Drop a damaged PDFTap to choose a damaged PDF
or click to browseor tap to choose
Attempt to recover and rebuild it as a clean file
Large files (over 50 MB) may take longer or run out of browser memory
Watch: Repair PDF in Action
Attempts to fix a PDF that won't open correctly or shows as corrupted — by parsing it in a lenient recovery mode that tolerates the kind of broken internal structure that trips up strict PDF readers, then rebuilding it as a clean file.
Upload the PDF that's giving you trouble and download the repaired version. A "What was fixed" line says what was done, for example that the cross-reference table (the file's index) was rebuilt and how many pages were recovered. If the file can't be recovered at all, you'll get a clear message rather than a broken download. If the file isn't a PDF at all, the message says what it looks like (plain text, a picture, a Word file) so you can open it in the right program.
This fixes structural corruption — broken cross-reference tables, malformed objects, the kind of damage that happens from an interrupted save or a bad transfer. A file that was cut off part-way (an interrupted download) is rebuilt from the pages that were saved before the cut; the pages after it were never written and cannot be recovered.
- Can this recover a PDF that's genuinely missing data, not just corrupted?
- No — it fixes structural corruption, like broken cross-reference tables or malformed objects from an interrupted save or a bad file transfer. It can't recover content that was never actually written to the file, like a truncated download.
- What happens if the file can't be repaired at all?
- You'll get a clear message saying so, rather than a broken or incomplete download that looks like it worked but didn't. A few other tools (PDF to Word, PDF to JPG) use different parsing paths and sometimes succeed at pulling out at least some content even when a full structural repair can't — worth trying if the file matters.
- How do I know the repair actually recovered everything?
- Check the reported page count against what you expect, and open the repaired file to confirm the content looks right — recovery is best-effort, not a guarantee. A structurally broken file can sometimes only be partially recovered even when the tool reports success.
- What kinds of PDF damage does this actually fix?
- Broken cross-reference tables, malformed PDF objects, and similar structural issues — the kind of damage that typically comes from an interrupted save, a bad file transfer, or a buggy program that wrote the file incorrectly. It parses the file in a lenient recovery mode that tolerates this kind of damage, then rebuilds a clean, valid PDF from what it can read.
- Are my files uploaded to a server?
- No — the entire repair attempt runs in your browser. Your damaged PDF never leaves your device.
- Will the repaired file look exactly the same as it should have?
- Usually, if the damage was purely structural — the actual page content is unaffected by cross-reference or object-table corruption, so a successful repair typically looks identical to what the file was meant to be. If content itself was genuinely lost (not just the structure around it), that content is gone regardless of how the file is repaired.
- Does a PDF that opens fine in my browser still need repairing?
- Not necessarily — browsers and some PDF readers are more lenient about minor structural issues than others. If a file opens and displays correctly everywhere you've tried, there's likely nothing to fix. Repair matters most when a file fails to open, shows an error, or displays incorrectly in a specific program.
- What should I do once the file is successfully repaired?
- If it needs to be protected before sharing, PDF Password works on the repaired file the same as any other PDF — repair itself doesn't add or remove any password protection.
- What kinds of damage can it fix?
- Structural problems such as a wrong or missing cross-reference index, junk before the header, a missing end marker, or a file cut short by an interrupted download. It rebuilds the document from whatever complete pages are still readable.
- What does Recovered 4 of 9 pages mean?
- The file declared nine pages but only four were still complete, usually because it was truncated. The missing pages were never fully written, so they cannot be restored. The tool flags the file as cut off in that case.
- Will the repaired file match the original exactly?
- Not necessarily. Recovery is best effort, and the page notes ask you to confirm the content is what you expect. Pages with a missing font, image or content stream are left out of a scan-based rebuild.
- Why was my file rejected as not a PDF?
- The first bytes were checked and did not look like a PDF header. The message names what the file seems to be, such as a Word, Excel or PowerPoint file, a JPEG, PNG or GIF picture, or an older Office file.
- What can I do if repair fails?
- You get a message saying no complete page could be found. Other tools read files differently, so the page offers links to PDF to Word and PDF to JPG, which can sometimes still pull out part of the content.
How to use Repair PDF
Drop the damaged PDF and press Attempt Repair; the status reads Attempting recovery while it works. The tool first re-opens the file in a tolerant mode that copes with a wrong or missing index, junk before the header or a missing end marker, and saves a clean copy. If that finds no pages, it falls back to scanning the raw bytes for every complete page object and assembling a fresh document from the pages that survived.
When it finishes, a sentence reports how many pages came back, for example 4 of 9 if the file was cut off, and a preview opens below. Look through it, compare with what you expect, then press the download button; the file is saved with repaired_ before the original name. If nothing can be rescued you get a message instead of a file, with links to PDF to Word and PDF to JPG.
When to use it
Try it when a PDF refuses to open with a complaint about its structure, opens blank or only in part, or arrived through an interrupted download, a truncated email attachment or a crash during saving. It is also worth a go on a file that was fine yesterday and was damaged by a copy, sync or disk hiccup, before you spend time hunting for another copy.
When not to use it
Do not treat it as a guaranteed rescue. If you can fetch the file again from the sender or restore a backup, a clean original beats any rebuilt copy. It cannot create pages that were never written to disk, and a renamed Word, image or ZIP file is named and turned away because that is not damage to a PDF.
Tips for best results
Keep the original untouched and work from the download. Compare the page count with what you expect: when the report reads N of M, the pages after the cut were never saved. Check the first and last pages of the result, plus images and fonts, because a page survives only if every part it needs is present. Expect bookmarks and fillable fields to be missing after the fallback scan.
Limitations
Repair is best effort and never guaranteed. It can mend structure but not content that never reached the disk. In the fallback route a page missing any font, image or content stream is dropped, and the new document is built from pages alone, so bookmarks, form definitions and similar document-level extras are not carried across. When no complete page exists, the result is an error rather than a file.
Privacy
The damaged file is read as raw bytes by your browser and the rebuilt PDF is produced in the same tab. While repairing, nothing is requested from any other server, no helper download happens, and this tool keeps no copy of your file or its name. Closing the tab discards the repaired result along with everything else held in memory.
More from PDF Tools: