Start with a readable scan
Capture printed text with even light, a complete crop and as little blur as possible. Rotate the scan into a readable orientation. A dark crease or reflection can hide letters before OCR even begins. If the source image is poor, recapturing it is often more useful than trying to fix the extracted text afterward.
Run text recognition
The app’s published features include OCR for extracting text from documents and TXT export. Use the text-extraction workflow for the scan you need. A PDF image and extracted text are different outputs: keep the original scan when the appearance or layout of the document matters.
Proofread before reusing
Compare the extracted text with the source, especially names, dates, decimal points and reference numbers. OCR can confuse similar letters and digits, and it may not preserve columns or complex tables. Correct mistakes before pasting the text into another document. Never assume recognition has verified the content of the original.