Compress PDF to Size

Every page is rendered by pdf.js to a canvas, encoded as a JPEG and reassembled into a new PDF by pdf-lib, and each attempt is weighed in bytes, so the number you are shown is the size of the file you download. The quality setting is binary-searched: a high-quality build first, then a low one to prove the target is reachable at all, then midpoints until the bracket closes to 0.07. If even the lowest quality overshoots, the pages are re-rendered at 55% of the scale and the search runs again. Six builds is the ceiling.

There is a floor. A sixty-page document cannot become 100 KB without turning every page into an unreadable smear, so the search stops rather than pretend. You are told the smallest size it managed and offered that file anyway. Raise the target, or split the document and compress each part on its own.

No. pdf.js renders the pages, the canvas API re-encodes them and pdf-lib writes the new file, all inside this tab, no request carries the PDF. Size limits usually come from portals asking for a passport or a signed contract, which is exactly the kind of file that has no business passing through someone else's server.

No, an encrypted PDF is refused before any work starts, so unlock it first and come back.