# How to redact a scanned PDF (and the hidden text most tools miss)

> A scanned page is a picture, so search finds nothing until OCR, and many scans hide a text layer behind the image. How to redact scanned PDFs properly.

Source: https://redactonce.com/blog/redact-scanned-pdf
Published: 2026-11-19
Publisher: RedactOnce (https://redactonce.com)

## In short

- On an image-only scan, search finds nothing until OCR reads the page and records where each word is.
- Many scans carry an invisible OCR text layer: covering the image leaves that text searchable and copyable.
- OCR misreads characters, so review close matches and box handwriting, signatures, stamps and photos by hand.
- RedactOnce’s free tool offers on-device English OCR and manual boxes, with fresh OCR verification before exporting recognized scans.

Scanned letters, signed contracts and photographed forms are some of the most sensitive documents people redact, and the easiest to get wrong. A scan behaves differently from a PDF made on a computer, in two ways that matter for redaction.

## Is it a scan? A quick test

Open the PDF and try to select a word.

- **Nothing selects**: the page is only an image. Search finds nothing, so you, or OCR, have to find every sensitive value.
- **Words select, but the highlight sits slightly off**: it’s a scan with an OCR text layer, a “searchable PDF”. The hidden text needs redacting as well as the image.
- **Words select neatly**: it’s probably a digital PDF. See [how to redact a PDF](https://redactonce.com/blog/how-to-redact-a-pdf).

## Problem 1: you can’t search an image

On an image-only page the computer sees pixels, not letters. Finding every mention of a name means reading every page yourself, and on page forty of a long file that is where mistakes happen.

Optical character recognition (OCR) reads the image and records each word with its position on the page. Once a scan has been through OCR, names and numbers can be matched and boxed automatically, the same way as in a digital PDF.

## Problem 2: the text you can’t see

Many scanners, and tools that “make PDFs searchable”, store OCR text in an invisible layer behind the image. A black box drawn over the image hides the pixels, but the invisible words stay: search finds them and copy-paste reveals them.

> **Covering the picture isn’t enough**
>
> On a searchable scan, a redaction has to remove the area from the image and the matching words from the hidden text layer. If your tool only paints over the image, search the result for what you removed.

## Problem 3: OCR makes mistakes

OCR misreads characters, especially on faint, skewed or low-resolution scans: `0` becomes `O`, `1` becomes `l`, and a name breaks across two words. Three habits help:

- Use matching that tolerates small spelling differences, then review the close matches yourself.
- Draw boxes by hand over handwriting, signatures, stamps and photos. OCR doesn’t read handwriting reliably.
- Scan at 300 dpi, straight and with good contrast. It improves recognition more than any setting.

## How RedactOnce handles scans

The [free tool](https://redactonce.com/redact-scanned-pdf) runs in your browser with optional English OCR, so names and numbers can be found on scans. You can also draw boxes over what to remove; the redacted areas are removed from the page, image and text layer alike.

In a bundle, scanned pages are read with OCR (English) on our servers. Names, account numbers, emails and the other built-in kinds are then found on scans just as in digital PDFs, and close spellings are listed for you to decide. Every redacted page is rebuilt from an image with the boxes painted in, plus an invisible text layer holding only the words that weren’t redacted, so the result stays searchable and nothing removed survives. Each export is read back and checked before you can download it.

**Redact a scan in your browser** Read scanned pages in English on your device, or draw boxes over what to remove. Review every mark. [Redact a scanned PDF](https://redactonce.com/redact-scanned-pdf)

## Checklist for scanned documents

1. Find out whether the scan has a hidden text layer.
2. Run OCR, or plan to read every page yourself.
3. Review close matches: OCR errors hide real names.
4. Box handwriting, signatures, stamps and photos by hand.
5. After redacting, search the result and copy its text to check nothing remains.

## Questions and answers

### Can you redact a scanned PDF?

Yes. Either draw boxes over what to remove by hand, or run OCR so names and numbers can be found automatically. Either way, the redacted areas must be removed from the image and from any hidden text layer.

### Why can I select text in a scanned PDF?

The scanner or a later step ran OCR and added an invisible text layer behind the page image. That text has to be redacted too, or it can still be copied and searched.
