A set of pre-trained machine-learning models that predict (im-)politeness scores in texts.