Pay only for what you use with Amazon Textract, a machine learning (ML) service that uses optical character recognition (OCR) to automatically extract text, handwriting, and data from scanned PDF documents, forms, and tables.