It is also honest about the limit. A file truncated before its content, or one whose page tree is destroyed, cannot be rebuilt from nothing, and Bindery tells you so instead of handing back an empty document. The rebuilt file is verified before it is offered for download. Because the parsing and the rewrite happen in your browser, a broken file — which often contains exactly the content you least want on a stranger’s server — is never uploaded for “analysis”. If the repaired PDF opens but a page looks odd, Extract pages and Organize let you salvage the good ones.
How it works
- Open Repair and choose the damaged file, or drop it anywhere
- Bindery re-parses, rebuilds and verifies the result on your device
- Download the repaired copy; the original stays untouched
Questions people ask
What kinds of damage can be repaired?
Broken or missing cross-reference tables, truncated trailers, trailing garbage after the end-of-file marker, and recoverable object corruption — the causes behind most “cannot open” errors. The rebuilt file gets an entirely fresh structure, which also fixes files that open in one viewer but not another.
What cannot be repaired?
A file with no recoverable page tree: for example a download that stopped before the page content arrived, or a file overwritten with something that was never a PDF. Bindery says clearly when that is the case rather than producing a blank document, and the original is left as it was.
Is the broken file uploaded to diagnose it?
No. Re-parsing and rebuilding run in your browser, and there is no upload endpoint for documents. Damaged files are often the most sensitive ones — a half-downloaded contract, a corrupted scan of records — which is exactly why the repair stays on your device.