A scanned PDF or a photo of a page has no text layer, so Silvertext reads it with OCR first and then checks the recognized text.
How it works
Upload the PDF or image as usual. If a PDF has little or no selectable text, it is routed to OCR automatically; PNG and JPEG images always are. The composer shows the file while the text is read, and the report opens on the recognized text.
Who can use it
OCR is for signed-in members. A free scan without an account refuses scanned PDFs and images with a prompt to sign up; the free account includes one full document.
What to expect
- Clear, typed pages come through accurately. Handwriting, low-resolution photos and skewed scans lose words, and a lost word can look like a spelling error in the report.
- Very long scanned PDFs are cut at a page limit so one read stays inside what the OCR step can carry. The report tells you where it stopped.
- Check the recognized text on the page before acting on findings. If a passage reads wrong, fix it in the composer and check again.
A better alternative
When you have the original Word or text file, upload that instead. Extracted text is exact; OCR is an approximation.