In this project, we focus on identifying toxic and harmful comments written in the Chittagonian language, which is a low-resource and underrepresented language. Because Chittagonian lacks large annotated datasets and NLP tools, detecting toxic content becomes a challenging task.