Document AI Tools Research

Potential Technology to use for capturing document citation information

Minimum Requirements

Tool	Site	Status	Notes

Tool	Site	Status	Notes
Google	https://cloud.google.com/document-ai	Rejected	can extract text but not reference with associated citation text with hierarchy. IE {“1. text”} versus {{“1.“}, “”text”). emit JSON API responses which require programming to present information. used to only support OCR of image files. Now, it supports OCR of image and PDF but not Microsoft file formats.
Microsoft	https://azure.microsoft.com/en-us/products/ai-services/ai-document-intelligence	Rejected	used to only support OCF of image files. It now supports OCR of images, PDF, and Microsoft Word.
AWS	https://aws.amazon.com/textract/	Rejected	used to only support OCR of image files and PDFs but maximum file size was too small for majority of our documents. Maximum file size resolved.