PDF Compare checks what changed between two PDFs on the visible page: it renders file A and file B page by page at the same DPI and compares them pixel by pixel, reporting how many pixels differ and what share of the page that is. It is especially useful for checking layout differences before and after a revision, before and after a stamp, or after a re-export.

The result can be viewed three ways: “Side by side”, “Overlay” and “Difference highlight”. In the difference image, red marks areas that A has and B does not, or that were changed, while blue marks what is new in B. Comparison, preview and diff generation all happen locally in the browser.

1. 01Prepare both files— add at least two readable PDFs at the top of the PDF workbench and choose them as “File A (reference)” and “File B (comparison)”.
2. 02Set the precision— choose 72, 96 or 144 DPI; raise “Difference sensitivity” when you need to catch subtler changes, keeping in mind that a high DPI costs memory and time.
3. 03Start the comparison— click “Start comparison”; the tool compares page by page and shows each page's differing pixel count and ratio in a results table; extra pages when the two counts differ are counted as wholly different.
4. 04Review and export— click “View comparison” to switch between side by side, overlay and difference highlight, or to show only the pages that differ; to keep a record, download a single diff image or click “Download diff images as ZIP”.

What is compared

the visible pixels rendered at the chosen DPI — not PDF object structure, metadata, hidden text or certificates, so machine-level differences stay invisible

Page alignment

page 1 against page 1, page 2 against page 2; inserting or deleting a page keeps the rest compared by position, and the tool never searches for the matching section

Precision and resources

72 / 96 / 144 DPI are supported; beyond 120 pages it warns about memory and time, though that is not a hard page ceiling

Anti-aliasing and fonts

a different generator, font rendering or a tiny layout shift can all produce pixel differences, and the higher the sensitivity the more of them get flagged

What it does not do

it does not track semantic edits to text, does not produce a Word-style record of insertions and deletions, and does not judge which version is correct; to pull text out of a file, use “PDF Extract Text”

Reading both PDFs, rendering them at the same DPI, comparing pixels and generating the diff images all take place locally in the browser; the file contents are never uploaded to a server.