KhetAI
KhetAI (AgriAdvisor) tackles a simple but costly problem: millions of Indian farmers in low-connectivity regions have no fast way to diagnose crop disease, leading to wrong treatments and lost yield. KhetAI fixes this by running a multimodal AI model entirely on-device - no cloud, no internet, no recurring cost. A farmer uploads or captures a photo of an affected crop through a lightweight HTML/Tailwind frontend. A FastAPI backend routes the image to a local Gemma model served via Ollama, which analyzes the image and returns a diagnosis along with practical treatment and prevention steps - delivered in the farmer's own language rather than English-only agri-advice. Because the entire pipeline runs locally, KhetAI is a strong fit for AMD's on-device inference story: no GPU cloud dependency, fast local turnaround, and a clear path to running efficiently on AMD hardware for real-time diagnosis in the field. The roadmap extends from single-machine MVP to village-level local hubs and eventually a native mobile app
Category tags:For Innovators & Creators