Skip to content

Compare PDF

Compare PDF Text

Only the words, so a reformatted document still reads as unchanged.

Processed locally in your browser. Your file is not uploaded to any server. How this works

  1. 01Add your files
  2. 02Compare PDF Text
  3. 03Download

Runs in your browser · nothing uploaded

About compare PDF text

This page locks the comparison to words, and that restriction is the feature. The situation it serves is the returned draft: counsel sends back a revised agreement, a client returns a marked-up proposal, a regulator issues an updated version of a standard, and you need to know what they actually changed rather than that they also happened to regenerate the PDF from a different template. Because only the extracted text is compared, everything presentational drops out — new fonts, different margins, a redesigned header, a change of page size — and what remains is the substantive edit. The text is normalised before alignment: hyphenated line breaks are rejoined, runs of whitespace collapsed, and typographic quotes and dashes folded to their plain equivalents, so a document retyped in a different editor does not light up with hundreds of cosmetic differences that bury the one clause that moved.

How to compare PDF text

  1. 01

    Add the original and the revision

    Original first, revised second, so additions and deletions read the right way round.

  2. 02

    Choose whether whitespace counts

    Leave it ignored unless indentation itself is meaningful in your document.

  3. 03

    Run the comparison

    Both text layers are extracted, normalised and aligned locally.

  4. 04

    Work through the changes

    Each difference is shown in context with its page number so you can verify it against the document.

What this tool does

  • Compares content only — restyling and re-templating are invisible to it
  • Hyphenation, whitespace and typographic punctuation normalised before alignment
  • Each difference reported with its page number and surrounding context
  • Insertions and deletions distinguished, not merged into a vague "changed" flag
  • Runs locally, so confidential drafts never leave your machine

Limitations worth knowing

Every PDF tool has constraints. Stating them plainly is more useful than discovering them halfway through your work.

  • Formatting, font, colour and layout changes are not detected at all — use Visual PDF Comparison for those.
  • Both documents need a real text layer; scans return nothing until OCR'd.
  • Table contents are compared as sequential text, so a value moved between columns may read as a deletion plus an insertion.
  • Text drawn as vector outlines is invisible to the comparison.

How your file is handled

This tool runs entirely inside this browser tab. When you choose a file, your browser reads it from your own disk and hands the bytes to JavaScript running on this page — no network request carries your document anywhere. You can confirm that yourself: open your browser’s developer tools, switch to the Network panel, and run the tool. You will see no upload.

Nothing is stored after the fact. Closing or reloading this tab discards the file, the result and everything derived from them, because none of it ever left your machine. Read how local processing works.

Questions about compare PDF text

The two files look completely different but the report is empty. Is it broken?

No — that is the expected result when only presentation changed. Identical wording in a new template produces no text differences by design.

Will a reflowed paragraph be reported as a change?

Not with whitespace ignored. Line wrapping is a layout artefact, and folding it away is precisely what stops reflow from hiding the real edits.

Can I compare a scan against a digital original?

Only after OCR'ing the scan, and be cautious: recognition errors will appear as differences that do not exist in the documents.

How is this different from PDF Text Diff?

Same method, different audience. This page is about reviewing a revision to approve or reject changes; PDF Text Diff presents a developer-style word-level diff of the two texts.

Read more about this