01 / Two review modes
Use Visual diff for rendered pages
Visual diff rasterizes the pages, highlights changed regions, and reports a per-page visual-match percentage. It can expose layout, image, chart, stamp, and signature-mark appearance changes that extracted text may miss.
The result describes rendered pixels. It does not validate a digital signature, prove document authenticity, or explain the semantic meaning of a visual change.
The free plan shows the first 3 pages in Visual diff and compares the extracted text of the whole document, for files up to 5 MB each. Pro shows every page and opens larger files.
- Review changed regions page by page.
- Use true visual evidence for scans and artwork.
- Treat signature appearance and cryptographic validity as different questions.
02 / Two review modes
Use Extracted text for wording
Extracted text keeps wording changes navigable across the document without reducing them to page pixels. This is useful when the question is whether clauses, labels, or paragraphs changed.
A PDF may have incomplete or unusual text encoding. Scanned pages without extractable text require OCR, and Diff Anything does not perform OCR, so Visual diff is the appropriate built-in evidence for those pages.
03 / Two review modes
Keep the review local and the conclusion narrow
Both PDF files are opened and compared on the Mac or Windows computer. Diff Anything has no account, cloud comparison service, analytics, or telemetry.
Use both modes when wording and presentation matter, then inspect the original documents for metadata, attachments, forms, digital-signature validity, or other properties that neither view claims to prove.