Repair PDF

Structural damage: a download that was cut off, a cross-reference table with wrong offsets, a missing or broken trailer, an index that no longer matches the objects. These are what usually stop a reader from opening a file, and the pages, fonts and pictures are still in the bytes behind them. Content that was overwritten or never arrived cannot be brought back.

Two repairers, in order. First qpdf checks the file and, if it reads it with warnings or errors, writes it out again with a fresh index. If qpdf can't read it at all, a second parser reads the file from the front object by object, keeps every page it finds and writes a new file, which qpdf checks again. Only a file that passes the check is called repaired.

Yes, for everything that survived - the page content isn't changed, only the structure around it. A truncated file loses the pages that were cut off; check the preview to see what is there.

No. The check and the repair run in your browser; the file never leaves your device.

It has to be opened first. An encrypted file is flagged rather than guessed at; unlock it in the list with its password and it is examined again from the readable bytes, and repaired if it needs it.

Because the file is sound, and handing back the same document under a new name would only look like something happened. If it still will not open in a particular program, the trouble is elsewhere: run Inspect to see what the file actually contains.