What the PDF tools cover
The PDF tools handle the routine jobs that normally send people to an upload site: merge several files into one, split one into several, compress a file that is too large to email, rotate, delete or extract pages, password-protect a document, add a watermark, convert JPG or PNG images into a PDF, and read, rewrite or strip the title, author and subject metadata. The newer tools go further: unlock a file whose password you know, reorder pages visually, convert pages to JPG or PNG, extract the text, pull out the embedded images, add page numbers, place a signature or logo, crop margins, flatten forms and resize every page to A4 or Letter. Three more cover the jobs that usually sit behind a subscription: sign a document, redact words so the text is genuinely removed rather than covered over and compare two versions page by page.
The file is read into your browser's memory and processed there, using local copies of pdf.js and pdf-lib. It is never uploaded, which is the right design for bank statements, contracts and anything else you would not hand to a stranger. Compression tries a lossless re-save, then recompression of embedded images, then full-page rasterisation, and keeps whichever result is smallest. It never returns a file larger than the one you gave it. Protection uses real AES encryption through qpdf compiled to WebAssembly, not a watermark pretending to be a password.
Limits
- Password-protected input files must be unlocked before any other tool can read them.
- Very large documents are bounded by your device's memory. A 300-page scan will be slower on a phone than on a laptop.
- Rasterised compression turns text into an image. The page tells you which strategy won so you know whether the text is still selectable.