Abstract: Despite the availability of information on zoonotic diseases, many South African communities remain uninformed due to limited language inclusivity. This project addresses that gap by using rabies as a pilot case to enhance knowledge accessibility for Sepedi-speaking communities. As Sepedi is a low-resource language, we fine-tuned two large language models—GPT-4o-2024-08-06 and Gemini-1.5-Flash-001—using translated and curated rabies-related data. The resulting models were deployed through an interactive, user-friendly Streamlit web application designed to deliver culturally relevant, accurate, and accessible health information. By leveraging web-based technologies, the platform enables real-time interactions with the models, aiming to raise awareness and support informed decision-making at the community level.