SmartAgro Ai

Frontend

Inspiration

In many regions, farmers do not have immediate access to agriculture experts when diseases or pest attacks appear. Delayed identification often leads to crop loss and wrong chemical usage. I wanted to build a tool that works like a field crop doctor — available instantly through a simple interface.

The idea was to combine modern AI models with voice interaction so even non-technical users can use it easily.

What it does

SmartAgro AI allows a farmer to:

Describe crop problems using voice input Type questions in natural language Upload crop/leaf images for analysis Get AI-generated diagnosis using Gemini Receive medicine and treatment suggestions See a simple risk score View a crop stage action calendar See live temperature context Listen to answers using text-to-speech Get responses in English or Hindi

How we built it

The project is built as a full-stack AI web app.

Backend

Python + Flask server Gemini AI API for reasoning and diagnosis Prompt engineering for structured agri advice Weather API integration Session-based chat memory

Frontend

HTML, CSS, JavaScript Responsive glass-style UI Browser Speech Recognition for voice input Speech Synthesis for read-aloud answers Mobile and desktop compatible layout

Deployment GitHub for version control Cloud hosting deployment Secure environment variables for API keys

Challenges we ran into

What I Learned During this project I learned: How to integrate large AI models into real web apps Prompt design for domain-specific answers Secure API key handling in deployment Voice input and speech output in browsers Flask production deployment workflow Designing AI tools for real-world users

Accomplishments that we're proud of

Some key challenges were: Handling API keys securely during cloud deployment Fixing environment variable issues on hosting platform Managing free hosting cold-start behavior Making voice features work reliably across browsers Keeping AI answers short but still practical Designing a UI simple enough for farmers

What we learned

Planned future upgrades include: Region-specific crop disease databases More local language support Offline/SMS interaction mode Fertilizer timing automation Pest risk prediction using weather trends Government scheme and subsidy suggestions