Abstract: Due to diverse preferences, netizens frequently experience a lack of informed consumption of Afrikaans music videos on social media platforms. This project addresses the issue of uninformed consumption of Afrikaans music videos on social media by developing a multimodal sentiment-based music recommendation system. The system is designed to enhance user experience by facilitating quick access to relevant music video recommendations by utilizing innovative data collection and multimodal sentiment analysis. It leverages advanced Natural Language Processing (NLP) and feature extraction techniques to analyze sentiments from song lyrics, and audio components, enhancing the understanding of user preferences and emotional resonance within Afrikaans music videos. Sentiment analysis results from multinomial Naive Bayse and Keras models are averaged to determine an overall sentiment for each song, which is then used as input for the recommendation model. Two versions of the recommendation system are deployed: Version 1 allows users to input their preferred sentiment, genre, and year to receive personalized music video recommendations; Version 2 accepts YouTube video URLs, provides song metadata, and suggests similar songs.