T-Rex (Trainable Relation Extraction) is a highly configurable machine learning-based Information Extraction from Text framework, which includes tools for document classification, entity extraction and relation extraction.