ML-based OCR solution?

anyone using a ML based OCR solution with erpnext?
been looking for a reliable way to extract data from pdf invoices

i.e. email → erpnext → attachment → ocr → populate invoice doctype


It’s pretty easy with external OCR API services. Sending PDF file, API service will return an structured object including all recognized data. Just some field mapping and logical process.

Azure, AWS and others offers free tier for 1000 pages / month

was more thinking if anyone had this rigged and ready?