BASALT · JOURNAL
OCR in Acrobat and on the Mac
Acrobat's OCR is mature and good, with language support, output options and control over how recognised text is stored. macOS has recognition built into the Vision framework, which any native application can use.
Both produce an invisible text layer over the page image. The differences are in control and in integration rather than in the concept.
Where Acrobat is stronger
Breadth of language support and control over output modes, including the option to replace the page image with cleaned up rendering rather than merely adding a layer beneath it.
Handling of unusual documents, where a mature engine with years of tuning generally does better than a general purpose framework.
Where the system framework is competitive
Speed and integration, since recognition happens locally with no separate engine to install or licence, and it is maintained as part of the operating system.
Privacy, since Vision performs recognition on device. There is no upload and nothing to configure to prevent one.
The thing that matters more than either
OCR adds real text to the file, and that text is a redaction hazard.
If a scanned page is redacted by destroying pixels, and the invisible OCR layer beneath still contains the words, the document looks correctly redacted and remains machine readable. Extracting the text recovers exactly what was supposedly removed.
This is not a criticism of any particular OCR implementation. It is a consequence of what OCR does, and it means the redaction step has to handle both the image and the text layer, and that verification has to check both.
Order of work
OCR first, then review and mark, then redact, then verify. Running OCR after redaction produces a text layer over a document you have already finished and can recognise text near the areas you removed.
What to check afterwards
Extract the text from your finished document and search for the terms you removed. On an OCRed scan this takes seconds and catches precisely the failure described above. If the extraction returns the removed words, the image was redacted and the text layer was not.
Frequently asked questions
Is Acrobat's OCR better than the Mac's built in recognition?
Acrobat has broader language support, more output control and years of tuning on difficult documents. The Vision framework in macOS is competitive on speed and integration, needs no separate engine, and performs recognition on device.
Does OCR create a redaction risk?
Yes. OCR adds real machine readable text over the page image, so a redaction that destroys the pixels but leaves the text layer intact produces a document that looks redacted while remaining extractable.
Should I OCR before or after redacting?
Before. OCR first so you can search and find the material, then mark, then redact both the image and the text, then verify. Running OCR afterwards re-reads a document you have already finished.
How do I check that an OCR layer was redacted properly?
Extract text from the finished file and search for the terms you removed. If they appear, the image was redacted and the invisible text layer was not.
Doing it in Basalt
Basalt is a native macOS PDF toolkit with eighteen tools in one window: merge, split, compress, OCR, Bates numbering, privilege logs, forms, signing, watermarks, comparison and more. Redaction destroys content rather than covering it, and an independent verifier re-opens every written file and proves the material is gone before the file is saved. If it cannot prove it, nothing is written. A one time $29 licence covers up to three Macs, it is free for the first 24 hours, there is no account, and the engine holds no network entitlement at all, which macOS enforces at the code-signature level. Download it at basaltformac.com.
Redaction that proves itself
Basalt destroys the content you mark, then re-opens the file it wrote and proves the content is gone before it saves anything. Your documents never leave your Mac.