Repair a damaged PDF
Rebuild a PDF that will not open, opens with an error, or shows blank pages because of a broken structure — and get a plain report of what was wrong, what was fixed, and how many pages came back. The file is scanned and rewritten inside this page, then checked by qpdf compiled to WebAssembly, so it never leaves your device. A repair cannot invent data that is missing, and this tool says so when that is the case.
Drop a damaged PDF here
or
Your file is repaired in this page only — it is never uploaded.
What this tool can repair
Most PDFs that refuse to open are not destroyed. The pages are usually still in the file; what has broken is the index that tells a reader where they are. A PDF ends with a cross-reference table listing the position of every object in the file, and a trailer pointing at the start of the document. When a download is interrupted, a file is copied badly, or a program edits a PDF without rewriting that table, the positions stop matching. Stricter readers refuse the file outright; others open it with errors or missing pages.
This tool ignores the old index and finds every object by scanning the file itself, then writes a new cross-reference table and trailer. It also corrects stream lengths that do not match their data, unpacks compressed object streams whose index was lost, finds the document catalog if the trailer no longer points at it, and rebuilds the page list when parts of it have gone.
What it cannot do
A repair can only work with the bytes that are there. If a file was cut off halfway through downloading, the second half was never saved, and no tool can bring it back — this one recovers the pages that are complete and tells you how many of the pages the file said it had actually came back. A page whose content was partly lost is kept with what survives and named in the report, because it may show partly or not at all.
It also cannot open a PDF that is password-protected and damaged: decrypting needs the very structure that is broken. And it does not fix what is drawn on a page — a scan that was blurry stays blurry. When there is nothing left to recover, the report says so plainly rather than handing you an empty file.
Reading the report
Every run ends with a list of what was found and done. A tick marks something that was fine or has been fixed, such as a rebuilt cross-reference table or a corrected stream length. An exclamation mark is a warning worth knowing — for instance that the file ends early, so something was cut off. A cross marks something lost for good, like pages whose content could not be recovered.
The headline gives the result in one line: repaired with every page recovered, partly repaired with fewer pages than the file claimed to have, or not recoverable. If the file turns out to be undamaged, it says that too — some files that misbehave are structurally fine, and the problem lies with the program opening them.
Why your file never leaves your device
The rebuilding is done by code running inside this tab, and the result is then read and rewritten by qpdf, a long-standing open-source PDF tool compiled to WebAssembly. qpdf is strict, so a rebuilt file it accepts is one that holds together. Nothing is uploaded, and once the engine has loaded the tool keeps working with the network switched off.
Once a file opens again, the other tools here work on it as normal — Organise PDF to drop any pages that came back blank, or Merge PDF to put a recovered document back together with the pages you still have elsewhere.
Frequently asked questions
How do I fix a PDF that will not open?
Drop it into the tool at the top of this page. It scans the file for every object, rebuilds the cross-reference table and trailer that readers rely on, and gives you a new copy along with a report of what was fixed. Most PDFs that fail to open have a broken index rather than lost pages, and those usually come back completely.
Can a corrupted PDF always be recovered?
No. A repair can only use the data still in the file. If the file was cut off or overwritten, the missing part is gone; the tool recovers the pages that are complete, tells you how many of the pages the file claimed to have came back, and names any page whose content was damaged.
What does 'rebuilt cross-reference table' mean?
A PDF keeps an index of where each object sits in the file. When those positions are wrong or missing, readers cannot find the pages even though they are there. Rebuilding means finding every object by scanning the file and writing a fresh, correct index.
Why does the tool say my PDF has no damage?
Because its structure read cleanly from end to end. If it still fails somewhere, the trouble is usually the program opening it, or something drawn on a page, rather than the file itself. The freshly written copy offered anyway sometimes satisfies a fussy viewer or printer.
Can it repair a password-protected PDF?
Only if the damage still lets the file be opened. Decrypting a PDF depends on the same structure a repair rebuilds, so a file that is both encrypted and badly damaged cannot be recovered here. Remove the password from an undamaged copy with the PDF password remover first.
Is my PDF uploaded to repair it?
No. The file is scanned, rebuilt and checked inside your browser tab, and never sent to a server. Once the page has loaded it works offline.