Repair PDF
Processed in your browser. The file never leaves your device.
Frequent questions
The ones that fail because of structure, not content: a broken cross-reference table, a file cut short in a download, junk in front of the header, wrongly declared stream lengths or dictionaries with bad keys. We rebuild the index and recover the pages that are still inside the file.
We recover the complete pages that survive and tell you how many there are. Whatever was lost when the file was cut cannot be rebuilt, and if what is missing is the document catalogue there is nothing to return: we say so instead of handing you an empty PDF.
Everything we return has been read again by qpdf, a reference tool for validating PDFs, and the list of problems it found is shown under the result. If the file was fine, you get the original back untouched.
No. The repair runs in your browser with qpdf compiled to WebAssembly, served from our own domain and downloaded only when you open this tool. You can disconnect after the page has loaded.
Not directly: an encrypted PDF has to be unlocked first with the remove-password tool, then repaired. If you try, we tell you it has a password.
Your machine’s memory is the limit. We accept up to 100 MB per file, one at a time.