ABSTRACT As the digital world evolves, the issue of cyberbullying continues to escalate, affecting individuals across diverse cultural and linguistic contexts. While significant progress has been made in developing efficient cyberbullying detection systems primarily in English, these solutions have reached a saturation point, limiting their applicability and effectiveness in non-English speaking environments. This paper presents a comprehensive multilingual cyberbullying detection system designed to identify and address instances of online harassments in two languages used in Nigeria –Pidgin and Igbo. A prototype is developed that operates across data sets created for these two languages. Using this prototype, experiments are carried out with Multinomial Naive Bayes (MNB), Logistics Regression (LR), and Stochastic Gradient Descent (SGD) algorithms to detect cyberbullying in these two languages. The results of our experiments show an accuracy up-to 97% and F1-score up-to 96% on datasets for both the languages.