"The file is damaged and could not be repaired" tells you nothing about what went wrong. In practice a PDF that will not open is usually missing one of a handful of landmarks: the %PDF header a viewer checks first, the %%EOF marker at the end, the startxref pointer that says where the object table lives, or the trailer that names the document root. This tool reports each one individually, says whether the parser could still build a page tree, and then tries two repairs in order: copying the object graph into a fresh document, and — if that fails — redrawing every page the renderer can still paint.
Features
Checks the header, trailer, cross-reference pointer and end-of-file marker
Reports a startxref that points past the end of a truncated file
Detects encryption, which has to be cleared before objects can be rewritten
Trims junk prepended by mail gateways or appended after the last %%EOF
Rebuilds from the object graph first, keeping text, fonts and vector artwork
Falls back to redrawing each page when the object graph cannot be copied
Tells you which of the two repairs was used, and what it cost
How to use the Repair PDF
Drop the PDF that will not open
Read the structure check to see which landmark is missing
Press Repair and let it try the object rebuild first
Download the repaired file and open it to confirm
Example
Input
invoice-scan.pdf — downloaded over a dropped connection
Output
End-of-file marker missing; startxref points to byte 481,209 in a 402,880-byte file. Rebuilt: 6 pages recovered.
A cross-reference offset past the end of the file is the signature of a truncated download.
Common errors & troubleshooting
The repair produced a document with fewer pages than expected. — Only the pages whose objects survived can be recovered. A truncated file loses whatever came after the cut, and there is nothing left to rebuild from.
Repair reports the file is encrypted. — Objects cannot be rewritten without reading them. Remove the password first, then repair.
The repaired file lost its selectable text. — That means the object rebuild failed and the pages were redrawn instead. Recognise the text afterwards to make it searchable again.
Nothing could be recovered at all. — If not a single page renders, the damage reaches the page content itself. Re-download the original or restore it from a backup — no reconstruction can invent the missing bytes.
Frequently asked questions
What usually breaks a PDF?
An interrupted download or copy, which truncates the file, and email gateways that prepend or append text. Both are visible in the structure check as a missing marker or a bad offset.
Is my broken file uploaded to be repaired?
No. The bytes are inspected and the rebuild happens in this tab, so a damaged file containing sensitive material stays with you.
What does "fast web view" in the report mean?
The file was arranged so a browser can show page one before downloading the rest. Rebuilding drops that arrangement, which affects loading speed and nothing else.
Will the repaired file be identical to the original?
The pages will be. The object layout is rewritten, so the byte size and any linearisation change — that is what repair means.
Can this recover a PDF I deleted?
No. This works on a file you still have; recovering a deleted file is a job for disk recovery software.
Related tools
PDF Metadata Editor — View, edit or strip a PDF’s title, author, subject, keywords and producer.
Decrypt PDF — Remove password protection from an encrypted PDF.
Compress PDF — Shrink a PDF by re-rendering each page to a JPEG at a chosen quality and resolution.
Merge PDFs — Combine multiple PDFs into one, with reordering.
PDF to Images — Render PDF pages to PNG or JPG images you can download.
Flatten PDF — Lock filled form fields into the page so a PDF looks the same everywhere.
PDF OCR — Recognise the text in a scanned PDF and make it searchable.