What the PDF Comparison Tool does
This tool compares two versions of a PDF and shows every word that was inserted, deleted or changed. It pulls the text out of each page, lines the two documents up, and marks the differences side by side with the words around them, in the full text, and as a count for each page. You can download a report as HTML or CSV.
It compares text only, so it answers "what did they change in the wording?" - a new figure, a dropped sentence or a reworded clause - even when the pages have reflowed. Both files are read in your browser and never uploaded.
How to use it
- Load the old version and the new version, or use the example with two drafts of an annual report.
- Tick Ignore case or Ignore punctuation if those changes do not matter to you, for example when only the quote style changed.
- Select Compare. Start with Side by side: each row is one change, with the old words struck through and the new words underlined, shown with a few words of context.
- Use Full text to read the new version with the changes marked in place, or By page to see where the changes are. Download the HTML report to share the result.
Reading the results
A change is one run of words that differ, with the words on either side the same. When some words were replaced it counts once, with the deleted and inserted words shown together. Unchanged is the share of words that match: twice the matching words, divided by the words in both documents together.
The documents are compared as one continuous text, not page against page. If an insertion pushes a paragraph onto the next page, the moved words are not reported as changes. Page numbers in the results show where each change sits in the old and in the new file.
The comparison works in two steps. Whole lines are matched first, then any lines that do not match are compared word by word. A reworded paragraph that wraps onto different lines is still compared word by word, so a single new word does not show up as a whole rewritten paragraph.
Worked example: two drafts of a library's annual report
The example compares two 2-page drafts. The tool finds 6 changes, with 16 words deleted and 8 inserted. Visitors went from 47,900 to 48,210, the increase from 5 to 6 per cent, e-book loans from 9,780 to 9,870, and "ten author talks" became "twelve author talks". Each of these swaps one word for another.
The fifth change turns "open a study room." into "open a study room and lend laptops.". That deletes one word and inserts four. The sixth change deletes the 11-word sentence thanking the Friends of Riverside Library and inserts nothing. So 1+1+1+1+1+11 = 16 words were deleted and 1+1+1+1+4 = 8 were inserted, which matches the counts.
Formulas and scoring rules
- Unchanged share
unchanged = 2 x matching words / (words in old + words in new)Rounded to one decimal place. Two empty documents count as 100 %.
Limitations: what the result does not prove
- Only text is compared. A changed image, chart, colour, font, table border or page layout is not detected.
- Scanned PDFs without a text layer have no words to compare. Run OCR on them first. Text is read in the order pdf.js extracts it, so in multi-column layouts or text boxes the order can differ from what you see on the page.
- Words split with a hyphen at the end of a line are treated as two pieces. If one version breaks the word and the other does not, it shows up as a small change.
- Up to 1,500 pages and 400,000 words per document can be compared. Password-protected files cannot be opened.
Privacy: where your data goes
Everything you paste, type or drop is processed in this browser tab. It is not uploaded, logged, stored or sent to analytics. Session recording and tag-manager scripts are switched off on this page.
Standards and sources
- Eugene W. Myers - An O(ND) Difference Algorithm and Its Variations (Algorithmica, 1986)
- Mozilla - pdf.js (text extraction)
- ISO 32000-2:2020 PDF 2.0, section 9.10 Extraction of text content
Frequently asked questions
How do I compare two PDF files for differences?
Load the old and new PDFs here and select Compare. Every inserted, deleted and changed word is listed side by side with context and page numbers. You can also read the full text with the changes marked, and download an HTML report.
Why does the tool say there are no differences when the PDFs look different?
It compares only the text. If the change was to an image, a colour, a font or the layout, the words are the same, so no text differences are reported. Compare the page previews by eye for visual changes.
Does adding a paragraph make every later page show as changed?
No. The whole text is compared as one sequence rather than page against page, so words that moved to another page still match. Only the paragraph you added is reported.
Can I ignore changes to capitals and punctuation?
Yes. Tick Ignore case to treat Report and report as the same word. Tick Ignore punctuation to ignore commas, full stops, dashes and the difference between straight and curly quotes.
Can it compare a Word document with a PDF?
Not directly. Save or print the Word document as a PDF first, then compare the two PDFs. The comparison uses only the text, so different layouts do not matter.
Are my documents uploaded to compare them?
No. Both files are read by pdf.js, which is served from this site and runs in your browser, and the comparison also runs in the page. Neither document leaves your device.
Last reviewed by the A2Z.Tools team against the sources listed above.
