Documents & upload

OCR and text recognition

How Dynafis makes scans, images and PDFs machine-readable.

OCR makes visible text from scans and image-based PDFs available for extraction. The result is a recognition proposal, not proof that every character is correct.

When OCR is used

OCR is required when a PDF has no usable text layer or the invoice is an image. It does not replace existing structured XML data.

Common error sources

Low resolution, shadows, skewed pages, small print and similar characters such as 0/O or 1/I can change values. Invoice numbers, IBANs, tax IDs and decimal separators need particular attention.

Verify the result

Compare extracted fields with the original view. Correct values in the review queue and document the decision instead of exporting recognised text unchecked.