The Self-Supervised Anomaly Detection (ssad) library is a Python framework for self-supervised training of reconstruction-based autoencoder models for anomaly detection.