Compare two PDFs
Drop in two versions and see what changed, with page numbers.
Using toolbook means you agree to our Terms and Privacy Policy.
or drag it here
One PDF file
or drag it here
One PDF file
Your files stay on your device
How to use it
Drop two versions of a PDF here to see every added and removed passage, with page numbers. Get an optional plain-English summary too. Your PDFs stay on your device.
Got a new version of a contract or report and need to know what changed? Drop both PDFs onto the box above and press Compare.
Every added and removed passage is highlighted, the exact words that changed are marked, and each change shows its page number.
Want the gist? Where your browser supports it (Chrome and Edge on desktop), you can get a short plain-English summary of the changes.
Only the words are compared — not pictures, fonts, colours, or layout. Scanned pages don’t have words to compare, but the tool can read them for you on a second pass.
FAQ
How do I compare two PDFs?
Drop both in and press Compare. Every change is listed with its page number.
Does it compare pictures or layout?
No, only the words. Two PDFs can look different and still match if the text is the same.
It says my PDF has no text. Why?
It’s a scan — pictures of pages. The tool can read those pages on a second pass.
Why is a whole paragraph marked when I only changed a word?
Check the paragraph setting is on. Line mode is for tables and code.
Do I need the AI summary?
No. The comparison works without it. The summary is an extra where your browser supports it.
Can anyone else see my PDFs?
No. They’re compared on your device and never sent anywhere.
How it worksFor the curious
Text is extracted with pdf.js and compared with the same Myers diff the text comparison tool uses, with word-level highlighting inside edited passages.
Comparison is by paragraph rather than line. PDF text is positioned, not flowed, so inserting one word near the top of a paragraph shifts every line below it; a line diff would report the whole paragraph as rewritten. Line mode remains for tables and code listings.
Pages with no text layer (scans) are named and can be read with Tesseract OCR on a second pass, which is opt-in because it downloads a recognition engine.
The summary uses your browser’s built-in language model, on your device. It receives only the changed paragraphs and their page numbers, never either whole document.