ContractorHandbook

Compare Two PDFs and See Exactly What They Changed

A general contractor sends the contract back signed and says they made a couple of small edits. Reading forty pages twice to find out which two sentences moved is not a job anybody does properly at six in the evening. Open both files here and the changed wording is listed for you, page by page, in about the time it takes to drag them in.

what you sentwhat came back9changes
Word by word, not page by page. A contract that comes back with one sentence moved and a payment term quietly rewritten looks identical at arm's length, and this is the thing that catches it.

How it works

  1. Choose the original, the version you sent out. It is opened on your device and never uploaded.
  2. Choose the one that came back. Both are read straight away and the wording is compared page by page.
  3. Read what changed. Switch to Marks on the page for scans and drawings, and download the change list if you need it in writing.

What you get

  • Every changed word listed, with what was removed and what replaced it
  • A few words of the sentence either side, so a change makes sense on its own
  • A count for the whole document, not just the page you are looking at
  • Marks on the page mode, which boxes anything that moved even on a scan
  • Says so when pages were added or taken off the end
  • A change list you can download as a PDF and attach to your reply

The two questions, and why they need two answers

What words changed is the question a contract argument is actually about. One sentence in the payment terms went from thirty days to sixty, one line in the scope of work grew four words, and those four words are now your problem. This is the default here because it needs no drawing at all, which is why the answer appears the moment the second file is open.

What changed on the page is a different question, and it is the only one that can be answered when the thing that came back is a scan. A scanned page has no words in it as far as any software is concerned, only a picture of words. Marks on the page draws both versions at the same size, compares them, and puts a box around every part that is not identical.

Marks on the page catches things the wording never will: a signature block that moved, a detail redrawn on a drawing, initials added in a margin, a box ticked that was not ticked before. Use the wording for a contract that is still a word processor file. Use the marks for anything that has been printed, signed and scanned.

What contractors actually use this for

The contract that came back. You sent a contractor agreement, they sent back a signed PDF, and somewhere in it the retainage went from five percent to ten. That is a real number on a real job and nobody mentioned it in the email.

The subcontract you were handed. A general contractor sends their own paper, you mark it up, they send back a clean copy to sign. Comparing your marked version against their clean one is the only way to know whether they took your changes or quietly put their own wording back.

The revised plan set. Sheet A2.1 came back as revision 3 and the cloud is either missing or covers half the sheet. Marks on the page boxes what is genuinely different between the two revisions instead of what somebody remembered to cloud.

The change order that grew. You wrote a number and a scope. Check that the copy coming back for signature still carries both of them before your signature goes anywhere near it.

How the wording comparison works, and what it is honest about

The text is pulled out of both files and compared word by word, not line by line. That matters, because adding one word at the start of a paragraph pushes every line down and a line comparison would tell you the whole paragraph changed. This tells you one word changed.

Changes are grouped the way somebody reads them. If a sentence has three edits a few words apart, they come back as one change with the removed wording and the new wording under each other, not as three fragments you have to reassemble in your head.

Where it is blunt: if a page was genuinely rewritten rather than edited, lining it up word by word produces noise, so it says the page was rewritten and how many words went each way. And a change to something that is a picture rather than text, a stamp, a logo, a signature, does not appear in the wording at all. That is what the other mode is for, and the page says so rather than letting you assume the wording caught everything.

Pages that were added or taken away

Two files with a different number of pages are the most common version problem there is, and it is also the easiest one to miss, because a forty one page document looks exactly like a forty page document in a mail client.

The pages both files have are compared normally. Anything past that is reported plainly: this many pages were added at the end, or this many were taken off. Those extra pages are not compared against anything, because there is nothing to compare them against.

One thing worth knowing: if a page was inserted in the middle rather than at the end, everything after it shifts by one, and page by page comparison will then report large changes on every page after the insertion. That is not a bug in the comparison, it is what actually happened to the document, and it is usually the most important thing anybody could tell you about it.

What to do once you know

Download the change list and reply with it. A message that says the payment terms on page 4 went from thirty days to sixty is a different conversation from one that says you think something changed. The list is a plain PDF with the removed wording in red and the new wording in green, page by page.

Keep both files. The version you sent is the evidence that you sent it, and the day somebody says the sixty day terms were in there all along is the day you will want it.

If you are the one making the changes, make that easy for the person on the other end. Put a DRAFT stamp on anything that is not final and number the pages before you send it, so a change can be pointed at by page number instead of described.

Questions people ask

Are the files uploaded?

No. Both are opened and compared in your browser, on your own device. There is no server here that could receive a document.

Does it work on a scanned contract?

Use Marks on the page for that. A scan has no text in it, only a picture of text, so the wording comparison will correctly report that neither file has any words in it. Marks on the page compares the pictures and boxes everything that is different.

Why does it say a page was rewritten instead of listing the changes?

Because that page has so little in common with the old one that lining it up word by word produces noise rather than an answer. It tells you how many words came off and how many went on, which is the useful part.

Does it show changes in the drawings?

Marks on the page does. It compares what is actually drawn, so a moved dimension, a redrawn detail or a new note gets a box around it whether it is text or not.

What if one file has more pages than the other?

The pages both files have are compared, and the count says how many were added or taken off the end. If a page was inserted in the middle, expect everything after it to be reported as changed, because everything after it moved.

How big a document can it handle?

The wording comparison reads every page of both files and is quick even on a few hundred pages, because it does not draw anything. Marks on the page draws two pages at a time, only the page you are looking at, so the size of the document does not slow it down.

Related