HARITRAKSHAK

AI-powered leaf health scanner that detects plant disease symptoms, explains confidence and severity, and suggests safe care steps.

Comment

Inspiration

HARITRAKSHAK was inspired by a real problem faced by plant owners, gardeners, and farmers: when leaves show spots, curling, yellowing, or pest marks, it is difficult to quickly understand what might be wrong and what to do next.

We wanted to build a simple AI-powered plant health scanner that can turn a leaf image into useful guidance. The project was also showcased at Government Polytechnic Ahmedabad and received local News18 Gujarati coverage, which motivated us to make it stronger for OpenAI Build Week.

What it does

HARITRAKSHAK lets users upload a leaf photo and receive an AI-assisted plant health analysis.

It shows the original image, Sobel edge detection, simulated thermogram, likely plant species, possible disease symptoms, confidence score, severity level, probable cause, and safe care recommendations.

The app also includes a follow-up chatbot, guest mode, analysis history for signed-in users, multilingual support, media spotlight, and a Netlify live deployment.

How we built it

We built HARITRAKSHAK using Next.js, React, TypeScript, Tailwind CSS, Firebase, Genkit, and Gemini API. The app is deployed on Netlify.

Codex helped us build and improve the project quickly. It assisted with the dashboard layout, upload flow, image analysis UI, result page, loading states, error handling, environment variable cleanup, Netlify deployment setup, README, and Devpost preparation.

GPT-5.6 is presented as the reasoning layer for clear explanations, follow-up questions, and care-plan guidance from model results and user context.

Challenges we ran into

Some challenges were making the scan flow smooth on both desktop and mobile, keeping the interface simple, handling temporary AI service errors, protecting API keys with environment variables, and preparing the project for Netlify deployment.

Another challenge was explaining plant health results responsibly. We wanted the app to help users triage symptoms without pretending to be a certified agricultural diagnosis.

Accomplishments that we're proud of

We are proud that HARITRAKSHAK is more than just “upload image → disease name.” It gives users a full workflow: visual filters, confidence, severity, probable cause, care steps, and follow-up chat.

We are also proud that this project was successfully showcased in college and received local media coverage.

What we learned

We learned that responsible AI apps need both useful answers and clear limits. For plant health, it is important to show confidence, ask for context, and recommend expert help when needed.

We also learned how much Codex can speed up real product work, from fixing bugs and building UI to preparing deployment and documentation.

What's next for HARITRAKSHAK

Next, we want to improve the disease dataset, add crop-specific models, include location and weather context, support more Indian languages, and build a stronger history system for tracking plant health over time.

We also want to make HARITRAKSHAK useful for schools, home gardeners, nurseries, and small farms.

Built With

Share this project:

Updates