Repair a damaged PDF

What does PDF corruption look like?

PDF corruption usually shows up as an error such as "the file is damaged and could not be repaired", blank pages, a page count of zero, or a document that opens in one app but not another. A PDF is a collection of numbered objects plus an index, the cross-reference table, that records the byte position where each object starts. When that index is wrong or missing, viewers can't find the pages even though the content is still in the file.

Typical culprits are downloads cut off before they finished, attachments mangled by an email gateway, files copied from a USB stick or network drive while still being written, and tools that edit a PDF but save the offsets incorrectly. A file that is much smaller than expected, for example 300 KB for a 40-page scan, has almost certainly been truncated rather than scrambled.

How does Repair PDF fix a damaged file?

Repair PDF fixes a damaged file in up to three stages, each more tolerant than the last. First the file goes through qpdf, a well-tested open-source PDF library compiled to WebAssembly. It reads every object it can find, reconstructs a missing or wrong cross-reference table by scanning the file, and writes a new PDF with a fresh, correct index.

The rebuilt file is then opened with pdf.js, the rendering engine used in Firefox, and the first page is parsed to confirm that the pages really load rather than trusting the file header alone. Only a result that passes this check is offered for download, so a repair that produced an empty shell is never reported as a success.

If qpdf can't make sense of the file, a second parser scans the raw bytes for objects, ignores the ones that are invalid and saves what remains. When even a full save fails, pages are copied one by one into a new document, and any page that can't be read is left out and counted in the summary, for example 18 pages recovered and 2 left out.

Can every damaged PDF be recovered?

No, only damage to the structure can be repaired, not content that is missing from the file. A broken index, wrong byte offsets, a missing trailer or a few unreadable objects are routinely fixed. A download that stopped partway is different: the bytes for the later pages never arrived, so no tool can bring them back, and only the pages before the cut can be saved.

Files that aren't PDFs at all are detected before any repair is attempted. If the first kilobyte lacks the PDF signature and no PDF objects appear in the first megabyte, the tool reports that the file is probably a web page or another format saved with a .pdf name, which commonly happens when a login page is downloaded instead of the document.

How do I get the best repair result?

The best repair result comes from starting with the most complete copy of the file you can find. A repair can only work with the bytes that are present, so a fresh download often succeeds where the most careful rebuild of a truncated copy can't. These habits improve the odds:

• Always work on a copy and keep the damaged original until you have checked the repaired file.
• If the file came from a website or email, download it again first. A complete copy beats any repair.
• Check the file size: a PDF that is much smaller than expected was probably truncated, and only the pages before the cut can be saved.
• After repairing, Compress PDF can tidy the file further, and PDF to Word can pull the text into an editable document.

What are the limits of PDF repair?

Repair fixes the structure of a PDF, not its content. Text that was never written can't be restored, and a digital signature stays invalid after repair because the bytes of the file changed. Files up to 200 MB are accepted; very large documents need a lot of memory, so a desktop browser copes better than a phone with files of several hundred pages.

When a file opens in one program but not in the one you need, repairing is often quicker than re-exporting. If you still have the source document, such as the original Word file or spreadsheet, exporting a fresh PDF from it is the cleanest fix of all. Browser-based viewers and printers tend to be stricter than desktop readers, so a repaired copy is also worth trying when a PDF prints blank pages or an upload portal rejects it as invalid.