How to repair a PDF
- Drop the broken file. Even one that no viewer will open.
- Click Repair. The structure is rebuilt first, since that keeps your text.
- Download the result. You're told whether it was repaired properly or salvaged as images.
Why a working PDF stops working
A PDF ends with an index of where everything in it lives. Damage that index and a viewer has no way to find page one, so it gives up on the entire document, even though every page is sitting there intact. This is far and away the commonest kind of PDF corruption, and it comes from dull causes: a download that dropped, a cloud sync that stopped halfway, a program that crashed mid-save.
Rebuilding is possible because the objects themselves are usually undamaged. Reading the file leniently, finding what is still there, and writing a fresh document with a correct index gets the original back with its text, fonts and images unchanged.
When rebuilding isn't enough
Sometimes the damage reaches further and nothing will parse the file properly. Rather than stopping there, the tool falls back to rendering whatever can still be displayed and assembling those pages into a new document. That gets your content back at the cost of the text layer: worse than a real repair, and much better than losing the document.
If the file opens but is merely enormous, that isn't damage; compression is the tool. If it opens but demands a password you have, unlock it first.
Common uses for repairing PDFs
- Fix an interrupted download: a file that arrived incomplete and now refuses to open.
- Recover after a crash: the document your editor was saving when it went down.
- Open a file from an old drive: an archived PDF that no longer loads.
- Rescue a synced file: a cloud copy that stopped uploading partway.
Frequently asked questions
How do I fix a corrupted PDF?
Drop the file and click Repair. The document is parsed as leniently as possible and written out fresh, which rebuilds the index viewers rely on. If that index was the only thing broken, and it usually is, the file opens again with its text intact.
What actually gets repaired?
The cross-reference table: the map telling a viewer where every object in the file lives. Interrupted downloads, failed syncs and careless exporters all damage it, and viewers refuse the whole document even though the pages underneath are fine. Rebuilding that map from the objects it can still find is what fixes it.
What if the file is beyond that?
There's a second attempt. Anything still renderable is recovered as images and assembled into a new document. You get your pages back, but as pictures, so the text stops being selectable. It's a salvage operation rather than a repair, and the tool tells you which of the two happened.
Can it recover a file that was never fully downloaded?
Partly. Whatever arrived can often be read; whatever never arrived is not there to recover. A truncated file usually comes back with its earlier pages and not its later ones.
Will repairing change my document?
The pages are copied across as they are. Repaired files are often slightly smaller, because rebuilding discards the accumulated layers of previous edits that a much-revised PDF carries around.
Is my damaged file uploaded?
No. Repair happens in this tab, which also means a file too sensitive to email to a recovery service can still be fixed.