Two kinds of comparison
A text comparison pulls the words out of both files and lines them up, showing added, removed and changed wording. A visual comparison overlays the pages as pictures and highlights where the pixels differ. Contracts, quotes and reports want the first. Drawings, logos and layout proofs want the second.
What a text comparison catches
- Clauses added or deleted.
- Numbers changed: a date, a price, a notice period.
- A single word flipped, such as "may" to "shall", which is the kind of change a quick read never finds.
It will not see a changed font, a moved image, or a signature block that shifted to the next page, because none of that changes the text.
What it does not catch
If the PDF is a scan, there is no text to compare: the page is a photograph of words. Run text recognition first, then compare the recognised text. Recognition is not perfect, so treat a handful of odd one-character differences as likely misreadings rather than real edits.
Reading the result
Comparisons work line by line, so one inserted sentence can push every following line out of step and light up the whole rest of the page. Look at where the difference starts, not at how much is highlighted. Reflowed paragraphs do the same thing when the margins change: the words are identical but every line break has moved.
Before you compare
- Check the page count. If it differs, something structural changed and that is worth finding first.
- Compare the word counts. A big gap says content moved; an identical count with differences says wording was swapped.
- Use the same section. Comparing a full document against an extract produces noise, not information.
Keeping it private
Contracts, payroll files and medical records are exactly the documents people upload to a stranger's server to check a difference. Comparison that runs in the browser never sends the file anywhere, which is the only version of this worth using for anything confidential.
Tools
Compare PDF Files lines up the text of two PDFs and marks what changed, in your browser. Diff Checker does the same for text you paste, PDF Word Counter gives the counts for a quick sanity check, and Image to Text recognises the words in a scanned page so there is something to compare.