This project explores the use of Large Language Models (LLMs) to provide multilingual, localized, and targeted agricultural advisory to smallholder farmers in Uganda. By utilizing LLMs, the project aims to bridge the gap between farmers and agricultural experts, offering timely and context-specific advice that can help address challenges such as food insecurity caused by diseases, pests, and poor yields. The project focuses on developing and analyzing an agricultural question-answer dataset to ensure that LLMs, such as GPT-4, can generate relevant and coherent responses to farmer inquiries in multiple languages, including Luganda. This solution leverages mobile technology to make advisory services accessible to rural farmers, overcoming limitations like the shortage of experts and the need for localized advice.