Compare Two PDFs and See Every Changed Word
or drop files here
Put the version you sent next to the version that came back. Struck words get a red box, new ones a green box, and both drafts are read inside this tab instead of on a stranger's server.
Two versions in, every edit out.
Drop the two versions
Original goes in first, revision second. The comparison starts as soon as both are in, page against matching page.
Read the red and green
Amber chips mark edited pages, green chips new pages, red chips missing ones. Open any page to see the boxed words and a plain list of what became what.
Keep a marked-up copy
Save the revised PDF with every new or reworded word under a green highlighter stroke, ready to send back with your questions.
Whenever a draft comes back changed.
Reading two long documents side by side with a ruler is slow and unreliable. Letting the words line themselves up takes seconds and does not get tired on page forty.
A contract that came back "clean"
The other side says they only changed the payment terms. Compare their copy with the one you sent and see whether a liability cap or notice period quietly moved as well.
Lease renewals
Landlords often reissue last year's lease with a new date. A word-level check shows whether rent, deposit, pet or break clauses were edited before you sign it again.
Printer proofs against the final file
Before approving a brochure or book proof, compare it with the file you supplied. Typesetting slips such as a dropped line or a doubled word stand out immediately.
Updated terms and policies
When a supplier, insurer or software vendor sends "revised terms", put the old PDF beside the new one and read only the sentences that actually changed.
Revised quotes and tenders
A second quote that looks identical can hide a new exclusion or a changed quantity. Red and green boxes on the figures make the difference hard to miss.
Paper and thesis revisions
Show a supervisor or journal reviewer exactly what was rewritten between drafts by sending the highlighted copy along with your response letter.
How the words get lined up.
A PDF does not store sentences. It stores short runs of characters, each with a position on the page. The tool asks pdf.js for those runs, sorts them top to bottom and left to right, and splits them on spaces into a plain list of words, remembering where each word sits so it can be boxed later. Sideways text, such as a vertical label in the margin, is skipped.
The two lists for a page are then aligned with Myers' O(ND) diff, published by Eugene Myers in 1986 and still the default inside git. It finds the smallest number of word insertions and deletions that turn the old page into the new one, and it stays fast when the pages are long but the edits are few, which is exactly the shape of most revisions.
A removal followed directly by an addition is listed as a replacement, old wording then new. Because pages are paired by number, an inserted page throws off everything after it, and pages laid out in two columns are read row by row across both, so a long edit in one column can make a few lines in the other look changed too.
Remember that this is a comparison of words. Anything that is only drawn, from a swapped photo to a clause turned red, is invisible to it.
Worked example
Original, clause 4.2
The tenant shall pay $1,200 on the first day of each month.
Revised, clause 4.2
The tenant shall pay $1,350 on the fifth day of each calendar month.
Change list: $1,200 → $1,350, first → fifth, and calendar added. Ten words are shared by both versions, so two get a red box and three get a green one.
| Between the two versions | Detected? |
|---|---|
| A word, number or name reworded | Yes, boxed on both sides |
| Text added or removed, even a single comma | Yes |
| Capitalisation changes ("Seller" to "seller") | Yes |
| A whole page added or removed at the end | Yes, page chip turns green or red |
| A paragraph moved elsewhere on the page | Shown as removed here, added there |
| Bold, italics, font or colour changes | No |
| Pictures, logos, charts or signatures swapped | No |
| Same words, different line breaks | No, reported as unchanged |
| Scanned pages without a text layer | No, run OCR first |
Questions answered.
Which file should I add first?
Does it notice formatting, images or colour changes?
Why do two scanned PDFs say there are no differences?
A page was inserted and now every later page shows as changed. Why?
What is in the PDF I save?
Why is the save button greyed out?
Are capital letters and punctuation compared too?
How does it decide what changed?
Do my two documents get sent anywhere?
More tools you might need.
Related Guides & Tutorials
How to Fix "At Least One Signature Has Problems" Error in Adobe Acrobat
What the yellow "At least one signature has problems" banner in Acrobat really means, how to tell a trust problem from a tampered file, and how to fix each case.
How to Repair and Recover Corrupted PDF Files That Won't Open (Recovery Guide)
Why PDFs get damaged, how to tell whether a broken file can be saved, and how to rebuild a missing cross-reference table or end marker to get the pages back.
How to Convert CMYK PDF to sRGB for Web and Social Media Sharing (True Color Guide)
Why a print-ready CMYK PDF looks off on screens and social media, and how to convert it to sRGB in InDesign, Illustrator, Acrobat Pro or Ghostscript, or export the pages as images.