Repair a Damaged PDF
No account. No upload. Just the tool.
A PDF you need won’t open. Adobe Acrobat shows an error. The print dialog hangs. Your browser preview shows a blank page where the PDF should be. Signegy re-parses and re-saves the file using tolerant parsing — most damaged PDFs are structurally damaged but content-intact, and a clean re-save fixes them. All in your browser. No upload.
How It Works
PDF files have a few layers:
- Content streams — the actual page contents (text, images, vector graphics).
- Object dictionaries — descriptions of every page, font, image, form field.
- Cross-reference table — an index pointing to where each object lives in the file.
- Trailer — the entry point that says “start with this object”.
Most “corrupt PDF” errors come from layers 3 and 4: a tool wrote a malformed cross-reference table, or the trailer entries don’t match the actual file offsets, or objects ended up in the wrong order. The page contents are usually fine — viewers just can’t navigate to them.
This tool re-parses the file using pdf-lib’s tolerant mode (which doesn’t trust the cross-reference table — it walks the file directly) and re-saves with a clean, conformant structure. Most files that strict viewers refused will open fine in any viewer after a repair pass.
When to Use This
- A PDF won’t open in Acrobat but does open (badly) in Preview or Chrome.
- A PDF prints correctly but throws an error when you try to fill a form on it.
- A scanner produced a PDF that’s larger than it should be and slow to open.
- A merged PDF has visible content but bookmarks, links, or page numbers don’t work.
- A PDF that was generated by an old or buggy tool throws “invalid object” warnings.
What This Tool Won’t Do
- Encrypted PDFs. If the file is password-protected and you don’t have the password, this tool can’t recover it.
- Truly destroyed files. If pdf-lib’s tolerant parser can’t find any valid PDF objects, there’s nothing to repair.
- Damaged content streams. If a page’s actual content stream is corrupted (a bad image, a malformed text operator), the page may render blank or partial after repair. This is rare — most “corruption” is structural.
- OCR or recovery of deleted content. Repair restores what’s in the file, not what used to be in the file.
Pair With Other Tools
After repairing, the file is in a clean state. Common follow-ups:
- Compress PDF — repair often shrinks files, but compress can shrink further by rasterizing.
- Sign PDF online — sign the repaired version to lock it in.
- Merge PDFs — combine the repaired file with others.
Signegy provides general information, not legal advice. Consult a qualified legal professional for advice specific to your situation and jurisdiction.
Frequently Asked Questions
What kinds of corruption can be repaired?
Tolerant parsing recovers PDFs with minor structural issues: bad cross-reference tables, malformed object streams, missing trailer entries, mismatched object IDs, and similar metadata-level damage. The tool re-walks the file and re-emits a clean version. It can't recover content that was actually deleted or overwritten.
What can't be repaired?
Three things: (1) PDFs that are completely unreadable — if pdf-lib can't even start parsing, there's nothing to recover. (2) Encrypted PDFs whose password is unknown. (3) PDFs where the actual page content streams are corrupted — repair restores the structure but not the visual content.
Does it work on PDFs that crash my viewer?
Often, yes. Adobe Acrobat and some viewers refuse to open files with even minor structural issues that pdf-lib's tolerant mode handles. Re-saving through Signegy produces a clean PDF that any viewer should accept.
What does the output look like compared to the original?
Visually identical for everything that parsed correctly. Smaller, often, because the re-save uses object streams to compress the structure (a feature most authoring tools don't enable). Content layout, fonts, images, and form fields are preserved.
Will the repair change my PDF's page count or content?
No. The same pages with the same content are preserved. Only the underlying file structure (cross-reference tables, object layout) is reconstructed.