Give old paperwork a searchable layer
Archives often contain image-only PDFs created by office scanners, multifunction printers or phone cameras. The page may be legible but impossible to search or copy. PDF Guard renders each image-only page inside the browser and runs printed-English OCR locally. The recognised text becomes available for the same checklist, detail and abbreviation checks used with digital PDFs. The original scan is not rewritten or uploaded.
A practical first pass for mixed-quality scans
Use the report to locate dates, reference labels, versions and expected headings across a set of pages. When a checklist item is marked unclear or not detected, inspect both the OCR text and the page image. Old typewriters, stamps, skewed pages, dot-matrix print, bleed-through and faded copies can all change recognition. The checker is most helpful when it narrows attention, not when it is treated as an archival transcription authority.
Improve the source before recognition
If possible, scan at a consistent orientation and roughly 300 dots per inch. Use grayscale or colour when faint marks matter, and avoid extreme thresholding that removes punctuation. Flatten folded pages, keep the camera parallel to the paper and crop away unrelated background. For a bound document, reduce curvature near the spine. A clearer input usually improves names, numbers and short codes more than rerunning the same poor image.
Local limits protect the browser
OCR is computationally intensive. PDF Guard limits a check to forty PDF pages and twenty image-only OCR pages. These limits reduce crashes on mobile devices and make it clear when a document should be split. The recognition model currently supports printed English. Handwriting and other languages require a different model and are not advertised as supported. Large pages may take several seconds each depending on the device.
No hidden archive
The browser downloads the application and recognition model as static files. Your selected scan and recognised text are not posted back to the website. There is no document account, history or cloud storage to revisit later. Download the local report before closing the tab if you want a record, and store it according to the same confidentiality rules as the source document.