Repair PDF - Free Online Browser Tool

Fix a PDF that will not open by rebuilding its cross-reference index, and say plainly when it cannot be saved.

A PDF that will not open is usually damaged in one of a handful of specific ways, and most of them leave the actual page content intact. That is what makes repair possible rather than hopeless: the file is indexed by a table at the end, and a great many files open perfectly once that table is rebuilt from what is actually there. This tool looks at the damage before it tries anything. It reports what is wrong - a cross-reference table pointing at the wrong offset, a missing or unparseable startxref, an object declared in the table that is not in the file, a page tree that does not reach every page - and then it rebuilds what it can. Objects are recovered by scanning the file itself for real object headers, so pages, text and images that were never reachable from the broken table are picked up rather than lost. What comes out is one fresh PDF with a new, correct index - the original file is never touched, so you keep the damaged copy until you have opened the repaired one and seen every page you expect. The honest limit: this repairs structure, not content. A file truncated in the middle of a stream still has a truncated stream at the end, and the tool will say how many objects could not be recovered rather than quietly dropping them. Encrypted files are reported as encrypted, not cracked - that is a different job and pretending otherwise would be worse than refusing. The whole repair happens on your own machine. A file that will not open may well be one you would rather not send anywhere, which is exactly when a repair tool earns its keep.

This repair pdf utility runs completely inside your web browser. Your document is opened locally and processed on your own device using JavaScript and WebAssembly, so your file is never uploaded to any server. There is no signup, no watermark and no limit on how many documents you can handle.

  • 100% client-side - files never leave your device
  • Instant processing, no upload wait
  • Works on desktop, tablet and mobile
  • Free forever, no account required

Common next steps: Merge PDFs, Split PDF, Count Pages & Words, Extract Pages.

How to Use the Repair PDF Tool

Five steps, and the whole thing runs locally on your own device.

  1. Drop in the file that will not open

    The tool reads it as bytes rather than trusting its own parser, so a file broken enough to crash a normal reader still reaches the repair path.

  2. Read the diagnosis

    Before changing anything the summary names what was found: where startxref points and whether it lands on a cross-reference table at all, how many objects the file declares against how many are really in it, whether a catalog can be found, and how much data sits past the end of the file.

  3. Rebuild the index

    Objects are recovered by scanning the file for genuine object headers and the cross-reference table is written afresh. No guesswork about what a broken offset was meant to be.

  4. See exactly what it did

    One repaired file comes back, and the summary records what was discarded, how many objects were reindexed, and every finding - so you can judge the repair before you replace the original with it.

  5. Save, and open the result before you trust it

    Open the repaired file and check the page count against what you remember before you overwrite anything. A repair that still will not open is reported as a refusal with the reason, never handed over as a success.

Frequently Asked Questions

Which kinds of damage does this actually fix?

The ones that leave the content intact: a wrong or unparseable cross-reference offset, a missing startxref line, a table that lists objects which are not in the file, a page tree that stops short of some pages, and objects that exist in the file but were never reachable from the broken index. Those are all index problems, and all of them are recoverable because the page data itself has not been touched.

Can it recover a file that was cut off mid-download?

Not the part that is missing. If a stream was truncated at 40 per cent, those 40 per cent may decode and the rest is simply not there. The tool recovers what exists, tells you how many objects could not be recovered, and will not claim a clean repair when there is nothing left to recover from. A partial file is a partial rescue.

Will it open an encrypted or password-protected PDF?

No, and it does not try to get around the password. An encrypted file is reported as encrypted with the encryption details found in it, because removing that is a different job with a different purpose. If the file is one you are entitled to open, open it with its password, save an unprotected copy, and bring that.

Is the original file changed?

No. Nothing is written back to what you dropped in - the repaired PDF and the loose pages are new files produced from what was read. Keep the damaged original until you have opened the result and confirmed the pages are all there.

What if the repaired file opens but a page is blank?

That page's content stream was damaged rather than its index, and a blank page in the output means the object was found but its bytes would not decode. The report distinguishes this from a page that was never recovered. Objects of this kind cannot be reconstructed from a damaged file - they have to come from the original application, so keeping the source document matters.