A PDF is a list of numbered objects plus a table at the end saying where each one is. When that table is wrong or missing, readers refuse the file even though the objects are intact. The tool reads every object it can find, discards references that lead nowhere, and writes a fresh file with a correct table.
It cannot invent pages that were never downloaded, decrypt a file without its password, or restore an image whose bytes are garbage. If the repair cannot produce a valid file, it says so rather than handing you an empty one.
An email attachment that was cut off mid-transfer. A scan from a ten-year-old app that never quite followed the spec. A file saved over itself forty times until the stale references piled up. Acrobat refuses all three with the same message, and all three are usually a few seconds from opening again. The repair is worth trying before the redownload, and essential when the redownload is not an option.
A thirty-second check before you start
Some "damaged PDFs" are not PDFs at all. A login page saved with a .pdf name, a download that stopped at byte zero, an email gateway's placeholder. The check below reads the first and last few kilobytes of the file on your machine and tells you which kind of problem you have.
- Header
%PDF-1.xin the first bytesConfirms it is a PDF <html>where the header should beAn error page, not a documentstartxrefand%%EOFat the endThe trailer survived- Header present, trailer missingInterrupted download: repairable
- A file of a few hundred bytesNothing to repair; redownload
How the repair works
Add the broken PDF
Drop it or browse. The parser accepts files most readers reject. It is read into browser memory; nothing is uploaded.
Let it scan
Every object the parser can find is read. References that lead nowhere are dropped, and the cross-reference table is reconstructed from scratch.
Download the rebuilt file
A fresh, structurally valid PDF built from whatever the source still had. Open it and page through to see what came back.
Pages that survive are byte-identical. The repair rewrites the index, not the pages.
Including files Acrobat, Preview and Chrome refuse to open.
The scan and rebuild run in your browser.
What breaks in a PDF
Every PDF ends with a cross-reference table listing the byte offset of every object, followed by a trailer that points at the table. Readers start from the end. If the offsets are wrong, the table is missing, or the trailer is cut off, a strict reader stops before it has looked at a single page. The pages are still there; the map to them is not.
- Interrupted download or email transferTrailer cut off
- File saved over itself many timesStale, conflicting references
- Old scanner or export softwareOut-of-spec structure
- FTP in text mode, strict gatewaysBytes altered in transit
- Truncated by a full diskEnd of file missing
What it can and cannot recover
Structure is wrong, content is present
- Missing or incorrect cross-reference table
- Truncated trailer after a cut-off download
- Malformed object streams from old software
- Stale references from repeated re-saves
- Files that strict readers refuse but lenient ones open
Content itself is gone
- The last N pages of a download that never arrived
- An encrypted file whose password is lost: repair is not decryption
- Image streams whose bytes are corrupt (the rest of the file survives)
- A file whose first bytes are wrong and contains no findable objects
- A file that was never a PDF
A repair tool is not a data-recovery service. The honest outcome is one of three: a complete rebuild, a partial rebuild with the pages that survived and a note about the rest, or a clear failure. The tool never writes out an empty document and calls it success.
Before you run it
- Keep the broken original. The repair writes a new file and does not modify the source, but a copy costs nothing and protects you if something unexpected happens.
- Try the redownload first, if you can. A clean copy from the source beats any rebuild. Repair is for when the sender is gone, the link has expired, or the file was born broken.
- Run the check above. An HTML error page renamed .pdf cannot be repaired, and knowing that in thirty seconds beats wondering.
- Try a lenient reader. If Chrome or Preview opens the file, printing it to a new PDF from there is an alternative rebuild.
Why it runs in your browser
A file that will not open is still a file full of content: a contract, a statement, a scan of something private. Uploading it to a repair service hands that content to a server in exchange for a structural rewrite that a browser performs in a second.
The tool is right above
Everything on this page runs in your browser. Scroll up, drop the file, and keep it on your machine.
Common questions
What people ask with a refused file in front of them.
01What kinds of damage can this fix?
Missing or wrong cross-reference tables, truncated trailers, and malformed object streams. Any damage that leaves most of the page content in the file can be rebuilt around.
02What can it not fix?
Pages that were never downloaded, files encrypted with a lost password, image streams whose bytes are corrupt, and files that are not PDFs at all. If the content is gone, no tool invents it.
03Will the repair change how the file looks?
No. Pages that survive are copied unchanged. The repair rewrites the index, not the content.
04How do I know if my file is really a PDF?
Use the check on this page, or open the file in a text editor: the first characters should be %PDF. If they are HTML, the file is a saved web page and needs redownloading.
05Why is my PDF broken in the first place?
Usually an interrupted download, an old application that wrote an out-of-spec file, or repeated saves that left stale references. Occasionally a transfer that altered bytes.
06Should I repair or redownload?
Redownload if you can; a clean copy from the source is always better. Repair when the source is gone, the link has expired, or the file was broken from the start.
07Are my files uploaded anywhere?
No. The scan and rebuild run in your browser.
08Is it free?
Free. Every tool, unlimited use, no signup.