EXplainable Artificial Intelligence is a subfield of AI, which has experienced significant growth in recent years. The aim is to develop methods and tools to explain the results produced by AI algorithms, in particular predictors built from data using machine learning approaches. Dedicated to this task, PyXAI is a Python library allowing to compute explanations for predictions made from several well-known tree-based supervised learning models: decision trees, random forests, and boosted trees). PyXAI supports wo popular machine learning libraries: Scikit-Learn and XGBoost. Several types of explanation can be calculated: abductive (why this prediction?) and contrastive (why not another prediction?). Various classes of abductive explanations are proposed to the user allowing to achieve different compromises in terms of size / computation time.