Why a PDF may contain no extractable text

A PDF can mix text, photographs and vector drawings. Visible letters may simply be pixels in a full-page scan, leaving no characters for a text extractor to read. Such pages need image OCR.

Even with a text layer, internal storage order may differ from reading order. Columns, running headers and table cells can become interleaved. Compare important passages with the original page.

Do not overwrite the original because an extraction is empty. Font character mapping can affect Korean text extraction. Murphkitsch’s text-to-PDF tool embeds a Korean font and has been checked by extracting the generated text again.