PDF Tools

Almost every PDF tool on the web works the same way: you upload your file to a stranger's server, wait while it is processed there, and download the result — a workflow that asks you to hand contracts, statements, scans of identity documents and unpublished manuscripts to a third party for the privilege of moving pages around. The tools in this category invert that model completely. Modern browsers can read, rebuild and render PDFs themselves, so every operation here runs on your own device: a file is opened in your browser's memory, worked on there, and offered back as a download from your own machine. Nothing is uploaded because nothing needs to be — and that is not a policy promise that could quietly change in terms of service, it is the architecture of the page, verifiable by loading a tool, disconnecting from the network, and using it exactly as before.

Under the hood the eleven tools run on two different engines, and the difference decides what each one can honestly promise. The structural tools — merge, split, rotate, delete pages, reorder, and the two stamping tools that add page numbers or a watermark — copy pages as objects from the original into a new document without re-rendering them, so text stays selectable, vectors stay vector, page sizes and rotations are preserved, and there is no quality change of any kind. The render tools — PDF to Images, PDF to Text, and Compress PDF — instead go through the page renderer: pages become pixels at a chosen DPI, or have their stored text layer read out. The consequence is sharpest for Compress, which rebuilds each page as a JPEG image: that process is lossy, it removes the text layer, and a PDF born as vector text usually comes out larger, not smaller, because its pages have been replaced by photographs of themselves. That warning is printed on the Compress page itself, with recorded figures.

Between them the eleven tools cover the working life of a PDF. Assembly: merge several files into one, build a PDF from images, and reorder, delete or split pages until the document says what it should in the order it should. Correction and marking: rotate sideways scans back upright by page range, stamp page numbers in a choice of formats and positions, and watermark every page, singly or tiled. Conversion and extraction: render pages to PNG or JPEG at 150, 200 or 300 DPI with the pixel dimensions stated in advance, pull the text layer out page by page — with the plain caveat that a scanned PDF has no text layer and no OCR is performed here — and compress image-heavy scans and photo PDFs down to a fraction of their size, with exact before and after byte counts on every result.