Continual anomaly detection is gaining attention in many domains where the notion of normality evolves over time and models must adapt without forgetting previously learned knowledge. To this end, we demonstrate pyCLAD, an open-source Python framework for continual anomaly detection. pyCLAD provides reusable abstractions and implementations for scenarios, strategies, models, metrics, and experiment analysis. The proposed library is the first framework to address the intersection of anomaly detection and continual learning, as existing libraries largely target offline anomaly detection or supervised continual learning. pyCLAD is available as a pypi package https://pypi.org/project/pyclad/.