How to compare two PDFs and find what changed

Reading two versions side by side works until the document is longer than a page or two. Then you miss the one clause that mattered — which is usually the one they changed on purpose.

Revised documents come back without tracked changes constantly. The other side exports a clean PDF, the version number stays the same, and you are left to work out what moved.

Two kinds of difference

Comparison tools look for one of two things, and knowing which you need saves a lot of confusion.

• Text differences — wording that was added, deleted or reworded. This is what you want for contracts, policies and anything legal.
• Visual differences — pixels that moved. This catches layout shifts, swapped images and changed formatting that a text comparison ignores entirely.

A clause reworded from 30 days to 60 days is a text difference. A logo replaced or a table redrawn is a visual one. Most disputes turn on the first, most design reviews on the second.

Why this is worth doing carefully

Contracts are exactly the documents you should not be uploading to a website to compare. Two versions of a commercial agreement, sent to a server you know nothing about, is a worse problem than the one you started with. There is more on that in what actually happens to files you upload.

Comparison here runs in your browser. Both documents stay on your device, which is the only sensible way to diff two versions of something confidential.

Doing it

1. 1Open Compare PDFs and add both versions.
2. 2Put the original on the left and the revision on the right — the direction decides whether a change reads as an addition or a deletion.
3. 3Work through the highlighted differences rather than skimming. Short changes are the easy ones to miss.
4. 4Check the page count first. A document that gained a page has probably gained a clause.

When comparison finds nothing but the file still differs

If both versions read as identical yet the files are clearly not, one of them is probably a scan. A scanned page holds an image of text, so there is no text to compare. PDF to Text will tell you within seconds whether a document contains real text or just a picture of it.

Metadata changes — author, creation date, producing software — will not show up in a content comparison either. A file can differ byte for byte while every visible word stays the same.

Common questions

Can I compare a scanned PDF with a digital one?

Not as text, because the scan has none — it is an image of a page. You would be comparing a picture against words, which no tool can do meaningfully.

Does comparing change either file?

No. Both are read only. Nothing is written back to either document.

What if the two versions have different page counts?

That is usually the most important finding on its own. Comparison still works, but check what was inserted or removed before reading the wording changes.