Hi, I’ve been testing different OCR and parsing tools for extracting data from invoices and receipts (images and PDFs). From my experience so far:
- LlamaParse: Very good at picking out fields like invoice numbers and dates, but struggles when the invoice image is blurry.
- Mistral OCR: Handles blurry invoice images better, but sometimes misses fields like invoice numbers or dates even when they’re clearly visible.
These are just my observations after testing multiple invoices.
I’m not talking about “information extractor” nodes or AI agents — I mean the OCR/Parser itself that turns the document into usable text or structured data.
So please which OCR/Parser tools have you personally found to be the most accurate and reliable for invoice data extraction?