Skip to content

OCR scanned pages

When to use this

You imported a scan or a photo of a document and ⌘F finds nothing on the page.

  1. Open Inspector → OCR.
  2. Under Pages, keep Scans only for the normal safe path. Orifold targets pages that do not already have useful searchable text.
  3. Choose All visible only when an existing text layer is wrong or incomplete. Orifold warns you before this broader pass.
  4. Leave Language on Automatic, or choose a specific language when you know the scan’s language and want Vision to use it directly.
  5. Keep Continue after a page fails on when a useful partial result is better than stopping the entire document. Turn it off when every page must succeed together.
  6. Click Make Searchable. Progress and Cancel stay visible while Orifold works.
Tip

OCR runs locally using Apple’s Vision framework. A scan of your passport, a lease, or a medical form never leaves the machine.

When the run finishes, the Inspector shows a receipt rather than silently declaring success. It records how many pages were requested and recognized, any skipped page numbers, recognized and low-confidence line counts, average confidence, and whether changed documents passed the structural postflight.

Review the highlighted pages when the receipt flags skipped pages or low-confidence text. A successful structural check means the updated PDF is readable; it does not mean every word in a difficult scan was recognized perfectly.

⌘F now finds recognized text on the page. Orifold adds a searchable layer to the original page instead of rebuilding it, preserving page geometry and existing annotations. The operation is also available to Undo and Redo like other workspace changes.

Note

Run OCR before editing page text or objects. If the workspace already contains committed text or object edits, undo those edits before OCR; Orifold blocks the operation so one page-rewrite path cannot overwrite another.

Edit existing PDF text

Editing text once it’s recognized.

Search inside your workspace

Searching across OCR’d and native pages alike.