Sentiment analysis of online comments has become an important part for understanding public opinion and its impact on public figures. However, the black-box nature of many accurate sentiment analysis models hinders trust and interpretability. This paper introduces XS2A, a platform for explainable sentiment analysis of Arabic comments written about public personalities. XS2A employs transformer-based models for sentiment analysis and emotion detection alongside with explainability methods to provide transparent and understandable explanations for model predictions. This paper discusses explainability methods and their integration within the architecture of XS2A, highlighting the platform’s applications in public relations, social media analysis, and market research. We also discuss future directions for improving XS2A.