Line up the pages of two PDFs, list the text that changed, and mark the pixels that differ in red.
This tool compares two PDFs by reading the text of every page and drawing every page, then showing the lines that changed and the pixels that differ. The text of each page is rebuilt from the position of its words, so lines come out in reading order, and the two pages are compared line by line. The words that changed are listed with the unchanged lines around them. For the appearance, page A is drawn 1000 pixels on its longer side and page B at the same zoom, and a pixel counts as different when its red, green, blue or transparency value differs by more than the sensitivity: 48 levels on Low, 24 on Normal, 6 on High, out of 255. The result is page A faded to 35 percent with every differing pixel in red, and the share of pixels that differ is shown. A page whose words match but which looks different, for example after a font or image change, is marked Looks different. Appearance is compared for the first 100 page pairs; later pairs are compared by text only, and the result says so.
If both PDFs have the same number of pages, page 1 is compared with page 1, page 2 with page 2 and so on. If the counts differ and you match by content, pages whose text is identical are matched first, and the pages between two matches are paired in order. Take PDF A with pages P, Q and R and PDF B with pages P, Q2, X and R, where Q2 is Q after an edit. P and R match exactly. Between them A has Q and B has Q2 and X, so Q is paired with Q2 and X is reported as only in B. A page that was edited and also moved far away is reported as removed in one place and added in another. Matching by page number is available when you want a fixed pairing. With very large PDFs, when the two page counts multiplied pass 4,000,000, pages are matched by number and the result says so.
It does not read scanned pages. A scan holds a picture of each page and no text, so there are no lines to compare. Choose Appearance only for those, or read them with PDF to JPG followed by Image to Text first. Text that was reflowed shows as changed lines even when the words are the same, because a new line break makes a different line. Pages are compared at the same zoom, so a page that is A4 in one PDF and Letter in the other shows as different. A shift of one pixel moves every edge on the page and raises the percentage. The tool does not compare images inside a page separately, fonts, bookmarks or metadata, and it does not merge or edit either PDF. A page with more than 60000 pieces of text counts as having none. A PDF over 200 MB or over 1000 pages asks you to confirm before it loads, and the comparison stops if you press Cancel and keeps the pairs done so far.
Both PDFs are opened and read by code running in this tab, with the PDF reader running in its own background worker. Pages are drawn on canvases here and compared one pair at a time. There is no upload step and no server that receives your files. The page counts visits with Google Analytics, using rough groups such as 1 to 10 MB, never a file name or any text from the documents. The differences are kept in memory while the page is open and are not stored or put in the address bar. You can download a plain-text report and a zip of the red difference pictures, and both are made in this tab.
Add the original as PDF A and the new version as PDF B, keep Text and appearance selected, and press Compare PDFs. You get a list of pages that differ. Open a page to see the changed lines or the picture with the differences in red. Download compare-report.txt for a text record.
Lines are compared as whole lines. If an edit makes a paragraph one word longer, every line after it may break in a new place, and each shows as changed. Look at the added and removed lines together. Ignore spaces at the start and end of lines, which is on by default, stops indentation and empty lines from counting.
With Match pages By content, pages with identical text are matched first and the pages between them are paired in order. A 4 page PDF compared with its 5 page revision reports the new page as only in B and does not shift the others. By page number compares page 1 with page 1 and so on.
It sets how far a pixel's colour must move, out of 255, to count as different: 48 on Low, 24 on Normal and 6 on High. Use Low to ignore faint shading, High to catch a slightly different grey. The share of differing pixels is shown for each page.
Only by appearance. A scan has no text layer, so the text comparison finds nothing and the tool says so. Choose Appearance only to see where the pictures differ. If the scans are slightly shifted, many pixels will differ even though the words match.
No. Both PDFs are read and drawn in your browser, and nothing is uploaded. The page counts visits with Google Analytics using rough size groups such as 1 to 10 MB, never a file name or any text from the documents. The report and the pictures are made on your device.