Amazon Textract

• automatically detect text, handwriting, and data
  • also the relationship between text (eg item/ price in invoice)
  • metadata (where text occurs)
  • document analysis (names, address, bday)
  • receipt analysis (price, vendor, line items, dates)
  • identity document (abstracts fields eg DocumentID)
• input = document (JPEG, PNG,PDF,TIFF)
• output = extracted text, structure of text & analysis
• synchronized (real-time) for most doc, asynchronized for large doc
• pay as you go
• integrate with your own applications