future OpenAI Vision or custom model integration.
  • Sustainability features: Dynamic UI components with actionable tips.

Challenges we ran into

  • Distinguishing between healthy and diseased leaves reliably in the prototype phase without a real AI model.
  • Creating realistic disease diagnosis flows for demos.
  • Sourcing high-quality, diverse images of coffee leaf diseases for future model training.

Accomplishments that we're proud of

  • Built a fully functional prototype with smooth UI/UX that simulates real AI behavior.
  • Designed a scalable architecture ready for integration with a production-grade AI model.
  • Developed a comprehensive disease and sustainability database specifically for coffee farming.

What we learned

  • The importance of clear feedback to users when diagnosing plant health.
  • How critical balanced datasets are to avoid prediction bias in future AI models.
  • Building inclusive, easy-to-understand interfaces can make technology accessible to uneducated farmers as well as professionals.

What's next for CoffeeGuardian AI

  • Integrate a real AI disease detection model using OpenAI Vision or a custom CNN trained on our disease image dataset.
  • Expand to support other crops grown in mountainous regions, like tea or cocoa.
  • Add offline functionality for farmers in low-connectivity areas.
  • Partner with local cooperatives to distribute the tool widely and collect more real-world data for continuous improvement.

Built With

Share this project:

Updates