Description of a "plain language" classifier I built, the idea being that the application can detect whether a document's text is written in accordance with federal readability standards.