E-krishi sheba

E-krishi Sheba

Inspiration

In Bangladesh, millions of farmers and rooftop gardeners lose crops every year because they cannot correctly identify plant diseases in time. Most rely on local shops or guesswork for pesticides, often using the wrong chemical or wrong dose. We wanted to build a simple, Bangla-first tool that helps both field farmers and city rooftop gardeners diagnose diseases quickly and get the right organic + chemical solution.

What it does

E-Krishi Sheba is an AI-assisted web platform that:

Diagnoses crop & rooftop plant diseases (rice blast, potato late blight, dragon fruit stem canker, citrus canker, tomato leaf curl, etc.)
Lets users scan/upload leaf photos (AI simulator) or use a 3-step symptom wizard
Gives dual remedies — organic (neem, Bordeaux) and chemical with exact dosage
Includes Tank-Mix compatibility checker, smart dosage calculator (tank / land / rooftop bottle), seasonal calendar, and nutrient deficiency guide
Fully in Bangla with community sharing

How we built it

We built a complete single-page web app using pure HTML, CSS, and JavaScript (no heavy frameworks) so it works even on low-end phones and free static hosting. All data (diseases, chemicals, PHI, calendar) is structured in JS. The AI photo scanner is simulated with sample results and file upload preview. LocalStorage is used for community posts. The entire app is bilingual-ready and supports light/dark theme.

Challenges we ran into

Making accurate dosage calculations for different units (decimal, katha, bigha, acre, and rooftop bottles)
Designing a simple interface that both elderly farmers and young rooftop gardeners can use
Keeping everything in one HTML file so it deploys easily on free hosts like tiiny.site
Balancing organic and chemical recommendations with proper safety (PHI) warnings

What we learned

We learned how important local language and simple UX are for real-world impact in agriculture. We also learned to structure agricultural knowledge (symptoms, weather conditions, tank-mix rules) into usable digital data, and how AI can assist diagnosis even with limited resources.