Skip to content

Compare PDF Files — Find Text and Visual Differences, In Your Browser

Compare two PDFs word by word and pixel by pixel: every added, removed and changed word marked, page by page, without uploading either file.

Found a problem, or something that could be better? We read every message and we fix things quickly.

This free tool compares two PDF files and shows exactly what changed: words added, removed and replaced, page by page, in your browser. Add the original and the revised PDF, and the text of every page is extracted, the pages of the two versions are matched up (an inserted or deleted page is recognised rather than shifting everything after it), and each pair is compared word by word. Removed words are struck through and marked [− …], added words are underlined and marked [+ …], side by side or inline, so the differences are readable without relying on colour. A visual mode renders the pages and highlights pixel differences, for changes the text cannot show: a moved logo, a new signature, a different chart.

It is honest about what a PDF comparison can and cannot do. A scanned PDF contains pictures of pages, not text, so the text comparison finds nothing in it; use the visual mode, or turn the scan into text first with the image to text (OCR) tool. Text that re-wraps onto the next line is handled, but text pushed onto another page appears as a removal on one page and an addition on the next. Both files stay on your device, which matters for the contracts, quotes and policies this is usually used for.

How to use

  1. Choose the original PDF as A and the changed PDF as B (or press Try two sample PDFs to see how it works).
  2. Read the summary: pages in each file, pages with changes, words added and removed.
  3. Step through the pages with the page menu or the Next change button. Switch between side-by-side and inline views; open Options to ignore case or punctuation, or to pair pages strictly by number.
  4. For layout, images or scanned files, switch to Visual differences, set the sensitivity threshold, and press Find changed pages to list every page that looks different.
  5. Download the report as HTML (for reading or printing) or plain text (wdiff style, [-removed-] {+added+}).

Frequently asked questions

How do I compare two PDF files for differences?

Add the older version as A and the newer as B. The tool extracts the text of each page, matches the pages of the two files, and marks every added, removed or replaced word. For differences that are not text, such as images, layout or a scanned page, use the visual mode, which compares the rendered pages pixel by pixel and boxes the regions that changed.

Are my PDFs uploaded?

No. Both files are opened with pdf.js, the PDF engine in Firefox, inside your browser tab, and all comparing happens there. Nothing is sent to a server, so confidential contracts and drafts never leave your device. The page keeps working offline once loaded.

Why does it find no text in my PDF?

The PDF is probably a scan: each page is an image of paper, with no text layer. Text comparison needs text. Use the Visual differences mode, which works on any PDF, or convert the scan to text with an OCR tool first. A PDF made by "Print to PDF" from a program, or exported from Word, has real text and compares fine.

Why does a small edit show up as many changes?

Usually because text moved between pages. If an edit pushes a paragraph from page 3 to page 4, the comparison sees it disappear from one page and appear on the next. Line re-wrapping within a page does not cause this, because words are compared, not lines; and words hyphenated at line ends are joined before comparing. Changes in tables or columns can also look larger, because a PDF stores them as positioned fragments with no reading order.

What happens if the PDFs have different page counts?

The pages are aligned by content. A page inserted in the new version is shown as "only in B (added)", a deleted page as "only in A (removed)", and the pages around it are still compared with their real counterparts. If you would rather compare page 1 with 1, 2 with 2 regardless, tick "Pair pages by number" in Options.

Can I save the comparison?

Yes. Download the report as an HTML file, which opens in any browser and prints cleanly with removed text struck through and added text underlined, or as a plain-text file using the [-removed-] {+added+} notation of GNU wdiff. If you ran the visual scan, each page also lists the share of pixels that differ.

Text comparison versus visual comparison

The two modes answer different questions. Text comparison asks "did the words change?" and gives an exact list: which word, on which page. It ignores fonts, colours, spacing and images, which is what you want when checking a contract or a policy for edited clauses. Visual comparison asks "does the page look different?" and catches what the text cannot: a swapped logo, a changed figure in a chart, a signature, a stamp, a shifted table, or any difference at all in a scanned PDF.

Use both when it matters. Run the text comparison to read the edits, then the visual scan to make sure nothing changed outside the text. A visual difference on a page with no text changes is worth a close look.

How the comparison works

• Text extraction: pdf.js returns each page's text as positioned fragments; fragments on the same baseline become a line, ordered left to right.

• Page alignment: pages are matched by the words they share (a dynamic-programming alignment, the same idea as aligning DNA sequences), so one inserted page is recognised as an insertion.

• Word diff: each pair of pages is compared word by word with the Myers diff algorithm (the one behind git diff), after optional case and punctuation normalisation.

• Visual diff: both pages are rendered at the same scale; a pixel counts as changed when any colour channel differs by more than the threshold, and changed pixels are grouped into numbered regions.

Limits worth knowing

Multi-column layouts and tables are compared in the order the PDF stores their text, which is usually but not always reading order. Two PDFs produced by different programs from the same document can differ in spacing, ligatures (fi as one character) or line breaks, and show changes where a reader sees none; ignoring punctuation helps with some of these. Very large pages that are almost entirely different are shown as replaced in full rather than diffed, so the browser stays responsive. Files are limited to 50 MB each.