Documents are the resources that you process in order to get classification results from the API. A document contains, and is created directly from, a file that you want to process (such as a PDF or Microsoft Office file).