Repair PDF Online – Fix a Corrupted PDF File That Will Not Open

Repair a PDF that will not open

Every so often a PDF that worked yesterday refuses to open today. The reader says the file is damaged, shows a blank page, or asks to "repair" it and then fails. If you need to fix a corrupted PDF, the tool on this page rebuilds the document from whatever is still readable inside it and saves a clean copy.

It is aimed at the ordinary causes: a download that was interrupted, an email attachment that was mangled in transit, a file copied off a failing pen drive, or a PDF produced by a faulty exporter. If the file is damaged in the part that tells a reader where everything lives, there is a good chance of a full recovery. If part of the actual page content is gone, you will get back less than you hoped, and this article explains how to tell the difference.

How to recover a damaged PDF, step by step

1. 1
   Drop the damaged file onto the box that says "Drop files here, or click to browse", or click the box and choose it. You can add more than one file.
2. 2
   Look at the message that appears under the box. For a damaged file it says that this is exactly what Repair PDF is for. If it says the file is not a PDF inside, or is empty, the tool cannot help, and the button stays disabled.
3. 3
   Press Repair PDF.
4. 4
   When "Done" appears, press Download. The repaired copy is called repaired.pdf, or the original name followed by "-repaired.pdf" if you added several files.
5. 5
   Open it and check the page count and the pages that matter. The tool does not report how many pages it recovered, so you are the final check.

What is actually inside a PDF, and what breaks

A PDF is a collection of numbered objects: pages, fonts, images, text streams and a catalogue that says which page comes first. At the end of the file sits a cross-reference table, often called the xref, and a short trailer. Together they work like a contents page, giving a reader the position of every object so it can jump straight to page 40 without reading the 39 before it.

Most damage happens at that end of the file. A download that stops a little early, a program that crashes while saving, or a transfer that adds stray bytes will often leave all the objects intact but the contents page missing or pointing to the wrong places. Strict readers give up at that point. The objects are still there, though, and a more patient program can find them by reading the file from the top.

How the repair works

The tool uses pdf-lib, an open-source PDF library, running in your browser. Instead of trusting the cross-reference table, pdf-lib reads the file object by object from the beginning. This version is told to tolerate objects it cannot parse and to skip them rather than stop. Once the document is in memory with a valid page tree, the tool writes it back out as a new file with a fresh cross-reference table and trailer.

That is all it does. It does not re-draw pages, run OCR, or reconstruct missing images. It rebuilds structure. This is why it works well on index damage and not at all on a file whose real content never arrived.

A worked example from a test file

To see what this approach recovers, a 10-page test document of 5,116 bytes was damaged in several ways and loaded with the same settings the tool uses.

With the cross-reference table and trailer deleted entirely, all 10 pages came back. With the start-of-index pointer set to a wrong number, all 10 came back. With the end of the file cut off just after the trailer began, all 10 came back. When a short stretch of about 20 bytes in the middle of the file was overwritten, 9 pages came back, because the damage landed on a page object that could not be read, so that page was skipped. And when the file was cut off partway through its objects, whether at the halfway point or just before the index, loading failed and nothing could be recovered.

The pattern is the useful part. Damage to the index is cheap to fix. Damage to a particular object costs you what that object held. Missing data cannot be repaired by anyone, because it is not in the file.

Reading the outcome

There are three outcomes. A new file downloads and opens normally: the structure was the problem and it is fixed. A new file downloads but has fewer pages or a blank or garbled page: the structure was rebuilt, and the objects that were damaged have been dropped. Or you see an error instead of a download.

The error has two versions. One says the file is encrypted rather than damaged, and that the contents are intact but locked. That file needs to be unlocked with its password, or Remove restrictions if it opens freely but blocks printing. The other says the file is damaged past rebuilding, meaning not enough structure is left to work with. Your best move then is to find another copy: the original download link, an email attachment, a cloud backup or a colleague's version.

Common reasons a PDF stops opening

• An incomplete download, usually after a dropped mobile connection. The file size is smaller than it should be.
• A WhatsApp, email or Drive transfer that was interrupted partway.
• A pen drive or memory card with bad sectors, or one that was pulled out before the copy finished.
• A scanner or app that crashed while writing the file, leaving a zero-padded or half-finished document.
• A file edited with two programs in sequence where the second one wrote the structure wrongly.
• A file that was renamed to .pdf from another format. Repair cannot turn a Word file or a photo into a PDF; the tool tells you the file is not a PDF inside.

Things to try before and after repairing

Before repairing, open the file in a different program, such as a web browser or a phone viewer. Some readers are stricter than others, and a file that Acrobat rejects may open elsewhere. Check the size against the original if you know it. A 4 MB file that shows 600 KB was cut short.

After repairing, check the pages. If one is missing and you only need the rest, that is a working result. If the repaired file is large, shrink it with the compressor so it is easier to send. If a few pages are unusable, delete the broken pages and keep the remainder as a tidy document. If only part of the file matters, extract the good pages into a new PDF.

If the repaired file opens but looks wrong, a picture-based workaround is available. Open it in any viewer that shows the pages and export them as JPG, then rebuild a clean document from the images.

Repairing several files in one go

The tool accepts more than one PDF at once, which is handy after a bad copy operation that left a whole folder of suspect files. Each file is rebuilt in turn, with its own result. Results are named after the originals, and when there is more than one you get a button to download them together as a ZIP.

Because a file that cannot be rebuilt stops the whole run with an error, rather than being skipped, a long queue is best split into small batches if you suspect that several are beyond saving. Repair the likely candidates first, then deal with the stubborn ones on their own with fresh copies.

Why your viewer's own repair prompt often is not enough

Many readers offer to repair a file when they open it, but they do so in memory and show you the pages without ever writing a fixed file. The next program you open it in hits the same fault. A repaired copy written to disk fixes the problem for every program, and it is the version to attach to an email or upload to a portal that checks the file itself, not your screen.

That matters most for portals. A form that parses your PDF on its own servers cannot rely on a viewer's forgiving behaviour, so a file that merely looks fine on your laptop can still be rejected at the upload counter. Repairing it first removes that uncertainty.

Mistakes that make recovery harder

• Saving over the damaged file. Keep the original untouched, because a different tool may do better with it than this one.
• Opening and re-saving the file in an editor first. A second write can destroy the very objects a repair would have found.
• Expecting repair to fix a password problem. A locked file is not broken, and rebuilding it only makes it worse.
• Assuming a successful download means a complete file. Always check the page count against what you expected.
• Repairing a copy that was truncated and then being surprised by the error. If the end of the file is simply missing, get the file again.

Limits and honest limitations

Each file can be up to 100 MB and 500 pages, because the work happens in your browser's memory. The tool recovers structure only. It cannot restore text, images or pages whose data is missing or overwritten, it does not tell you how many pages it dropped, and it does not repair a file that is cut off in the middle of its content.

A repaired file is also a rewrite, not a forensic copy. It is meant to open and print correctly. If you need evidence-grade preservation, work from the original bytes and a specialist tool.

Privacy: the file never leaves your browser

The repair runs entirely in your browser tab. Your file is read from your device, rebuilt in memory and handed back as a download. It is not uploaded to a Jobsixo server, which is the sensible way to handle an exam result, a contract or an identity document that happens to be damaged.

Related tools

Once the file opens, you may want to reorder its pages, add the repaired pages to another document, or save it in an archival format. Other tools throughout the site point here whenever they meet a damaged file, so you can always come back, repair, and retry.