PDF Text Comes Out Garbled, Missing or Out of Order: How to Fix It

TL;DR

Check the page limit first (free reads 5 pages), then the recognition languages, render resolution and level, then try Raw line breaks. Re-run extraction after each change; the project does not update by itself.

The problem

The output has the wrong words, whole pages missing, lines in the wrong order or a table smeared into text. It isn’t obvious whether the PDF, the settings or the app is at fault.

Checklist, in order

  1. Pages missing after page 5? That’s the free limit. The project shows how many pages are locked. Pro plus Run extraction again reads them.
  2. Text is gibberish or wrong script? Set Recognition Languages to what’s actually in the document (default is Simplified Chinese + English).
  3. Small print wrong? Raise Render Resolution to 400 DPI and keep level on Accurate.
  4. Names, codes or numbers altered? Turn Language Correction off.
  5. Headings, lists or paragraphs wrong? Switch Format to Raw line breaks, which keeps every line as read.
  6. Table is a pile of lines? Rebuild Tables is a Pro feature. Check it’s on, and re-run.
  7. “N pages could not be read”? Re-run at a higher resolution; if it still fails, the page may be corrupt. Re-export the PDF from its source.
  8. Result “came from an earlier version”? Run extraction again to rebuild the structure.

About scans

A crooked, shadowed or low-contrast scan limits any OCR. Re-scan flat with good light if the content matters. Check the output against the original for anything legal or financial.

Remember

Changing a setting doesn’t change an existing result. Always re-run the project.

FAQ

Why did only 5 pages come out? The free version reads the first 5 pages. Go Pro and run extraction again to read the rest.

Why is the text gibberish? Usually the wrong recognition language or a low render resolution. Fix both and re-run.

Why are some pages listed as failed? Those pages could not be read; they are listed and marked in the output so the gap is visible. Try again at a higher resolution.

Why do numbers change? Language correction can alter numbers and names; turn it off.

The app PDFExtract →