Recognising scanned PDFs
A scanned PDF is a stack of images. Until the images become characters, nothing in it can be searched, selected, quoted or asked about.
Before you start
A scanned PDF open in MindMax. It will open fine as images — you can read it that way without recognising anything.
Starting recognition
Recognition is a button you press — Recognize text with AI, in the banner at the top of the open scan. It does not happen automatically and it does not run across your folder.
That is deliberate: recognition costs credits, and you are the one who knows which document is worth it.
Flip through the scan first. Badly skewed pages, tight gutters on a bound book and heavy bleed-through from the reverse side all cap the quality you can get — and they are visible before you spend anything.
While it runs
It works in batches and reports progress as it goes. Pages that are done are kept, so closing the app or losing power does not throw away completed work — restarting resumes.
For a long book this takes a while. You can keep working elsewhere.
Reading the quality report
Do this before you start quoting from the document.
Each page gets its own record, and pages are flagged when the output looks wrong. The check looks at the shape of the result rather than searching for expected words, so it behaves the same way on any language.
Why per page and not one number for the document: a page that fails completely is harmless, because you can see it is nonsense. A page that comes out mostly right is the one that damages your work, because you will quote from it without noticing the digit that flipped.
Re-running the bad pages
Re-run recognition on the flagged pages only. If a page fails again, that is the scan's fault rather than the recogniser's — and now you know to check quotes from that page against the image.
After recognition
The document behaves like any other: search reaches into it, you can highlight and annotate it, and answers about it carry quotes that link back to the page.
Your original PDF is not modified. The recognised text is kept by MindMax in the hidden .mindmax
folder and shown in text mode, which is also where you highlight and annotate a scan.