Prerequisites

Updated on Jun 3, 2026

Optical Character Recognition (also known as OCR) is the electronic or mechanical conversion of images of typed, handwritten, or printed text into machine-encoded text from a scanned document.

This topic describes the requirements for the PdfProcessing library to use OcrFormatProvider.

The default Tesseract implementation is at this point Windows and Linux-only. You can still use the OCR feature with a custom implementation.

Used images must be 300 DPI for best results.

Required Packages

To use OcrFormatProvider, add the following packages:

.NET Framework	.NET Standard-compatible
Telerik.Windows.Documents.Core	Telerik.Documents.Core
Telerik.Windows.Documents.Fixed	Telerik.Documents.Fixed
Telerik.Windows.Documents.Fixed.FormatProviders.Ocr	Telerik.Documents.Fixed.FormatProviders.Ocr
Always add this reference as a NuGet package. It adds the required Tesseract references and files automatically. Otherwise, a manual setup might be required.
Telerik.Windows.Documents.TesseractOcr	Telerik.Documents.TesseractOcr
To export images different than Jpeg and Jpeg2000 or ImageQuality different than High, add a reference to the following assembly:
-	Telerik.Documents.ImageUtils
-	SkiaSharp Telerik.Documents.ImageUtils depends on SkiaSharp.
-	SkiaSharp.NativeAssets.* (version 3.119.1) May differ according to the used platform. For Linux (starting with Q2 2025) use SkiaSharp.NativeAssets.Linux.NoDependencies and execute the required commands.
-	SkiaSharp.Views.Blazor and wasm-tools For Blazor Web Assembly.

Verify that all Tesseract dependencies are properly set up.

Language Data Setup

Create a "tessdata" folder and populate it with the desired languages. The languages are in the form of .traineddata files and are crucial for Tesseract OCR because they contain the machine learning models that Tesseract uses to recognize text. English (eng.traineddata) is always required by default. You can download the language data files from the official Tesseract GitHub repository. Results may vary depending on the language version:

The "tessdata" folder placement is determined by the user. The DataPath property of the TesseractOcrProvider points to the parent folder that contains "tessdata", which allows the provider to locate and use it.

"tessdata" Structure:

plaintext

tessdata
├── due.traineddata
├── eng.traineddata     
└── spa.traineddata

Manually Set Up the Tesseract Native Assemblies

Verify that the following files exist in the root directory of your project:

• The Tesseract.dll assembly.

• The Tesseract native assemblies (x86, x64):

  

If these requirements are not met, go through the following steps:

1. Download the tesseract50.dll and leptonica-1.82.0.dll native assemblies from the listed links:
   • https://github.com/charlesw/tesseract/tree/master/src/Tesseract/x64.
   • https://github.com/charlesw/tesseract/tree/master/src/Tesseract/x86.
2. Create the following structure and add the two folders to the root of the application.
   • Folder Structure:
   plaintext

   RootFolder
   ├── x64
   │   ├── tesseract50.dll
   │   └── leptonica-1.82.0.dll
   └── x86
       ├── tesseract50.dll
       └── leptonica-1.82.0.dll

Linux-Specific Steps

Execute the following commands in the environment:

Ubuntu	Alpine	Fedora
sudo apt update	sudo apk update	sudo dnf install tesseract
sudo apt install tesseract-ocr	sudo apk add tesseract-ocr	sudo dnf install leptonica
sudo apt install libleptonica-dev	sudo apk add leptonica

If the generated tesseract/leptonica .so files cannot be found, they were likely installed with different names than expected. Copy their names and location, and set them to the corresponding properties:

• TesseractEnvironment.TesseractUnixLibName
• TesseractEnvironment.LeptonicaUnixLibName
• TesseractEnvironment.CustomSearchPath

See Also

• Using OcrFormatProvider
• Implementing a Custom OCR Provider
• Cross-Platform Images