This article shows the problems of applying toxicity detection models trained on contemporary social media content to medieval literature.