KRISHOKBONDHU

• 

  App UI

Inspiration To empower smallholder farmers in Bangladesh by bridging the literacy and technology gap through AI, local market data, and weather alerts in their native language.

What it does KRISHOKBONDHU is an offline-first mobile app featuring:

AI Plant Scanner: Takes photos of crops to identify diseases and prescribe exact medicine dosages.

Audio Accessibility: A prominent "Listen" button triggers clear Bangla Text-to-Speech for low-literacy users.

Hyper-Local Weather: Upazila-level forecasts for smart harvesting.

Live Market Prices: Real-time values for rice and vegetables to stop middleman exploitation.

How we built it Frontend: Flutter for high performance on budget Android devices.

Audio/UI: flutter_tts optimized for Bangla, paired with huge, easy-to-tap grid buttons.

Data & Storage: Local caching (SQLite) for seamless offline usage in rural areas.

Challenges we ran into UI Design: Rethinking standard mobile layouts to prioritize massive touch-targets and voice integration over text menus.

Bangla TTS Tuning: Dialing in the speech rate so complex agricultural instructions sound clear and natural.

Accomplishments that we're proud of Building a high-impact AgTech architecture from scratch.

Getting fully localized Bangla narration to work seamlessly within a clean dashboard layout.

What we learned User Empathy: Tech solutions must be designed for real-world limitations like slow networks and low-end hardware.

What's next for KRISHOKBONDHU Offline TFLite Model: Embedding custom crop-disease AI models directly into the app for 100% offline scanning.

SMS Fallover: Sending agricultural solutions over text messages when internet is completely unavailable.

Built With

• antigravity
• next.js