Repair PDF tries to fix PDFs that won't open, show errors or display blank pages. It rebuilds the file's internal structure to keep text and quality, and if that isn't possible it recovers every readable page as an image. You'll see what was wrong and how many pages came back — and the file never leaves your browser.
| Symptom | Likely cause |
|---|---|
| 'File is damaged' | Broken cross-reference table |
| Opens with blank pages | Missing or invalid page objects |
| Won't open at all | Truncated download |
| Fails after email | Extra bytes before %PDF |
Not always. Badly truncated or encrypted files may only partly recover.
Yes if the structure can be rebuilt; if pages are recovered as images, text becomes part of the picture.
Interrupted downloads, failed saves, email encoding problems and disk errors.
Yes. Drop them together.
No. Repair happens in your browser.
Repair first diagnoses the file: bytes before the %PDF header (common when a PDF is saved from an email or web page) are removed, and a missing %%EOF marker or cross-reference table is noted as a sign of truncation. It then opens the PDF with pdf-lib in a lenient mode that skips invalid objects, copies every readable page into a fresh document and saves it with a newly built cross-reference table — this keeps text, fonts and vector graphics intact. If the structure is too damaged for that, it falls back to pdf.js, which tolerates far more corruption: each page it can read is rendered at 144 DPI and placed into a new PDF as an image, so the content survives even though text is no longer selectable. The result says which method was used and how many pages were recovered.