Why a PDF stops opening
A PDF is not read from the front. It ends with a table saying where everything in it lives, and a reader starts at the back, looks up what it needs and jumps straight to it. That is why a four hundred page document opens instantly on page one.
It also means the last few kilobytes matter more than all the rest. Lose bytes to a transfer that stopped, a failing drive, a full disk or a crash while saving, and that table points at the wrong places. The reader jumps, finds nothing where it expected something, and refuses the file.
The document is usually still in there. It is the index that broke.
How this tool works
It walks the file from the start, finds every object actually present, and writes a fresh index describing what it found. The pages themselves are untouched: text stays text, drawings stay vector, nothing is rasterised or re-encoded.
The engine is qpdf, which is the reference tool for this sort of thing, compiled to run in the browser. It downloads once when you use the tool and runs on your machine.
It counts the pages before it claims anything
This is the part that matters, and it is where a repair tool can do real harm. A rebuild always produces a file. It will happily write out a valid document containing two of the forty pages that used to be there, and a tool that puts "repaired" over that has told somebody their dissertation is fine.
So nothing is claimed until the result has been opened and counted, and the page count is the headline. If pages are missing, it says how many and says plainly that they were not in the file to recover.
What cannot be recovered
Bytes that are gone. If a download stopped at 40%, the other 60% of the document does not exist anywhere in your copy. Nothing recovers it, here or anywhere, and any site claiming to is selling something.
Damage inside the content. If the page descriptions themselves are corrupted rather than the index, a rebuild produces a file that opens onto damaged pages. That is still better than nothing, and it is not a repair.
A file that was never a PDF. Sometimes the download saved an error page. If the file is a few kilobytes and starts with HTML, there was never a document there.
Try the original again first
Worth ruling out before assuming damage. Download it again if you can, since a truncated transfer is the most common cause by a distance. Try a different reader, because they vary enormously in how much they tolerate. And check the file size against where it came from: a document that should be two megabytes and is forty kilobytes did not arrive.
You cannot check what is in a file you cannot open
The file is rebuilt on this page using your own machine. Nothing is uploaded, which matters here in a way people do not always think about: a damaged file is one you cannot open to check what is in it, so uploading it means sending a stranger a document whose contents you no longer know.
Common questions
How do I fix a PDF that will not open?
Drop it in and press rebuild. Most files that refuse to open have an intact document with a broken index at the end, and that index can be rebuilt from what is actually in the file. Nothing is uploaded.
Can it recover a file that was cut off halfway?
Only the part that is still there. If a transfer stopped halfway, the second half of the document does not exist in the file and no tool can put it back. This says how many pages came out so you can see what survived.
Why does my PDF open in one program but not another?
Readers differ in how much damage they tolerate. Some rebuild a broken index silently, others refuse outright. Rebuilding the file often makes the strict one accept it.
Will it change how the document looks?
No. The pages are kept exactly as they are: text stays text, drawings stay vector. Only the structure that says where everything lives is rewritten.
What if it does not work?
Then the damage is in the content rather than the index, and the honest answer is that it is not recoverable here. The tool tells you what came out rather than claiming success on a file that lost most of itself.