Choose original and revised files
Select two readable PDFs and confirm which is the baseline. Compare equivalent versions; an inserted middle page can shift the following page-by-page matches.
A changed page can look different even when its extracted words are the same. Conversely, a text change may be hard to spot in a thumbnail. This comparison combines pixel differences with normalized extracted-text checks. It compares pages by their physical index; it does not align moved sections or provide a semantic legal redline.
Select two readable PDFs and confirm which is the baseline. Compare equivalent versions; an inserted middle page can shift the following page-by-page matches.
Adjust render scale and the pixel difference threshold. Different fonts, anti-aliasing or page dimensions can produce visual differences without a meaningful wording change.
Review changed, added and removed page counts. Text checks use extracted PDF text, not OCR, so an image-only scan depends mainly on its rendered appearance.
Save the JSON report or ZIP with the report and available difference PNGs. Open the originals at full size to confirm important findings; the report is an aid to review.
Each input can be up to 150 MiB, with no explicit page-count cap. The workflow compares through the larger page count; missing trailing pages are marked added or removed.
Difference previews are limited to the first 30 changed pages, while summary results cover the comparison. Export is JSON or ZIP, not a rewritten redline PDF. It does not compare every form value, metadata property, signature or document structure.
Compare the old and new PDF to locate changed pages before proofreading the actual wording.
Look for shifted artwork or altered spacing after changing an export setting. Confirm visual differences in the full pages.
Identify missing or additional trailing pages. Use manual review for a new page inserted inside the document.
Both PDFs are rendered and compared in the browser. Source bytes, extracted text and difference images are not uploaded for comparison. The downloaded report includes local source-file references.
Select a question to view the answer.
No OCR is performed. Scanned pages are compared visually.
No. It compares original and revised pages by physical page index.
The available outputs are a JSON report and a ZIP containing the report and available PNG difference images.
No. Render and extracted-text checks are not a byte-level or complete structural identity test.
Open another Fylvexa tool that works well with this task.