How it works

A partial-recovery pass for damaged PDFs: re-parses the file with tolerant settings and re-saves a clean copy.

When a PDF won't open, the cause is usually a bad xref table, malformed objects, or a truncated download. The tool loads the file via pdf-lib with throwOnInvalidObject set to false and ignoreEncryption set to true, then re-saves with useObjectStreams: false (which writes a simpler, more compatible file structure). For files that are just slightly broken (often the case for downloads that failed mid-stream or PDFs generated by a buggy library), this is enough to get back a working document. For files that are genuinely destroyed (truncated to 10% of original, garbage content, etc.), the tool will fail with an honest error. Output filename has -repaired appended. The result may be larger than the input because object streams are unpacked.

How to use it

1. Drop the broken PDF. Truncated downloads and files with bad xref tables are the best candidates.
2. Run the repair. The file is re-parsed locally with tolerant settings and rebuilt with a simpler structure.
3. Download the repaired copy. If the file was salvageable, the -repaired PDF downloads; truly destroyed files return an honest error.

Frequently asked questions

Can this recover any corrupt PDF?

No. It can fix files with minor structural issues (bad xref, malformed objects). Truly destroyed files (large truncations, corrupted page content) are beyond what client-side tools can do.

Will the repaired PDF be the same as the original?

If the original was salvageable, the page content is preserved. The internal structure is rebuilt, so file size and object layout may differ.

Is the file uploaded for repair?

No. Repair runs locally via pdf-lib in your browser.

Why is the repaired file larger?

The repair re-saves without object streams (a compression scheme used by modern PDFs). Trading a bit of size for maximum compatibility.