The rapid spread of hateful content on online platforms poses significant challenges to maintaining civility and safety. While extensive research has focused on high-resource languages like English, there is a lack of studies on low-resource languages. This review evaluates hate speech detection techniques specific to under-resourced languages in India, summarizing recent advancements and exploring text preprocessing, feature selection, and three categories of machine learning models: conventional machine learning, deep learning, and transformer learning. The study also addresses model generalization, available datasets, and limitations of previous research, with a particular focus on challenges faced by the Gujarati language. Additionally, it proposes solutions to enhance efforts against online hate speech highlight opportunities for future research in sentiment analysis and hate speech detection in less-resourced Indian languages specifically Gujarati language.