Compare two PDFs
Drop in both versions and see what changed, paragraph by paragraph, with page numbers.
By opening this website or using its tools, you agree to our Terms and Privacy Policy.
or drag it here
PDF — nothing is uploaded
or drag it here
PDF — nothing is uploaded
Your file never leaves your browser during processing
How it works
To compare two PDFs, drop both onto this page — the text of each is extracted in your browser with pdf.js, joined back into paragraphs, and diffed, so you get every added and removed passage labelled with its page number, and neither file is uploaded.
Both PDFs are read in your browser and their text is pulled out with pdf.js, then compared with the same Myers diff the text comparison tool uses. Additions, removals, and the words that changed inside an edited passage are highlighted, and every change is labelled with the page it sits on.
The comparison happens by paragraph rather than by line, and that choice matters more than it sounds. Text in a PDF is positioned, not flowed — the line breaks are wherever the original typesetter ran out of column width. Insert one word near the top of a paragraph and every line below it shifts, which a line-by-line diff reports as the whole paragraph being rewritten. Joining the lines back into paragraphs first means a one-word edit reads as a one-word edit. Line mode is still available for documents where the line genuinely is the unit, like tables of figures or code listings.
Only text is compared. A changed image, a moved logo, a different font, a colour change — none of that reaches the text layer, and the tool says so on every result rather than reporting "no differences" as though it had checked. Pages that carry no text layer at all, which is what a scan is, are named individually and can be read with OCR on a second pass; that step downloads a recognition engine, which is why it is opt-in rather than automatic.
The plain-English summary underneath is optional and runs on your browser’s own built-in model where there is one — currently Chrome and Edge on desktop. It is given the changed paragraphs and their page numbers, never either whole document, and nothing is ever sent to a server. On a browser with no model the diff above is unchanged; it needs no model to work.
FAQ
How do I compare two PDF files?
Drop both onto this page and press Compare. The text of each is extracted in your browser and diffed, and the result lists every added and removed passage with its page number. Nothing is uploaded.
Does it compare images or layout?
No — only text. Two PDFs can look very different and still match here if the words are the same. The result states this on every comparison so it is never mistaken for a full visual check.
It says my PDF has no text. Why?
It is almost certainly a scan: a picture of a page with no text layer behind it. Those pages are named in the result and can be read with OCR, which is offered as a second pass because it downloads a recognition engine and is much slower.
Why are whole paragraphs marked as changed when I only edited a word?
That is what line-based comparison does to a PDF, and it is why this tool compares by paragraph instead. If you are seeing it, check that the paragraph toggle is on — line mode is there for columnar documents where the line is the real unit.
Is the AI summary required?
No. The comparison is done entirely without a model. The summary is an optional layer that runs on your browser’s own on-device model where one exists, and the page tells you plainly when your browser has none.
Are my PDFs uploaded?
No. Both files are read and compared in your browser, and the optional summary runs on a model your browser provides and manages. The network counter under the tool stays at zero bytes uploaded while your files are processed.