A revised document can change in more than one way
A supplier changes a paragraph. A public notice arrives as a scan. A chart changes while every extracted word stays the same. A useful PDF monitor needs to match the kind of evidence you want to review.
PageFluxer now offers three capture choices: native text, optional English OCR, and visual comparison of one PDF page. Each begins with a preview you can inspect before saving a baseline.
Keep the relevant pages in view
Text setup accepts physical page numbers and ranges. Choose 1, 3-5 to follow those positions, or leave the field blank for the complete document. The preview identifies the selected pages and whether each used native text or OCR.
Page selection includes page boundaries and the document page count in the text comparison. It helps focus the wording you review while keeping added or removed pages visible. Positions can shift when a publisher inserts pages, so confirm coverage again after structural changes.
Bring a scanned notice into the text workflow
Choose OCR fallback when a selected page has sparse native text, or OCR every selected page when a scan sits beneath an otherwise readable header. English printed text is supported on up to five selected pages.
Review the wording and confidence before saving. OCR can misread small print, columns, numbers and unusual fonts. Low-confidence or unreadable results keep the previous baseline and report a capture issue. OCR does not interpret diagrams or make handwriting reliable.
See a chart change that text extraction misses
For appearance, choose one complete physical PDF page. Future captures compare its rendered pixels. Open a change to inspect the highlighted diff, compare side by side, adjust an overlay or zoom into the retained images.
Try the synthetic chart before and after examples. Their wording is the same, while a bar changes size and colour. The PDF workflow guide explains how to serve those versions at a stable URL and test a real alert.
A visual monitor watches only its chosen page. This release does not compare every PDF page, understand chart values or authenticate signatures. The existing one-Cloud-monitor-per-URL rule means choosing text/OCR or one visual page for a document.
Keep the evidence and its limits together
PDFs remain public, unencrypted documents up to 5 MiB and 100 pages. OCR runs in the isolated document processor; page images are not sent to an AI provider. Text history contains extracted wording, while visual history uses private workspace images and the existing retention and storage limits. Neither is an archive of the original PDF.
Native text and OCR share the HTML check allowance. Visual PDF checks use Browser capacity and visual/PDF slots, including Free’s Browser trial rules. Existing text monitors keep their extraction until edited. A capture-engine upgrade quietly refreshes its baseline instead of presenting engine differences as a document revision.
Start with one PDF, verify an unchanged check and a deliberate update, and confirm notification delivery. Read the complete PDF guide for page coverage, OCR limits, usage and recovery.