Manatee

How to repair a corrupted PDF on a Mac

· 6 min read

A corrupted PDF is usually a file whose internal table of contents (the cross-reference table that tells a reader where every object lives) is missing, truncated, or pointing to the wrong places. The page content is often still present. A repair that scans the file for the objects that survive and rebuilds the table around them can recover most or all of the pages. If the file was cut short, the pages that never arrived are gone, and a good tool says so rather than guessing.

First, confirm it is corrupted and not something else

Preview’s error messages are terse, and several non-corruption problems produce the same “couldn’t open” dialog.

  1. Check the size. Command-I in Finder. A 0-byte PDF has nothing in it; a 2 KB “PDF” for a 40-page report is almost certainly an error page saved with the wrong extension.
  2. Check that it is actually a PDF. Open Terminal and run head -c 8 /path/to/file.pdf. A real PDF begins with %PDF-1. followed by a version digit. If you see <html or a doctype declaration, the download gave you a web page. If you see PK, it is a ZIP or an Office document renamed.
  3. Check the end. Run tail -c 64 /path/to/file.pdf. A complete PDF ends with %%EOF (possibly followed by a newline or two). If it ends with page content, or with nothing recognizable, the file was truncated.
  4. Try Quick Look. Select the file and press the space bar. Quick Look uses the same engine as Preview but is sometimes more forgiving about damaged trailers.
  5. Try a browser. Drag the PDF into Safari or another browser. Browsers carry their own PDF engines, and some tolerate errors that Preview rejects.

If the file is the right size, starts with %PDF, and still fails everywhere, you have real corruption. Preview can’t open this PDF walks through the non-corruption causes in more detail.

How PDFs get corrupted

A PDF is a collection of numbered objects (pages, fonts, images, text streams), a cross-reference table listing the byte offset of each one, and a trailer at the very end pointing to that table. Readers start at the end, find the table, and jump to each object on demand.

Anything that damages the end of the file breaks the whole thing, because the entry point is gone:

  • Interrupted downloads and email attachments. The most common cause. The file saved but the last part never arrived.
  • Interrupted saves. A crash or a full disk while an app was writing the PDF.
  • Incremental updates gone wrong. Every time a PDF is annotated and saved, a new section is appended. A failure mid-append leaves a file with a valid old state buried under a broken new one.
  • Transfer corruption. A flaky USB drive or a network copy that flipped bytes.
  • Scanner and printer software bugs. Some scanners write malformed tables that strict readers reject.

The encouraging part: objects earlier in the file are usually intact. A repair rebuilds the table by scanning for them.

Try the source first

Before any repair, check whether a clean copy exists. This is always better than reconstruction.

  • Re-download the file from the original email, website, or shared folder.
  • If you made the PDF, open the original document (Word, Pages, a scan, an export) and export again.
  • Check whether the sender has a copy, and ask for it to be re-sent as a link rather than an attachment if it is large.

Repair the PDF with Manatee

If no clean copy exists, use the Repair mode in Manatee. It scans the file for every object it can identify, rebuilds the cross-reference table and trailer, and writes a new PDF.

  1. Open Manatee. Everything is free for the first 24 hours, no card or email.
  2. Drop the PDF in, or right-click it in Finder and choose Fix with Manatee.
  3. Wait for the job to finish. PDFs repair quickly; a large scanned document may take a bit longer.
  4. Read the outcome. Fixed means every page was recovered. Partial gives a count, for example 34 of 41 pages. Not fixable means no page objects could be located.
  5. Open the repaired copy, written alongside the original. The original is never touched.

The repair runs entirely on your Mac. For contracts, medical records, and anything else you would not paste into a web form, that is the point.

What a partial result means for a PDF

A Partial PDF repair means some page objects were recovered and others were not. The missing pages were in the part of the file that did not arrive, or they depend on a shared resource (a font or an image) that was lost. Pages that come back are complete; there is no such thing as a half-recovered page with some text missing. The page count in the outcome is exact.

What repair will not do:

  • Recover pages that were never in the file. A download that stopped at 60 percent contains 60 percent.
  • Remove a password. An encrypted PDF is not corrupted, and Manatee will not attempt to open it without the password.
  • Fix a PDF that is really an error page. Go back to the download source.
  • Repair the disk the PDF sits on. If the drive is failing, copy everything off first.

Partial recovery explained describes how to read these outcomes.

If it opens but misbehaves

Some PDFs open but display blank pages, garbled text, or missing images, and some refuse to print. These are usually malformed rather than truncated, and a rewrite often cures them. Two options:

  1. If Preview opens the file, choose File, Export as PDF, and save a new copy. Preview rewrites the structure cleanly.
  2. If Preview opens it but the export fails or reproduces the problem, run it through Manatee’s Repair mode, which rebuilds from the objects rather than trusting the existing table.

Questions

Will the repaired PDF look the same? Yes, for the pages that are recovered. Repair does not re-render or recompress anything; it reuses the original page objects, fonts, and images. Text remains selectable and searchable where it was before.

Preview opens it but says some content could not be displayed. Is that corruption? Usually a font or an image resource that is missing or malformed. Try the Export as PDF route first; if the warning persists, a repair can sometimes reconnect the resource. If the resource itself was truncated, nothing can bring it back.

The PDF is from a scanner and every page is blank. Can repair fix that? Sometimes. Scanner software occasionally writes the image objects with a broken reference, and rebuilding the table reconnects them. If the images were never written (the file is far too short for a scan), rescan.

Can I repair a PDF on my iPhone or in the browser? Manatee is a Mac app, so the repair happens on the Mac. You can AirDrop the file from your phone, repair it, and AirDrop it back.