Scanned PDF vs searchable PDF: how to tell, and how to make a scan searchable

A scanned PDF is a photograph of a page. A searchable PDF has real text behind the page, so you can select it, search it and copy it. To turn the first into the second, you run OCR (optical character recognition: software that reads the letters in a picture). OCR adds the text as an invisible layer over the scan, so the page looks exactly as before.

In BuzPulse, the place to do this is Edit PDF. Open the scan, press Run OCR (English + Arabic), then download the file. The reading happens in your browser, and the scan is not uploaded.

You cannot tell by looking. A scan and a PDF exported from Word look the same on screen. Three quick tests work in any PDF viewer:

1. 1

   Try to select a word

   Drag across a line of text with the cursor. On a searchable page the words highlight one by one. On a scan, nothing highlights, or a whole block or the whole page turns blue at once.

2. 2

   Search for a word you can see

   Press Ctrl+F (⌘F on a Mac) and type a word from the page. "No results" for a word that is plainly there means there is no text layer.

3. 3

   Paste it somewhere

   Select what you can, copy it, and paste it into a notes app. Empty or garbled paste confirms it.

1. 1

   Open the scan in Edit PDF

   Choose the file on the Edit PDF page. When a page has no text layer, the editor tells you it has no editable text yet and offers OCR.

2. 2

   Press Run OCR (English + Arabic)

   The first time, the reading engine downloads (about 8 MB, then it is cached). After that, each scanned page is read in turn and you see "page 2 of 5" as it goes.

3. 3

   Read the result note

   When it finishes, the editor says how many lines of text it found. It also warns that OCR can misread, and that is worth taking seriously: check a few lines against the picture.

4. 4

   Download

   You do not need to edit anything. The lines OCR found that you did not change are saved into the download as invisible text, so the file can be searched and copied. Open the result and repeat the Ctrl+F test from above.

The same step lets you fix a mistake. Click a line OCR found, correct it, and your replacement is drawn over the scanned words. Lines you correct are drawn visibly, so only the lines you leave alone become invisible text.

The layer is only as good as the reading. Lines the reader is not confident about are left out rather than guessed, which means a faint or crooked scan can end up with a partly searchable page. Handwriting is not something to expect OCR to read.

OCR works from the picture, so a clearer picture helps. The PDF Enhancer redraws each page at a higher resolution and sharpens it. It does not add text: every page comes back as an image, so anything that was searchable before becomes part of the picture.

The order matters. If you enhance a scan, enhance first and run OCR on the enhanced file afterwards. Running OCR first and enhancing second throws the text layer away.

• It reads English and Arabic only. Other languages are not set up.
• It does not turn a scan into a Word or Excel file. The PDF to Word and PDF to Excel converters read the text layer, so they cannot read a scan, and OCR in Edit PDF does not change that.
• It does not guarantee accurate text. Check names, dates and amounts against the picture before relying on them.
• It skips rotated pages. The editor does not edit pages that are rotated, so turn them upright first with Rotate PDF and save.
• It does not rebuild the layout. Tables, columns and fonts stay a picture. You get lines of text, not a reformatted document.