OCR & Transcription

Detectors work on text. A lot of important content, though, isn’t text to begin with — it’s locked inside scanned PDFs, screenshots, images, and audio or video files. Two optional switches let Classifyre unlock that content so your detectors can see it.

By default both are off: Classifyre reads the text that’s already text and leaves the rest as metadata-only. Turn these on per source when there’s value hidden in non-text content.

OCR — read images and documents

OCR (Optical Character Recognition) extracts readable text from images and supported binary documents before the text-capable detectors run.

Turn it on when a source holds content like:

• Scanned PDFs and faxes
• Screenshots and photos of documents
• Images that may contain text (whiteboards, ID cards, forms)

With OCR on, a screenshot that contains a leaked API key or a scanned form with personal data becomes searchable by your detectors — instead of slipping through as “just an image.”

Trade-off: OCR does extra work per image, so scans take a little longer. Enable it on sources where image content genuinely matters, and leave it off where everything is already text.

Transcription — turn speech into text

Transcription converts the speech in audio and video files into text, so detectors can run over what was said, not just the file’s metadata.

Turn it on when a source holds content like:

• Meeting or call recordings
• Video posts and webinars
• Voice notes and podcasts

With transcription on, a recorded meeting where someone reads out a customer’s details becomes text your detectors can flag.

Trade-off: Transcription is the heaviest content step — it’s noticeably slower than reading text or doing OCR, and it relies on the transcription capability being available in your deployment. Reserve it for sources where spoken content is worth the extra processing.

How they fit with the rest of a source

• Both are part of a source’s sampling settings — they decide what content is read from each sampled item, while sampling decides which items are read.
• Extracted text flows into exactly the same detectors as ordinary text — there’s nothing extra to configure on the detector side.
• They’re independent: turn on either, both, or neither.

Next: confirm a source connects, and put scans on a schedule — Testing & Scheduling.