Best-effort salvage for damaged PDFs: the reader engine rebuilds what it can, and every recoverable page's text is extracted for download.
PDFs break in characteristic ways: truncated downloads lose the cross-reference table at the end, email systems mangle bytes, drives corrupt quietly. Recovery here works because the pdf.js engine, the same reader built into Firefox, rebuilds a damaged document's structure from the surviving objects rather than trusting the broken index, and many 'corrupt' files are structurally fine except for that index. This tool loads the file with error tolerance, then extracts the text of every page that renders, and reports honestly which pages came back and which did not.
The honest boundary: this salvages content, primarily text. It does not restore a bit-perfect original, formatting, images and interactive elements of recovered pages are not rebuilt into a new PDF, because a 'repaired' file that silently drops content would be a worse outcome than an honest text extraction. Scanned-image PDFs contain no text layer to salvage (run recovered pages through the OCR tool separately if the file opens there). Encrypted PDFs need their password first, the Unlock PDF tool handles those you own.
The tool recovered 7 of 10 pages, where are the other 3? Those pages' content streams were damaged beyond reconstruction. The report names the page numbers so you know exactly what is missing rather than guessing.
My PDF opens but shows errors, same thing? Usually yes: viewers paper over damage on screen. If the content matters, extract it now; files that throw errors tend to get worse, not better.
Can it recover a deleted PDF? No, deleted-file recovery is a disk-level job for dedicated undelete software, done as soon as possible after deletion. This tool works on a file you still have, however damaged.