How can I convert a scanned PDF into a searchable/text-based PDF in Calibre?

ehsangh70 · 05-28-2026, 04:40 AM

Hello everyone,

I have a scanned PDF file where the pages are basically images, and I want to convert it into a searchable/text-based PDF with selectable text.

Is there a way to do this using Calibre, or do I need an OCR plugin/tool alongside it?
If possible, could someone explain the best method or workflow?

Karellen · 05-28-2026, 05:13 AM

You can't. You need to OCR the document. You end up with text and then it is up to you what you do with it next - create an epub or a pdf.

Like all OCR, it is not perfect, so a proof read and formatting will be required.

The method I use is detailed here...
https://www.mobileread.com/forums/sh...3&postcount=23

Jaws · 05-29-2026, 02:04 AM

As an alternative to Karellen's method, I've had good success using PDF24, a free alternative for Windows systems at

https://tools.pdf24.org/en/ocr-pdf

There is both an online version (highly dependent upon how fast your connection is) and an installable program. On the good side, it's pretty decent with math symbols and superscripts and subscripts (fairly important in chemistry, for example); however, it's also quite font-dependent for non-Roman/Greek characters — it doesn't handle san serif well, and results are a lot better if the exact font used in the source is already installed (and active) in the \windows\font folder. Unfortunately, that means "exact" as in "same vendor, same character set" — "Minion Pro" is not the same as "Minion", even though they're from the same vendor.

That said, it's really essential if you're trying to get decent, reflowable, selectable-text output (whether as PDF or in another format) to use a specialist program. Calibre is many things, but not everything...

feuille · 05-29-2026, 03:00 AM

I have been using a plugin for this purpose for a while now—one that I haven't released yet because I still want to add a proofreading step. It is based on OCRmyPDF, which, in turn, relies on Tesseract for the OCR component. If anyone wants to try this out, I could release it even without the proofreading feature (which utilizes the hOCR files generated by Tesseract). For high-quality scans, the results are already excellent out-of-the-box.

rantanplan · 05-29-2026, 03:15 AM

I’d love to have that as one of my projects is to scan the original German translation for the A Song of Ice and Fire books.

feuille · 06-09-2026, 07:48 AM

OCR plugin is released: https://www.mobileread.com/forums/sh...76#post4590976

05-28-2026, 04:40 AM	#1
ehsangh70 Junior Member Posts: 2 Karma: 10 Join Date: May 2026 Device: PC	How can I convert a scanned PDF into a searchable/text-based PDF in Calibre? Hello everyone, I have a scanned PDF file where the pages are basically images, and I want to convert it into a searchable/text-based PDF with selectable text. Is there a way to do this using Calibre, or do I need an OCR plugin/tool alongside it? If possible, could someone explain the best method or workflow?

05-29-2026, 03:00 AM	#4
feuille Connoisseur Posts: 70 Karma: 666 Join Date: May 2020 Location: Germany Device: android smartphone + tablet with Moon Reader and ReadEra Apps.	OCRthisPDF (not yet published( I have been using a plugin for this purpose for a while now—one that I haven't released yet because I still want to add a proofreading step. It is based on OCRmyPDF, which, in turn, relies on Tesseract for the OCR component. If anyone wants to try this out, I could release it even without the proofreading feature (which utilizes the hOCR files generated by Tesseract). For high-quality scans, the results are already excellent out-of-the-box.

06-09-2026, 07:48 AM	#6
feuille Connoisseur Posts: 70 Karma: 666 Join Date: May 2020 Location: Germany Device: android smartphone + tablet with Moon Reader and ReadEra Apps.	OCRthisPDF OCR plugin is released: https://www.mobileread.com/forums/sh...76#post4590976

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Convert from Text(Markdown) to PDF: Can not copy text in the PDF	hellomichibye	Conversion	2	08-26-2019 12:29 PM
Non-searchable text and Scroll-bar inside the converted PDF output, unusual? Why?	Amortization	Calibre	6	05-07-2017 01:53 AM
Convert epub to pdf, with notes with main text in the pdf?	8140david	ePub	1	06-18-2015 01:13 PM
Urgent: How To Convert Wikibook PDF Into a Searchable Index?	deerayolia	Kindle Formats	4	05-28-2012 06:52 AM

05-28-2026, 05:13 AM	#2
Karellen Wizard Posts: 1,856 Karma: 9600930 Join Date: Sep 2021 Location: Australia Device: Kobo Libra 2	You can't. You need to OCR the document. You end up with text and then it is up to you what you do with it next - create an epub or a pdf. Like all OCR, it is not perfect, so a proof read and formatting will be required. The method I use is detailed here... https://www.mobileread.com/forums/sh...3&postcount=23

05-29-2026, 02:04 AM	#3
Jaws JCL Punch-Card Collector Posts: 109 Karma: 606560 Join Date: Jun 2014 Location: Antarctica Device: Aggressively Device Independent	As an alternative to Karellen's method, I've had good success using PDF24, a free alternative for Windows systems at https://tools.pdf24.org/en/ocr-pdf There is both an online version (highly dependent upon how fast your connection is) and an installable program. On the good side, it's pretty decent with math symbols and superscripts and subscripts (fairly important in chemistry, for example); however, it's also quite font-dependent for non-Roman/Greek characters — it doesn't handle san serif well, and results are a lot better if the exact font used in the source is already installed (and active) in the \windows\font folder. Unfortunately, that means "exact" as in "same vendor, same character set" — "Minion Pro" is not the same as "Minion", even though they're from the same vendor. That said, it's really essential if you're trying to get decent, reflowable, selectable-text output (whether as PDF or in another format) to use a specialist program. Calibre is many things, but not everything...

05-29-2026, 03:15 AM	#5
rantanplan Weirdo Posts: 1,124 Karma: 12503116 Join Date: Nov 2019 Location: Wuppertal, Germany Device: Kobo Sage, Kobo Libra 2, reMarkable PaperPro	I’d love to have that as one of my projects is to scan the original German translation for the A Song of Ice and Fire books.