Scanned PDFs are pictures of pages. Computers can't read them. Search returns nothing, copy-paste returns nothing, screen readers stay silent.
OCR recognises the text in the image and adds it as an invisible layer. The scan looks identical; the document becomes fully usable.
Prepare the scan
OCR works best on clean input. Deskew crooked pages. Improve contrast if the scan is faded. Crop out borders that confuse text detection.
A clean scan gives 95%+ accuracy. A messy one drops to 70-80%.
Run OCR with the right language
Open the PDF in Flint's editor and run OCR. Specify the document language. English is the default; most major languages are supported.
For multi-language documents, specify all relevant languages. Accuracy drops slightly versus single-language but still works.
Verify and correct
After OCR, try searching for a known phrase. If it finds correctly, OCR worked.
For critical fields (numbers, names, dates), spot-check the recognised text. Edit any errors in the editor — the OCR text layer is editable.
Save with the text layer
Save the file. The output looks like the original scan but is now searchable. Any future operations — search, copy, screen reader, find-and-replace — work on the recognised text.
For accessibility, also set the document language so screen readers pronounce the text correctly.
FAQ
What languages does OCR support?
Major Latin-alphabet languages (English, French, German, Spanish, Portuguese, Italian, etc.) plus many others. Cyrillic, Arabic, CJK — varies by tool.
How accurate is OCR really?
Clean printed text: 95%+. Low contrast or skewed: 75-90%. Handwriting: 40-70%. Always proofread critical data.
Will OCR work on photographs of documents?
Yes. Convert the photo to PDF first, then OCR. Phone photos in document mode work well.
Does OCR change the visual appearance?
No. The original scan is preserved; the text layer is invisible behind it.
Can I OCR a PDF with mixed scans and live text?
Yes. OCR adds text where there's none; existing live text is preserved.
OCR turns a picture into a document. Run it in Flint's editor and your scan becomes searchable, copyable, and accessible.