What Repair PDF does
When a PDF will not open — “file is damaged”, “could not repair”, a blank window, or a viewer that hangs — the fault is usually in the file's bookkeeping rather than in its pages. This tool tries two things in order: first it parses the file leniently and copies every page into a brand-new, cleanly structured PDF; if the page objects themselves are too damaged for that, it renders each page with PDF.js (which tolerates errors that stricter viewers refuse) and rebuilds the document from those images. Both paths run in your browser.
What usually breaks
| Damage | Symptom | Repair path |
|---|---|---|
| Corrupt cross-reference table or trailer | 'Damaged file' error, file opens in some viewers but not others | Structural rewrite — content fully preserved |
| Truncated download (file cut off) | Last pages missing or error at open | Rewrite recovers the pages that exist |
| Bad incremental update from an editor crash | Opens to an older version, or fails | Rewrite keeps the latest valid objects |
| Wrong line endings after email or FTP in text mode | Binary streams corrupted | Often unrecoverable; rasterize path may salvage what renders |
| Damaged content stream or font | Blank or partly drawn pages | Rasterize path — pages become images |
| Encrypted with an unknown password | Password prompt | Not damage — needs the password (Unlock PDF) |
The two outcomes
- “Structure repaired, content untouched” — the good case. Text stays selectable, images keep their resolution, and the file is usually smaller than before because leftover objects are dropped.
- “Pages re-rendered as images” — the fallback. You get a readable, printable document, but text is no longer selectable and the file is larger. Run OCR on it if you need the text back.
The message under the download button tells you which path was taken.
What cannot be repaired
- A file that is not a PDF at all — a renamed HTML error page, a zero-byte download, or a file that starts with anything other than
%PDF. - Data that was never downloaded: if the download stopped at 40%, the missing 60% of pages are gone. Re-download from the source.
- Pages that reference fonts or images whose data was overwritten.
- Documents corrupted by malware or deliberate tampering.
Before you repair
- Try re-downloading or asking the sender to re-export. A fresh copy beats any repair.
- Check the file size against the original; a file that is much smaller than expected is truncated.
- Try opening in a different viewer (Chrome, Firefox, Preview). If one succeeds, print or export from there.
- Keep the damaged file — repair produces a new file and never modifies the original.
Frequently asked questions
The repaired file has fewer pages than the original.
Only pages whose data exists in the file can be recovered. The missing pages were truncated or their objects were destroyed.
Why is the repaired file bigger?
It went through the rasterize path — pages are now images. The rewrite path usually makes files smaller.
Can it repair a password-protected PDF?
It can rebuild the structure of an encrypted file, but if the content is encrypted you still need the password. Unlock it first if you know it.
Will my form fields and links survive?
On the rewrite path, page-level annotations and fields are copied. On the rasterize path, they are lost.
How is this different from Compact PDF?
Compact PDF also rewrites the structure, but assumes the file loads normally. Repair PDF uses lenient parsing and falls back to re-rendering when loading fails.
Is the file uploaded?
No. Parsing, rewriting and rendering all happen in your browser.