Repair a PDF that won’t open
Re-parsed, rebuilt and rewritten with a clean structure — on your device.
Free: 3 jobs per UTC day, 25 MB per file. PDF exports include a visible Made with Bindery watermark. Pro exports clean PDFs.
Prefer the full studio? Open Repair PDF in the app →
A PDF that refuses to open often has a broken cross-reference table, a truncated trailer or junk appended after the end of the file. Bindery re-parses the document — leniently if it must — recovers every readable page, drops what cannot be read, and writes a fresh file with a clean structure.
Honest scope: a file whose page tree is completely destroyed cannot be reconstructed, and Bindery will tell you so instead of pretending. Bindery runs entirely in your browser — your file is never uploaded, so there is nothing to delete from a server afterwards.
How it works
- Open Repair and choose the damaged file — raw files are accepted
- Bindery re-parses, rebuilds and verifies the result
- Download the repaired copy; the damaged original stays untouched
Why a PDF stops opening
A PDF ends with a cross-reference table that tells readers where every object lives, and a trailer that points to it. Interrupted downloads truncate that table; email gateways and some cloud sync tools append junk after the end-of-file marker; generators with bugs write offsets that are off by a few bytes; a crash mid-save leaves half a file. Strict readers refuse all of these. Bindery re-parses the file leniently, walks the objects it can find, recovers every readable page, drops what cannot be read, and writes a fresh file with a clean cross-reference table and trailer. The repaired copy is then re-opened and checked before it is offered for download.
What can and cannot be recovered
- Fixable: broken or missing cross-reference tables, truncated or malformed trailers, garbage after the end-of-file marker, recoverable object corruption, wrong offsets.
- Partly fixable: a file cut short mid-way — the pages before the cut usually survive, the rest is gone.
- Not fixable: a file with no recoverable page tree, an encrypted file whose password you do not have (use Unlock if you know it), or a file that was never a PDF — a renamed Word document, an HTML error page saved with a .pdf extension.
Bindery says which of these you have rather than producing an empty document. If the file is a download, downloading it again is worth trying first.
Repairing in the browser versus uploading
| Repair PDF | Bindery | Upload-based tools |
|---|---|---|
| Where the file is processed | In your browser tab, on your device | Uploaded to the service’s servers |
| Intake | Raw bytes accepted, even when normal opening fails | Upload, then a server attempts recovery |
| Damaged files | Often contain exactly the content you least want on a server | Uploaded regardless |
| Verification | The repaired file is re-opened and checked before download | Varies |
| What exists afterwards | The original and the new file, both on your device | A server copy until the retention window ends |
| Works offline | Yes, once the app has loaded | No |
Questions people ask
What kinds of damage can be fixed?
Broken cross-reference tables, truncated trailers, trailing garbage data and recoverable object corruption. The rebuilt file gets a completely fresh structure.
What can’t be repaired?
A file with no recoverable page tree — for example one truncated before its content — cannot be rebuilt from nothing. Bindery says so clearly rather than producing an empty file.
Is my broken file uploaded for analysis?
No. The parsing and rebuilding run in your browser. Damaged files often contain exactly the content you least want on a server.
The file opens but a viewer says it is damaged — should I repair it?
Yes. Repair rewrites the structure cleanly, which removes the warning and often fixes printing and form problems caused by a malformed table.
Will repair change how the pages look?
No. Page content is copied as found; only the file structure is rebuilt. Pages that could not be read are dropped and reported.
My PDF is 0 bytes or a few hundred bytes — can it be repaired?
No — there is no content to recover. That file was never saved completely; look for the original or download it again.