Compare PDF

How to compare two PDFs and find the differences

1. Choose the old version and the new one (or pick both at once in the first box: the first file is taken as the old). Got them the wrong way round? Tap ⇅ Swap.
2. Choose how to compare: ignore capitals, ignore punctuation, and whether to compare the look of every page as well as its words.
3. Tap Compare. Every difference is listed in words, such as Page 2: Rs. 5,000 → Rs. 6,000; tap one to see it on the pages, or step through them with Previous and Next.
4. Save a short report of the differences as a PDF, to keep or share, if you need one.

The two versions are shown page by page, side by side (on a phone the old above the new): words removed are marked red on the old page, words added green on the new, and parts that look different are outlined in amber. Pages added or removed are said plainly. Everything happens in your browser: neither PDF is uploaded to ap7am or anywhere else.

How the words are compared

The words of each PDF are read in reading order, line by line, and the shortest list of words removed and added that turns the old into the new is worked out (the same method as the programs that compare versions of computer code). Each change is then put back on its page, at the words themselves.

• Spacing and line breaks never count: a paragraph that flows differently, or moves on to the next page, is the same. Neither do ligatures (fi and fi) or full-width letters.
• Capitals and punctuation count unless you tick the boxes to ignore them.
• Pages are paired by the words they share, so a page added in the middle does not make every later page look changed.
• A paragraph moved to another place is shown as removed where it was and added where it is now.

How the look is compared

Scanned pages and photos have no words to compare, so their look is compared instead; tick the box to do it for every page (a new stamp, a signature, a changed picture, bold or colour). Both pages are drawn at the same scale, 144 dots an inch, laid over each other, and the areas whose pixels differ are outlined. Letters drawn a shade lighter or half a pixel over do not count, and you choose how small a difference to show.

What it cannot do

• Two separate scans of the same paper never line up exactly, so their look differs almost everywhere: the look comparison is for versions made from the same file.
• Some PDFs, often older Telugu ones, store their letters in a way that cannot be read back; their words may be compared wrongly. Use the look comparison for those.
• A change in a word's font, size or colour is not a change in the words: the look comparison shows it.

Questions

To black out details before you share a PDF, use Redact PDF; to take the words out of a PDF, PDF to Text.

Made with pdf-lib (MIT licence) and Mozilla's pdf.js (Apache 2.0 licence), both running in your browser; the report's Telugu lines in fonts from Google Fonts (SIL Open Font License).