Skip to main content
PDFBase

Compare PDF

Local

See exactly which words changed between two PDFs

All Tools
Compare PDF
Local
Processing limits
  • Rendered PDFs: up to 500 pages, 8,192 px per side, 32 MP per page, 100 MP in total, and 250 MiB of output.
  • These are memory budgets, not upload limits — a large file is only ever read by your own browser.

Choose two PDFs. Both are read on your own device, neither is uploaded, and neither is changed.

Tool Details
Input formatsPDF
Output formatTXT
Max file size50 MB
Max files2
PDF render limits

500 pages · 8,192 px/side · 32 MP/page · 100 MP total · 250 MiB output

ProcessingLocal

Privacy First

Your files are processed entirely in your browser. Nothing is uploaded to any server.

How do I compare two PDF files?

Add two PDFs and the comparison runs on its own. Every changed page can be opened to show the exact words that were added and removed, with unchanged text collapsed so the edits are what you see. Pages added or deleted between the two versions are reported as such rather than shifting everything after them. Both documents are read and compared inside this browser tab, and you can copy or download the result as a plain-text report.

About the Compare PDF

Compare two PDFs in your browser and see the actual differences: which pages were added, removed or edited, and which words changed inside each one, highlighted line by line. Pages are matched by content rather than by position, so inserting a cover sheet does not report the whole document as changed, and a page with no text layer is reported as unreadable rather than as identical.

  • Check a returned contract against the version you sent and read the exact clauses that were altered, before signing it.
  • Find which pages changed between two revisions of a long document so a reviewer reads three pages instead of ninety.
  • Confirm that a re-exported file is the same document, not just the same length — reordered or duplicated content is caught, not scored as a match.
  • Compare confidential drafts without uploading either version to a document-comparison service.

How to use the Compare PDF tool

  1. 1

    Choose two PDFs

    Select both PDFs, or drag them into the upload area. The first is treated as the original and the second as the revision; a Swap button reverses them.

  2. 2

    Read what changed

    The comparison runs as soon as both files are chosen. Pages are matched by content, so an inserted page is reported as one insertion rather than as ninety changes, and each changed page can be opened to see the added and removed wording.

  3. 3

    Export the report (optional)

    Copy or download the comparison as a plain-text report, with or without the changed wording included.

  4. 4

    Nothing is uploaded

    Both documents are read on your device and neither is modified. Only the text layer is compared — a scanned page with no text layer is reported as unreadable, never as identical.

What exactly does the comparison measure?

The extracted text, word by word. Pages are first matched to each other by content, then each matched pair is diffed at word level, so the output is the added and removed wording in place rather than a page number and a percentage. It is a text-level diff, not a visual one: it will not flag a changed logo, a different font, a moved image, or altered colours. What it answers well is the question people usually have — did the wording change, and where.

Why does inserting one page matter so much?

Because the obvious way to compare two documents — page 1 against page 1, page 2 against page 2 — falls apart the moment a page is added or removed. Every page after the insertion is compared against the wrong page and reports as changed, which turns a one-page insertion into a ninety-page diff and buries the real edits. Pages here are aligned by content first, using identical pages as anchors and scoring the rest against each other, so an insertion is reported as one insertion.

What happens with a scanned PDF?

A scan is a picture of a document with no text layer, so there is nothing for a text comparison to read. Those pages are labelled as having no text layer and are excluded from the similarity figure — they are never counted as identical, which is the failure mode that would matter most, because a comparison tool reporting "identical" about two documents it could not read is worse than one that refuses to answer. To compare scanned documents, add a text layer with OCR first.

What can you use the comparison report for?

Reading the exact clauses that changed in a returned contract instead of re-reading the whole thing, confirming that a re-exported document matches the version you approved, or verifying that a supposedly minor revision touched only what it claimed to. The report is plain text, so it can be attached to an email or pasted into a ticket as a record of what was checked — and the changed wording can be left out of it when the wording itself is confidential.

Are my files uploaded when comparing two PDFs?

No. Both files are read and compared locally, and the report is generated in this tab. Comparison is inherently something you do with two versions of a sensitive document — contract drafts, agreements, policy revisions — so the tool never needs, and never has, a copy of either one.

How to do this offline, on the command line

Comparing extracted text catches content changes; comparing rendered pages catches layout changes.

pdftotext + diff

macOS: brew install poppler · Debian/Ubuntu: apt install poppler-utils

diff <(pdftotext -layout v1.pdf -) <(pdftotext -layout v2.pdf -)

Same idea as this page. Catches wording changes; blind to a font or margin change that leaves the words identical.

diff-pdf

macOS: brew install diff-pdf · Debian/Ubuntu: apt install diff-pdf

diff-pdf --output-diff=changes.pdf v1.pdf v2.pdf

Renders both files and highlights pixel differences, so it sees layout and image changes that a text diff cannot.

Frequently Asked Questions

Does it show which words changed, or just a percentage?
The actual words. Each changed page can be opened to show the added and removed wording in place, with unchanged stretches collapsed the way a code diff collapses them, so a one-clause edit in a long page is visible immediately instead of being buried. The counts and the similarity figure are there too, but they are the summary, not the answer.
What happens if a page was inserted or deleted?
It is reported as one insertion or one deletion. Pages are matched by their content before anything is compared, so adding a cover sheet to a ninety-page contract reports one added page rather than ninety changed ones. Positional comparison — page 1 against page 1, page 2 against page 2 — is what makes most comparison tools useless on exactly the documents you most need to check.
Can it compare scanned PDFs?
No, and it says so rather than guessing. A scanned page is a picture of words, not words, so there is nothing to extract and nothing to compare. Those pages are labelled "no text layer" and are excluded from the similarity figure — they are never reported as identical, which is the mistake that matters most here. Run the scans through OCR first if you need the wording checked.
Is a 100% match proof the documents are the same?
It is proof the extracted text is the same, word for word and in the same order. Fonts, images, colours, layout, form fields, annotations and anything drawn as a picture of text are outside what a text comparison can see. Two files can read identically and still look different on paper.
What format is the comparison result?
The comparison is shown on the page. You can also copy it or download it as a plain-text report to archive alongside the documents or paste into a ticket, with or without the changed wording included — leave it out when the wording itself is confidential.
Are the files uploaded?
No. Both PDFs are read in this browser tab, compared there, and discarded when you leave the page. Neither file is modified and neither is transmitted, which is the point when the two documents are drafts of something you are not free to share.

Need More Tools?

Explore our collection of 58 free PDF tools, all with privacy-first processing.

Browse All Tools