Repair PDF
Repair PDF attempts to recover a file that will not open properly. It parses the document in a tolerant mode, rebuilds the cross-reference table and salvages every page it can still read, then writes a clean new PDF. Recovery is best-effort: badly truncated files may come back with fewer pages than the original.
1 Add your file
Enter the open password to use it here. It is checked in your browser and never sent anywhere.
2 Choose your options
Recovery is best-effort, not magic. If a file was truncated during a download or a disk error ate part of it, the bytes that are gone are gone — we can only rebuild what is still present. Expect a good result from a file with a broken cross-reference table, and a partial one from a file that is physically incomplete. The tool always reports how many pages it managed to recover.
How to repair PDF
Add the damaged PDF
Even a file that your usual reader refuses to open is worth trying here — our parser is more forgiving than most.
Choose a strategy
Start with Rebuild structure, which keeps text intact. If that produces nothing, fall back to Salvage as images.
Download what we recovered
The report tells you how many of the original pages came back, so you know exactly what you have.
What repair can and cannot rebuild
A PDF finds its contents through a cross-reference table at the end of the file, which lists the byte offset of every object. Lose or corrupt that table and a reader cannot locate anything — even though the objects themselves are usually still sitting intact a few kilobytes earlier. Most "damaged" PDFs are damaged in exactly this way.
Repair ignores the broken index and walks the file from the beginning, finding object definitions directly and recording where they really are. It then rebuilds the page tree from what it found and writes a fresh, valid document. In short, it stops trusting the map and surveys the ground.
A worked example
A 12 MB contract downloaded over a dropping connection arrives at 11.4 MB and will not open. The missing piece is the end of the file, which is where the index lives — so the pages are almost certainly fine. A rebuild typically recovers everything except perhaps the last page. Before running it, try downloading the file again: a complete copy costs a minute and fixes this outright.
Limitations worth knowing
- Content that was never written to disk cannot be recovered. If half the file is missing, half the document is missing.
- Bookmarks, tagging, form field definitions and annotations are frequently lost, because each depends on references across many objects.
- A file that is not actually a PDF cannot be repaired. Failed downloads often save an HTML error page under the original name — a real PDF begins with %PDF-.
- Repair rewrites the document, so any digital signature it carried is invalidated.
- The result should be checked page by page; a rebuild can succeed structurally and still be missing content.
Further reading
Why PDF Files Get Corrupted, and What Repair Can Actually Recover
Related tools
Repair PDF — frequently asked questions
Broken or missing cross-reference tables, damaged trailers, files that were edited by software that wrote them badly, and files with unreferenced objects. These are the most common causes of 'this file is damaged and could not be repaired' and they usually come back completely.
Bytes that are not there. A download that stopped halfway, a file truncated by a full disk, or a file encrypted with a key nobody has. We recover the pages that survive and tell you how many that is.
With Rebuild structure, yes — the page content objects are the originals. With Salvage as images, the pages look the same but are pictures, so text is no longer selectable or searchable.
No. Repair runs in your browser like every other tool here. Nothing about the file, damaged or not, leaves your device.
Because the data for the missing pages could not be read at all. With Skip unreadable pages ticked we return everything we could rescue rather than failing outright, which is usually more useful than nothing.