Compare two versions of a document

Someone sends back the contract and says they made "a few small changes". Reading both versions side by side to find them is an hour you will not get back, and the change that matters is usually the one word you skim past. This finds them.

or drop a file anywhere on this panel

opened locally · 0 bytes sent

How it works

  1. 01Open the original.
  2. 02Open the revision as a second file.
  3. 03The changed words are marked on the pages and listed page by page.
  4. 04Export the marked-up revision, or just read it.

An inserted page is one insertion, not forty rewrites

This is where comparison tools fall down. Compare page one to page one, page two to page two, and a single inserted paragraph pushes everything down — so every page after it is reported as rewritten. That is technically true of the page images and useless to a person. Here the pages are matched up first, by their content, so an inserted page is reported as an inserted page and the rest are left alone.

A page that only reflowed is not a change

Re-exporting a document from a different program moves the line breaks without moving a word. Comparison ignores whitespace and capitalisation, so that shows as no change — which it is. What it does not ignore is punctuation, because in a contract a comma can be the whole argument.

It compares words, and says so

Two documents whose text matches can still differ in layout, fonts, images or signatures. The panel says this rather than leaving you to assume a clean result covers everything. A tool that quietly implies more than it checked is worse than one that checks less and tells you.

Two scans are not "identical"

If neither document has extractable text — two scans, say — there is nothing to compare, and a word comparison would find no differences and report a match. That would be a confident lie about documents nobody has actually read. Instead it tells you plainly that both look like scans and suggests running OCR on them first.

Both files stay on your machine

Sending two versions of a contract to a comparison website means uploading the confidential document twice. Here both are read by your own browser and neither leaves it, which you can confirm by disconnecting from the network before you start.

Questions

Which file is treated as the original?
The first one you open. The second is the revision, and the changes are marked on it. You can flip which one the pages show without reopening anything.
Why are deleted words not highlighted on the page?
Because they are not on the page — they were deleted. Deleted text is listed in the panel beside the page it was removed from, which is the only honest place to put it.
Can it compare images or layout?
No. It compares text. A visual comparison of two pages is a different tool and is not available yet.
How many pages will it compare?
Up to three hundred from each document. Matching pages between two long documents gets expensive quickly, and a partial comparison you can trust beats a complete one that locks the tab.