Drop in a PDF that won't open, shows errors or comes up blank. This tool rebuilds it and recovers every readable page inside your browser — no server, no account.
or choose a file from your device
The file loads into the page's memory only. No network request is made with your document.
The tool checks the header, end marker and cross-reference table, then rebuilds the structure and tests every page to see which ones really open.
You get a clean PDF with a report of what was wrong and how many pages came back.
Damaged files are often the important ones. Yours is never sent to a third party's server.
Processing happens in your browser's memory, so there's no server queue or file-size paywall.
Once this page is loaded, repairing a file doesn't need the network again.
Missing or broken cross-reference tables, a missing end-of-file marker (common with interrupted downloads), junk bytes before the header, bad object pointers, and files where some pages are corrupt but others are fine.
If the damage is structural, usually yes. If the actual page data is corrupt (a truncated download, a failing disk), only the pages that can still be read are recovered. The report tells you exactly how many came back.
Leave it on Automatic. It rebuilds the file two different ways, tests each result page by page, and keeps the better one. If neither opens, it falls back to rescuing pages as images.
"Rescue as images" redraws each readable page as a picture. It recovers how the pages look, but not the underlying text. Use it only when the text-preserving methods fail.
Not yet. Remove the password first with the tool you normally use, then repair the file. Encrypted structures can't be safely rebuilt without the key.
No. Everything runs in your browser tab using pdf.js and pdf-lib. You can check the Network tab in your browser's developer tools and see that nothing is sent.
If the report says no objects or pages were found, the file may be mostly empty or overwritten, and nothing can be recovered from it. Try an earlier copy, a backup, or the original source, such as the email attachment or cloud version.
No. Your original is never modified. The repaired PDF is saved as a new file ending in "-repaired".