How to Compare Two PDF Files for Differences
September 23, 20265 min readHow text-based PDF comparison works, how to read an inline word diff, and where a text-only comparison falls short of a visual diff.
“What actually changed?” is a surprisingly hard question to answer just by skimming two versions of a document side by side — small wording changes hide easily in a wall of text. A proper diff makes them impossible to miss.
Text comparison, not visual comparison
The Compare PDF tool compares the extracted text of two PDFs, word by word — it doesn’t compare pixels, layout, fonts, or images. That’s a deliberate choice: for documents like contracts, articles, and reports, what matters is almost always the wording, and a text diff surfaces that far more reliably than trying to spot visual differences between two page renders.
How to read the result
Once both PDFs are uploaded, the comparison view highlights:
- Green text — words added in the revised version.
- Red, struck-through text — words removed from the original.
- Plain text — content that’s unchanged between both versions.
Everything is shown inline in reading order, so you can scroll through the whole document and see changes exactly where they occur, rather than as a separate list disconnected from context.
Good use cases
- Checking what a counterparty actually changed in a contract redline.
- Reviewing edits between two drafts of an article or report.
- Confirming that a “minor update” to a policy document really was minor.
Where a text diff falls short
If the difference you care about is visual — a shifted logo, a changed color, reformatted tables — a text-only diff won’t catch it, since it deliberately ignores everything but the words. Similarly, a scanned PDF with no embedded text layer will extract as little or nothing, which can make an unchanged scanned page look like a huge diff simply because there’s no real text to compare. Try Repair PDF first if a file seems to extract strangely — a subtly damaged PDF can sometimes affect text extraction.
A note on privacy
Both PDFs are read and compared entirely in your browser — extracting text and computing the diff all happen locally, with nothing uploaded to a server.