PaperKit

Compare PDFs

Load two versions of a document, flip through pages side by side, and toggle a difference overlay that highlights changed areas in red.

Original PDF

Revised PDF

Page /

% of pixels differ on this page

Comparing loads two versions of a document side by side, page by page, with an overlay that highlights the areas that differ in red. It answers the question that comes up whenever a contract, a specification or a policy comes back from someone else: what actually changed? Reading two long documents in parallel is slow and unreliable — the human eye is very good at seeing what it expects — while a visual diff surfaces a changed figure or an inserted clause immediately. The comparison is visual rather than semantic: it renders both pages and looks for regions that do not match, which catches everything from a reworded sentence to a moved table, but flags reflowed text as changed even when the words are identical. Both files are processed in your browser, so two versions of a confidential agreement are never uploaded.

How the difference is computed

Both pages are rendered to bitmaps at the same scale and compared region by region; areas that do not match are tinted. This is deliberately dumb and therefore complete — it cannot miss a change, because any visible change alters the pixels.

The trade is that it does not understand what changed. A paragraph inserted on page 3 pushes everything after it down, and every subsequent page then reads as almost entirely different, even though the text is unchanged. That is the main thing to be ready for when comparing documents that have grown.

Reading the result

Work page by page and treat the highlights as candidates rather than conclusions. Small isolated marks are usually the real edits: a changed number, a new clause, a different date, a swapped name.

Large uniform highlighting across a page usually means reflow rather than rewriting. When you see it, find the first genuine change above it — that is the insertion or deletion that shifted everything, and the pages after it are often unchanged in substance.

What this is good and bad at

It is good at documents from the same source with the same layout: two exports of a contract, two revisions of a drawing, a policy before and after amendment. It catches everything, including changes to images, tables and figures that a text diff would miss entirely.

It is poor at documents that were regenerated with different fonts or margins, and at comparing a scan to a digital original, where nothing lines up and everything highlights. For those, extracting the text from both and comparing that is the more useful approach.

Frequently asked questions

Are my PDF files uploaded to a server?

No — this tool runs entirely in your browser using WebAssembly and JavaScript. Your files never leave your device, which also means there are no size limits and no queues.

How does the difference highlighting work?

Both pages are rendered at the same scale and compared pixel by pixel; areas that differ are marked red and a change percentage is shown. It catches layout, image and text changes alike.

Why is a whole page highlighted when I only changed one line?

Because the comparison is visual. An insertion pushes the following text down, so every line after it sits in a different place and reads as changed. Look for the first real difference — the pages below it are often unchanged in substance.

Can it compare a scan against the digital original?

Poorly. Scans differ in alignment, contrast and scale, so almost everything highlights. Extract the text from both and compare that instead when the two versions did not come from the same source.

More PDF tools

Advertisement