Find every wording change between two drafts of a PDF, in your browser. No upload, no file size cap, no account. Download either side with the changes marked.
🔒 Runs in your browser
The files you most need to compare are the ones you least want to upload
Two drafts of a contract. A settlement before and after the other side's edits. A filing you think is the same as the one you approved.
Finding what changed between them is ordinary work, and the ordinary free answer is to upload both documents to a website. For a menu or a manual that is fine. For a contract under negotiation it is a decision most people would rather not make - and one well-known service states plainly that free comparisons are kept for thirty minutes.
Here both files are read in your browser and the comparison happens on your machine. Nothing is transmitted, so there is nothing to retain, and there is no size cap or one-at-a-time queue, because there is no server doing the work.
What you get
Every added and removed word, grouped into the places they occur, each with its page number. Then either document to download with the changes highlighted in place - the revised copy showing what was added, or the original showing what was taken out. Which one you want depends on who you are sending it to.
⚠️ It compares words, not appearance
This is the honest limit and it is worth stating twice.
The comparison reads the text of both documents and lines it up. A change to a font, a colour, an image, a chart, or a table redrawn with the same numbers does not change the words, so it will not appear here. Neither will anything on a scanned page, because a scan has no text in it at all.
If a document's layout is the thing under negotiation, this is not the tool for that.
It will not tell you two scans are identical
A scanned PDF is a picture. There are no words in it to compare, so a naive comparison of two completely different scans would report no differences - which reads as reassurance and would be the most harmful thing this tool could say.
So it checks. When either file is a scan it offers to read the pages first, with the same recognition OCR PDF uses, running here in your browser. If only some pages are scans, the comparison runs on the rest and tells you how many pages it could not see.
A comparison of two scans is evidence, not a verdict
Once recognition is involved, the danger flips round. The risk is no longer that a real change is missed - it is that a change appears which nobody made, because recognition misread one page. It confuses rn with m, and l with 1, and on a contract "this clause was altered" is about as damaging wrong as it is right.
So the result is marked in two ways. Words the recogniser was unsure about are labelled unclear and are not counted as edits, and because replacing a word produces a removal and an addition together, a doubt about either half marks the whole change. And a comparison built on recognition always says so above the results, where it cannot be missed.
Read it as a list of places to check, and check the ones that matter against the pages themselves.
This tool runs entirely inside your browser. Your file is never uploaded — it never leaves your device, so there is nothing for us or anyone else to read, store or leak.
Common questions
Are my documents uploaded to compare them?
No. Both files are opened and read in your browser and the comparison is arithmetic done on your own machine. This matters more here than on most tools: the documents people need to compare are contracts, settlements, offers and filings, and the usual free option asks you to upload both of them. One well-known comparison service keeps the uploaded files for thirty minutes after the job. Nothing here is transmitted, so there is nothing to keep.
Is there a file size or page limit?
No. There is no cap because there is no server paying for the work - your own machine does it, so there is nothing to ration. Free comparison services typically cap the upload at around 10 MB and allow one comparison at a time, which is a billing decision rather than a technical one. Very large documents take longer here, and that is the only difference.
What kind of changes does it find?
Wording. Every word that was added and every word that was removed, in the order they appear, with the page number. Changes of case and punctuation count, because in a contract "shall" and "Shall" or a missing full stop can matter. What it does NOT find is anything that is not text: a changed font, a different colour, a moved or edited image, a redrawn chart, or a table rebuilt with the same numbers in it. It is a wording comparison, not a visual one, and it is better to know that up front.
Can it compare two scans?
Yes, with a caveat worth reading. A scanned page is a photograph of a document - the words are pixels, so there is nothing to line up until they are read. The tool detects that and offers to read both files here in your browser, using the same recognition as OCR PDF; nothing is uploaded and the first scan downloads about 4 MB of recogniser, once. The caveat is that recognition is a guess. It can confuse "rn" with "m", and a misreading on either side looks exactly like an edit, so a comparison of two scans is evidence to check rather than a verdict. Anything read unclearly is marked "unclear" and is not presented as a change. It will never silently report "no differences found" between two scans, because that reads as reassurance and would be the worst thing this tool could do.
What if the two files are not really versions of each other?
It says so. The comparison lines the two word streams up against each other, which works well when the files are drafts of the same document and stops making sense when they are unrelated. Rather than returning many thousands of meaningless differences, or freezing your browser trying, it stops and tells you the two files could not be lined up.
Is this a legal redline?
No. It shows you which words differ. It does not interpret them, it does not know which changes are material, and it has no view on whether a clause moved or was rewritten. Treat it as a fast way to find the places worth reading carefully, not as a substitute for reading them.