Migrating from Tesseract to IronOCR

The Fundamental Problem

Every Tesseract application starts the same way: specifying a tessdata path that must be correct on every machine the application runs on.

Tesseract Approach:

// TessDataPath must exist and be populated — breaks on first clean deployment
private const string TessDataPath = @"./tessdata";

public static string ExtractText(string imagePath)
{
    // Runtime failure if eng.traineddata is missing from TessDataPath
    if (!Directory.Exists(TessDataPath))
        throw new DirectoryNotFoundException(
            $"Tessdata not found at {TessDataPath}. " +
            "Download from https://github.com/tesseract-ocr/tessdata");

    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var img   = Pix.LoadFromFile(imagePath);  // Leptonica Pix object
    using var page  = engine.Process(img);
    return page.GetText();
}

// TessDataPath must exist and be populated — breaks on first clean deployment
private const string TessDataPath = @"./tessdata";

public static string ExtractText(string imagePath)
{
    // Runtime failure if eng.traineddata is missing from TessDataPath
    if (!Directory.Exists(TessDataPath))
        throw new DirectoryNotFoundException(
            $"Tessdata not found at {TessDataPath}. " +
            "Download from https://github.com/tesseract-ocr/tessdata");

    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var img   = Pix.LoadFromFile(imagePath);  // Leptonica Pix object
    using var page  = engine.Process(img);
    return page.GetText();
}

Imports System.IO
Imports Tesseract

' TessDataPath must exist and be populated — breaks on first clean deployment
Private Const TessDataPath As String = "./tessdata"

Public Shared Function ExtractText(imagePath As String) As String
    ' Runtime failure if eng.traineddata is missing from TessDataPath
    If Not Directory.Exists(TessDataPath) Then
        Throw New DirectoryNotFoundException(
            $"Tessdata not found at {TessDataPath}. " &
            "Download from https://github.com/tesseract-ocr/tessdata")
    End If

    Using engine As New TesseractEngine(TessDataPath, "eng", EngineMode.Default)
        Using img As Pix = Pix.LoadFromFile(imagePath) ' Leptonica Pix object
            Using page As Page = engine.Process(img)
                Return page.GetText()
            End Using
        End Using
    End Using
End Function

IronOCR Approach:

// No tessdata folder. No path. No file check. Just OCR.
var text = new IronTesseract().Read("document.jpg").Text;

// No tessdata folder. No path. No file check. Just OCR.
var text = new IronTesseract().Read("document.jpg").Text;

' No tessdata folder. No path. No file check. Just OCR.
Dim text As String = New IronTesseract().Read("document.jpg").Text

The entire TessDataPath constant, the Directory.Exists guard, the Pix object, and the three-level using nesting disappear. Language data is embedded in the NuGet package.

Step 1: Replace NuGet Package

Remove the charlesw Tesseract wrapper:

dotnet remove package Tesseract

dotnet remove package Tesseract

Install IronOCR from NuGet:

dotnet add package IronOcr

Language packs install as separate packages when needed:

dotnet add package IronOcr.Languages.French, IronOcr.Languages.German

Step 2: Update Namespaces

Replace the Tesseract namespace with the IronOCR namespace:

// Before
using Tesseract;

// After
using IronOcr;

// Before
using Tesseract;

// After
using IronOcr;

Imports Tesseract

' After
Imports IronOcr

Step 3: Initialize License

Add license initialization once at application startup, before any IronTesseract calls:

IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY";

IronOcr.License.LicenseKey = "YOUR-LICENSE-KEY"

A free trial runs without a key during development. Production deployments require a valid key from the licensing page.

Tessdata Path Elimination and Engine Initialization

The most immediate change is removing TesseractEngine initialization and all the tessdata validation code that surrounds it.

Tesseract Approach:

// Every class that uses OCR must handle this initialization block
private const string TessDataPath = @"./tessdata";

public string RecognizeInvoiceNumber(string imagePath)
{
    // Check tessdata presence — missing file = silent runtime failure
    foreach (var lang in new[] { "eng" })
    {
        if (!File.Exists(Path.Combine(TessDataPath, $"{lang}.traineddata")))
            throw new FileNotFoundException(
                $"Missing {lang}.traineddata. " +
                "Download from https://github.com/tesseract-ocr/tessdata");
    }

    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);

    // Pix is a Leptonica wrapper type — not a standard .NET image
    using var img  = Pix.LoadFromFile(imagePath);
    using var page = engine.Process(img);

    string text = page.GetText();
    float  conf = page.GetMeanConfidence();

    return conf > 0.7f ? text : string.Empty;
}

// Every class that uses OCR must handle this initialization block
private const string TessDataPath = @"./tessdata";

public string RecognizeInvoiceNumber(string imagePath)
{
    // Check tessdata presence — missing file = silent runtime failure
    foreach (var lang in new[] { "eng" })
    {
        if (!File.Exists(Path.Combine(TessDataPath, $"{lang}.traineddata")))
            throw new FileNotFoundException(
                $"Missing {lang}.traineddata. " +
                "Download from https://github.com/tesseract-ocr/tessdata");
    }

    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);

    // Pix is a Leptonica wrapper type — not a standard .NET image
    using var img  = Pix.LoadFromFile(imagePath);
    using var page = engine.Process(img);

    string text = page.GetText();
    float  conf = page.GetMeanConfidence();

    return conf > 0.7f ? text : string.Empty;
}

Imports System.IO
Imports Tesseract

' Every class that uses OCR must handle this initialization block
Private Const TessDataPath As String = "./tessdata"

Public Function RecognizeInvoiceNumber(imagePath As String) As String
    ' Check tessdata presence — missing file = silent runtime failure
    For Each lang In New String() {"eng"}
        If Not File.Exists(Path.Combine(TessDataPath, $"{lang}.traineddata")) Then
            Throw New FileNotFoundException(
                $"Missing {lang}.traineddata. " &
                "Download from https://github.com/tesseract-ocr/tessdata")
        End If
    Next

    Using engine As New TesseractEngine(TessDataPath, "eng", EngineMode.Default)
        ' Pix is a Leptonica wrapper type — not a standard .NET image
        Using img As Pix = Pix.LoadFromFile(imagePath)
            Using page As Page = engine.Process(img)
                Dim text As String = page.GetText()
                Dim conf As Single = page.GetMeanConfidence()

                Return If(conf > 0.7F, text, String.Empty)
            End Using
        End Using
    End Using
End Function

IronOCR Approach:

using IronOcr;

public string RecognizeInvoiceNumber(string imagePath)
{
    var result = new IronTesseract().Read(imagePath);

    // Confidence property returns 0-100 double
    return result.Confidence > 70 ? result.Text : string.Empty;
}

using IronOcr;

public string RecognizeInvoiceNumber(string imagePath)
{
    var result = new IronTesseract().Read(imagePath);

    // Confidence property returns 0-100 double
    return result.Confidence > 70 ? result.Text : string.Empty;
}

Imports IronOcr

Public Function RecognizeInvoiceNumber(imagePath As String) As String
    Dim result = New IronTesseract().Read(imagePath)

    ' Confidence property returns 0-100 double
    Return If(result.Confidence > 70, result.Text, String.Empty)
End Function

The FileNotFoundException guard, the tessdata constant, the Pix object, and the three-level nesting are gone. IronTesseract is constructed without arguments because language data is embedded. See the IronTesseract setup guide for configuration options when you need non-default behavior, and the confidence scores guide for the full confidence API.

Multi-Page TIFF Processing with Preprocessing Pipeline

Multi-frame TIFF files — common in scanned document archives and fax systems — require explicit frame iteration with Tesseract. IronOCR loads all frames in one call and applies the preprocessing pipeline uniformly.

Tesseract Approach:

using Tesseract;
using System.Drawing;
using System.Drawing.Imaging;

private const string TessDataPath = @"./tessdata";

public static string ExtractFromMultiPageTiff(string tiffPath)
{
    var allText = new System.Text.StringBuilder();

    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var tiffImage = Image.FromFile(tiffPath);

    int frameCount = tiffImage.GetFrameCount(FrameDimension.Page);

    for (int i = 0; i < frameCount; i++)
    {
        tiffImage.SelectActiveFrame(FrameDimension.Page, i);

        // Must save each frame to disk — Pix.LoadFromFile requires a path
        string tempPath = Path.GetTempFileName() + ".png";
        try
        {
            tiffImage.Save(tempPath, ImageFormat.Png);

            using var img  = Pix.LoadFromFile(tempPath);
            using var page = engine.Process(img);
            allText.AppendLine(page.GetText());
        }
        finally
        {
            File.Delete(tempPath); // Uncleaned temp files fill disk on failure
        }
    }

    return allText.ToString();
}

using Tesseract;
using System.Drawing;
using System.Drawing.Imaging;

private const string TessDataPath = @"./tessdata";

public static string ExtractFromMultiPageTiff(string tiffPath)
{
    var allText = new System.Text.StringBuilder();

    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var tiffImage = Image.FromFile(tiffPath);

    int frameCount = tiffImage.GetFrameCount(FrameDimension.Page);

    for (int i = 0; i < frameCount; i++)
    {
        tiffImage.SelectActiveFrame(FrameDimension.Page, i);

        // Must save each frame to disk — Pix.LoadFromFile requires a path
        string tempPath = Path.GetTempFileName() + ".png";
        try
        {
            tiffImage.Save(tempPath, ImageFormat.Png);

            using var img  = Pix.LoadFromFile(tempPath);
            using var page = engine.Process(img);
            allText.AppendLine(page.GetText());
        }
        finally
        {
            File.Delete(tempPath); // Uncleaned temp files fill disk on failure
        }
    }

    return allText.ToString();
}

Imports Tesseract
Imports System.Drawing
Imports System.Drawing.Imaging
Imports System.IO
Imports System.Text

Private Const TessDataPath As String = "./tessdata"

Public Shared Function ExtractFromMultiPageTiff(tiffPath As String) As String
    Dim allText As New StringBuilder()

    Using engine As New TesseractEngine(TessDataPath, "eng", EngineMode.Default)
        Using tiffImage As Image = Image.FromFile(tiffPath)
            Dim frameCount As Integer = tiffImage.GetFrameCount(FrameDimension.Page)

            For i As Integer = 0 To frameCount - 1
                tiffImage.SelectActiveFrame(FrameDimension.Page, i)

                ' Must save each frame to disk — Pix.LoadFromFile requires a path
                Dim tempPath As String = Path.GetTempFileName() & ".png"
                Try
                    tiffImage.Save(tempPath, ImageFormat.Png)

                    Using img As Pix = Pix.LoadFromFile(tempPath)
                        Using page As Page = engine.Process(img)
                            allText.AppendLine(page.GetText())
                        End Using
                    End Using
                Finally
                    File.Delete(tempPath) ' Uncleaned temp files fill disk on failure
                End Try
            Next
        End Using
    End Using

    Return allText.ToString()
End Function

IronOCR Approach:

using IronOcr;

public static string ExtractFromMultiPageTiff(string tiffPath)
{
    using var input = new OcrInput();
    input.LoadImageFrames(tiffPath);  // Loads all frames at once
    input.Deskew();                   // Applied to every frame uniformly
    input.DeNoise();

    var result = new IronTesseract().Read(input);
    return result.Text;
}

using IronOcr;

public static string ExtractFromMultiPageTiff(string tiffPath)
{
    using var input = new OcrInput();
    input.LoadImageFrames(tiffPath);  // Loads all frames at once
    input.Deskew();                   // Applied to every frame uniformly
    input.DeNoise();

    var result = new IronTesseract().Read(input);
    return result.Text;
}

Imports IronOcr

Public Shared Function ExtractFromMultiPageTiff(tiffPath As String) As String
    Using input As New OcrInput()
        input.LoadImageFrames(tiffPath)  ' Loads all frames at once
        input.Deskew()                   ' Applied to every frame uniformly
        input.DeNoise()

        Dim result = New IronTesseract().Read(input)
        Return result.Text
    End Using
End Function

No frame iteration. No temp file creation. No cleanup logic. The preprocessing pipeline applies to every frame without an additional loop. The TIFF and GIF input guide covers multi-frame handling in detail, including selective frame ranges for large archive files.

Searchable PDF Generation

Converting a scanned PDF into a searchable PDF requires Tesseract to render each page to an image (via an external PDF library), run OCR, then reconstruct a PDF with a text layer — a multi-library, multi-step process. IronOCR handles input, OCR, and output in a single pipeline.

Tesseract Approach:

// Requires: PdfiumViewer + Tesseract + a PDF writer library (iText, PdfSharp)
// Each library adds its own native dependencies and license considerations

using Tesseract;
// using PdfiumViewer;  // Comment: must add NuGet + deploy native pdfium.dll
// using iText.Kernel.Pdf;  // Comment: AGPL or commercial license required

private const string TessDataPath = @"./tessdata";

public static void CreateSearchablePdf(string inputPdfPath, string outputPdfPath)
{
    // Step 1: Render PDF pages to images (requires PdfiumViewer)
    // Step 2: Run OCR on each image (Tesseract)
    // Step 3: Write text positions back into PDF (requires iText or PDFsharp)
    //
    // Total: ~150 lines across three libraries
    // Native binaries required: tesseract*.dll, leptonica*.dll, pdfium.dll
    // License risk: iText is AGPL unless you purchase a commercial license

    throw new NotImplementedException(
        "Requires PdfiumViewer + Tesseract + a PDF writer. " +
        "No single-package solution exists with this stack.");
}

// Requires: PdfiumViewer + Tesseract + a PDF writer library (iText, PdfSharp)
// Each library adds its own native dependencies and license considerations

using Tesseract;
// using PdfiumViewer;  // Comment: must add NuGet + deploy native pdfium.dll
// using iText.Kernel.Pdf;  // Comment: AGPL or commercial license required

private const string TessDataPath = @"./tessdata";

public static void CreateSearchablePdf(string inputPdfPath, string outputPdfPath)
{
    // Step 1: Render PDF pages to images (requires PdfiumViewer)
    // Step 2: Run OCR on each image (Tesseract)
    // Step 3: Write text positions back into PDF (requires iText or PDFsharp)
    //
    // Total: ~150 lines across three libraries
    // Native binaries required: tesseract*.dll, leptonica*.dll, pdfium.dll
    // License risk: iText is AGPL unless you purchase a commercial license

    throw new NotImplementedException(
        "Requires PdfiumViewer + Tesseract + a PDF writer. " +
        "No single-package solution exists with this stack.");
}

Imports Tesseract
' Imports PdfiumViewer  ' Comment: must add NuGet + deploy native pdfium.dll
' Imports iText.Kernel.Pdf  ' Comment: AGPL or commercial license required

Private Const TessDataPath As String = "./tessdata"

Public Shared Sub CreateSearchablePdf(inputPdfPath As String, outputPdfPath As String)
    ' Step 1: Render PDF pages to images (requires PdfiumViewer)
    ' Step 2: Run OCR on each image (Tesseract)
    ' Step 3: Write text positions back into PDF (requires iText or PDFsharp)
    '
    ' Total: ~150 lines across three libraries
    ' Native binaries required: tesseract*.dll, leptonica*.dll, pdfium.dll
    ' License risk: iText is AGPL unless you purchase a commercial license

    Throw New NotImplementedException("Requires PdfiumViewer + Tesseract + a PDF writer. " & _
                                      "No single-package solution exists with this stack.")
End Sub

IronOCR Approach:

using IronOcr;

public static void CreateSearchablePdf(string inputPdfPath, string outputPdfPath)
{
    using var input = new OcrInput();
    input.LoadPdf(inputPdfPath);
    input.Deskew();    // Correct scanned page skew before OCR
    input.DeNoise();   // Remove scanner artifacts

    var result = new IronTesseract().Read(input);
    result.SaveAsSearchablePdf(outputPdfPath);
}

using IronOcr;

public static void CreateSearchablePdf(string inputPdfPath, string outputPdfPath)
{
    using var input = new OcrInput();
    input.LoadPdf(inputPdfPath);
    input.Deskew();    // Correct scanned page skew before OCR
    input.DeNoise();   // Remove scanner artifacts

    var result = new IronTesseract().Read(input);
    result.SaveAsSearchablePdf(outputPdfPath);
}

Imports IronOcr

Public Shared Sub CreateSearchablePdf(inputPdfPath As String, outputPdfPath As String)
    Using input As New OcrInput()
        input.LoadPdf(inputPdfPath)
        input.Deskew() ' Correct scanned page skew before OCR
        input.DeNoise() ' Remove scanner artifacts

        Dim result = New IronTesseract().Read(input)
        result.SaveAsSearchablePdf(outputPdfPath)
    End Using
End Sub

One method call produces the searchable PDF with an embedded text layer. No external PDF library, no native pdfium binary, no license entanglement with AGPL dependencies. The searchable PDF how-to guide documents the output format, and the PDF OCR example walks through a complete scanned document pipeline. For broader context on what IronOCR can do with PDF input, the PDF OCR use case page covers production architecture patterns.

Structured Data Extraction from Scanned Documents

Tesseract exposes word-level data through ResultIterator, which requires a do/while loop with manual bounding box extraction. IronOCR exposes a document hierarchy — pages, paragraphs, lines, words — as strongly-typed collections with coordinates already populated.

Tesseract Approach:

using Tesseract;

private const string TessDataPath = @"./tessdata";

public static void ExtractStructuredData(string imagePath)
{
    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var img    = Pix.LoadFromFile(imagePath);
    using var page   = engine.Process(img);
    using var iter   = page.GetIterator();

    iter.Begin();
    do
    {
        if (iter.IsAtBeginningOf(PageIteratorLevel.Para))
            Console.WriteLine("-- New Paragraph --");

        if (iter.TryGetBoundingBox(PageIteratorLevel.Word, out var bounds))
        {
            string word       = iter.GetText(PageIteratorLevel.Word);
            float  confidence = iter.GetConfidence(PageIteratorLevel.Word);
            Console.WriteLine(
                $"Word: '{word?.Trim()}' " +
                $"at ({bounds.X1},{bounds.Y1})-({bounds.X2},{bounds.Y2}) " +
                $"conf={confidence:P0}");
        }
    }
    while (iter.Next(PageIteratorLevel.Word));
}

using Tesseract;

private const string TessDataPath = @"./tessdata";

public static void ExtractStructuredData(string imagePath)
{
    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var img    = Pix.LoadFromFile(imagePath);
    using var page   = engine.Process(img);
    using var iter   = page.GetIterator();

    iter.Begin();
    do
    {
        if (iter.IsAtBeginningOf(PageIteratorLevel.Para))
            Console.WriteLine("-- New Paragraph --");

        if (iter.TryGetBoundingBox(PageIteratorLevel.Word, out var bounds))
        {
            string word       = iter.GetText(PageIteratorLevel.Word);
            float  confidence = iter.GetConfidence(PageIteratorLevel.Word);
            Console.WriteLine(
                $"Word: '{word?.Trim()}' " +
                $"at ({bounds.X1},{bounds.Y1})-({bounds.X2},{bounds.Y2}) " +
                $"conf={confidence:P0}");
        }
    }
    while (iter.Next(PageIteratorLevel.Word));
}

Imports Tesseract

Private Const TessDataPath As String = "./tessdata"

Public Shared Sub ExtractStructuredData(imagePath As String)
    Using engine As New TesseractEngine(TessDataPath, "eng", EngineMode.Default)
        Using img As Pix = Pix.LoadFromFile(imagePath)
            Using page As Page = engine.Process(img)
                Using iter As ResultIterator = page.GetIterator()

                    iter.Begin()
                    Do
                        If iter.IsAtBeginningOf(PageIteratorLevel.Para) Then
                            Console.WriteLine("-- New Paragraph --")
                        End If

                        Dim bounds As Rect
                        If iter.TryGetBoundingBox(PageIteratorLevel.Word, bounds) Then
                            Dim word As String = iter.GetText(PageIteratorLevel.Word)
                            Dim confidence As Single = iter.GetConfidence(PageIteratorLevel.Word)
                            Console.WriteLine($"Word: '{word?.Trim()}' " &
                                              $"at ({bounds.X1},{bounds.Y1})-({bounds.X2},{bounds.Y2}) " &
                                              $"conf={confidence:P0}")
                        End If
                    Loop While iter.Next(PageIteratorLevel.Word)
                End Using
            End Using
        End Using
    End Using
End Sub

IronOCR Approach:

using IronOcr;

public static void ExtractStructuredData(string imagePath)
{
    var result = new IronTesseract().Read(imagePath);

    foreach (var page in result.Pages)
    {
        Console.WriteLine($"Page {page.PageNumber} — confidence: {result.Confidence}%");

        foreach (var paragraph in page.Paragraphs)
        {
            Console.WriteLine($"  Paragraph at ({paragraph.X},{paragraph.Y}):");
            Console.WriteLine($"  {paragraph.Text}");

            foreach (var word in paragraph.Words)
            {
                Console.WriteLine(
                    $"    Word: '{word.Text}' " +
                    $"at ({word.X},{word.Y}) " +
                    $"size {word.Width}x{word.Height} " +
                    $"conf={word.Confidence:P0}");
            }
        }
    }
}

using IronOcr;

public static void ExtractStructuredData(string imagePath)
{
    var result = new IronTesseract().Read(imagePath);

    foreach (var page in result.Pages)
    {
        Console.WriteLine($"Page {page.PageNumber} — confidence: {result.Confidence}%");

        foreach (var paragraph in page.Paragraphs)
        {
            Console.WriteLine($"  Paragraph at ({paragraph.X},{paragraph.Y}):");
            Console.WriteLine($"  {paragraph.Text}");

            foreach (var word in paragraph.Words)
            {
                Console.WriteLine(
                    $"    Word: '{word.Text}' " +
                    $"at ({word.X},{word.Y}) " +
                    $"size {word.Width}x{word.Height} " +
                    $"conf={word.Confidence:P0}");
            }
        }
    }
}

Imports IronOcr

Public Module Extractor
    Public Sub ExtractStructuredData(imagePath As String)
        Dim result = New IronTesseract().Read(imagePath)

        For Each page In result.Pages
            Console.WriteLine($"Page {page.PageNumber} — confidence: {result.Confidence}%")

            For Each paragraph In page.Paragraphs
                Console.WriteLine($"  Paragraph at ({paragraph.X},{paragraph.Y}):")
                Console.WriteLine($"  {paragraph.Text}")

                For Each word In paragraph.Words
                    Console.WriteLine(
                        $"    Word: '{word.Text}' " &
                        $"at ({word.X},{word.Y}) " &
                        $"size {word.Width}x{word.Height} " &
                        $"conf={word.Confidence:P0}")
                Next
            Next
        Next
    End Sub
End Module

The ResultIterator loop disappears entirely. The document hierarchy is a set of enumerable collections — no iterator state, no manual level tracking, no bounding box extraction by output parameter. Each word object carries its own coordinates and confidence. The read results guide documents every level of the hierarchy, and the OcrResult API reference lists all available properties.

Multi-Language OCR Without Tessdata File Management

Adding a language to a Tesseract application means downloading a .traineddata file, placing it in the tessdata folder, updating every deployment manifest that includes that folder, and modifying the engine initialization string. With IronOCR, it is a single NuGet package reference.

Tesseract Approach:

using Tesseract;

private const string TessDataPath = @"./tessdata";

public static string ExtractFromEuropeanDocument(string imagePath)
{
    // Before this call works, these files must exist:
    // ./tessdata/eng.traineddata  (~15 MB, from GitHub)
    // ./tessdata/fra.traineddata  (~15 MB, from GitHub)
    // ./tessdata/deu.traineddata  (~15 MB, from GitHub)
    // ./tessdata/spa.traineddata  (~15 MB, from GitHub)
    // Total: ~60 MB to download, version-match, and deploy to every environment

    foreach (var lang in new[] { "eng", "fra", "deu", "spa" })
    {
        if (!File.Exists(Path.Combine(TessDataPath, $"{lang}.traineddata")))
            throw new FileNotFoundException(
                $"Download {lang}.traineddata from " +
                "https://github.com/tesseract-ocr/tessdata " +
                $"and place in {TessDataPath}");
    }

    // Language string is a concatenation — order affects recognition priority
    using var engine = new TesseractEngine(TessDataPath, "eng+fra+deu+spa", EngineMode.Default);
    using var img    = Pix.LoadFromFile(imagePath);
    using var page   = engine.Process(img);
    return page.GetText();
}

using Tesseract;

private const string TessDataPath = @"./tessdata";

public static string ExtractFromEuropeanDocument(string imagePath)
{
    // Before this call works, these files must exist:
    // ./tessdata/eng.traineddata  (~15 MB, from GitHub)
    // ./tessdata/fra.traineddata  (~15 MB, from GitHub)
    // ./tessdata/deu.traineddata  (~15 MB, from GitHub)
    // ./tessdata/spa.traineddata  (~15 MB, from GitHub)
    // Total: ~60 MB to download, version-match, and deploy to every environment

    foreach (var lang in new[] { "eng", "fra", "deu", "spa" })
    {
        if (!File.Exists(Path.Combine(TessDataPath, $"{lang}.traineddata")))
            throw new FileNotFoundException(
                $"Download {lang}.traineddata from " +
                "https://github.com/tesseract-ocr/tessdata " +
                $"and place in {TessDataPath}");
    }

    // Language string is a concatenation — order affects recognition priority
    using var engine = new TesseractEngine(TessDataPath, "eng+fra+deu+spa", EngineMode.Default);
    using var img    = Pix.LoadFromFile(imagePath);
    using var page   = engine.Process(img);
    return page.GetText();
}

Imports Tesseract
Imports System.IO

Private Const TessDataPath As String = "./tessdata"

Public Shared Function ExtractFromEuropeanDocument(imagePath As String) As String
    ' Before this call works, these files must exist:
    ' ./tessdata/eng.traineddata  (~15 MB, from GitHub)
    ' ./tessdata/fra.traineddata  (~15 MB, from GitHub)
    ' ./tessdata/deu.traineddata  (~15 MB, from GitHub)
    ' ./tessdata/spa.traineddata  (~15 MB, from GitHub)
    ' Total: ~60 MB to download, version-match, and deploy to every environment

    For Each lang In New String() {"eng", "fra", "deu", "spa"}
        If Not File.Exists(Path.Combine(TessDataPath, $"{lang}.traineddata")) Then
            Throw New FileNotFoundException(
                $"Download {lang}.traineddata from " &
                "https://github.com/tesseract-ocr/tessdata " &
                $"and place in {TessDataPath}")
        End If
    Next

    ' Language string is a concatenation — order affects recognition priority
    Using engine As New TesseractEngine(TessDataPath, "eng+fra+deu+spa", EngineMode.Default)
        Using img As Pix = Pix.LoadFromFile(imagePath)
            Using page As Page = engine.Process(img)
                Return page.GetText()
            End Using
        End Using
    End Using
End Function

IronOCR Approach:

// Install language packs once per project:
// dotnet add package IronOcr.Languages.French
// dotnet add package IronOcr.Languages.German
// dotnet add package IronOcr.Languages.Spanish
using IronOcr;

public static string ExtractFromEuropeanDocument(string imagePath)
{
    var ocr = new IronTesseract();
    ocr.Language = OcrLanguage.English;
    ocr.AddSecondaryLanguage(OcrLanguage.French);
    ocr.AddSecondaryLanguage(OcrLanguage.German);
    ocr.AddSecondaryLanguage(OcrLanguage.Spanish);

    return ocr.Read(imagePath).Text;
}

// Install language packs once per project:
// dotnet add package IronOcr.Languages.French
// dotnet add package IronOcr.Languages.German
// dotnet add package IronOcr.Languages.Spanish
using IronOcr;

public static string ExtractFromEuropeanDocument(string imagePath)
{
    var ocr = new IronTesseract();
    ocr.Language = OcrLanguage.English;
    ocr.AddSecondaryLanguage(OcrLanguage.French);
    ocr.AddSecondaryLanguage(OcrLanguage.German);
    ocr.AddSecondaryLanguage(OcrLanguage.Spanish);

    return ocr.Read(imagePath).Text;
}

Imports IronOcr

Public Shared Function ExtractFromEuropeanDocument(imagePath As String) As String
    Dim ocr As New IronTesseract()
    ocr.Language = OcrLanguage.English
    ocr.AddSecondaryLanguage(OcrLanguage.French)
    ocr.AddSecondaryLanguage(OcrLanguage.German)
    ocr.AddSecondaryLanguage(OcrLanguage.Spanish)

    Return ocr.Read(imagePath).Text
End Function

The tessdata folder, the file-existence loop, the path concatenation string, and the deployment manifest updates are all replaced by PackageReference lines in the .csproj. Adding a language to Docker means one additional dotnet add package — not a Dockerfile COPY step. The multiple languages guide covers the full 125+ language catalog and CJK character sets, and the languages index lists every available language pack.

Issue 1: Tessdata Path Reference Remains After Migration

Tesseract: TessDataPath constants, Directory.Exists(TessDataPath) guards, and File.Exists(Path.Combine(TessDataPath, lang + ".traineddata")) checks appear throughout the codebase and in project files as <Content Include="tessdata\**"> build items.

Solution: Search for all occurrences and remove them along with the tessdata folder itself:

# Find all tessdata references in source
grep -r "TessDataPath\|tessdata\|traineddata" --include="*.cs" .
grep -r "tessdata" --include="*.csproj" .

# Find all tessdata references in source
grep -r "TessDataPath\|tessdata\|traineddata" --include="*.cs" .
grep -r "tessdata" --include="*.csproj" .

After removing the path constants and file guards, delete the tessdata folder from the project. Remove any <Content Include="tessdata\**" CopyToOutputDirectory="..." /> lines from .csproj files. Dockerfile COPY ./tessdata lines and ENV TESSDATA_PREFIX environment variable declarations are also safe to remove.

Issue 2: Pix Object Type Cannot Be Resolved

Tesseract: Pix is a Leptonica image wrapper type from the Tesseract namespace. References appear in variable declarations (using var img = Pix.LoadFromFile(...)), method signatures that accept Pix parameters, and any code that calls Pix.LoadFromMemory() or Pix.LoadFromBitmap().

Solution: Replace Pix.LoadFromFile(path) with input.LoadImage(path) on an OcrInput instance. Replace Pix.LoadFromMemory(bytes) with input.LoadImage(bytes). The OcrInput class accepts file paths, byte arrays, streams, and System.Drawing.Bitmap objects directly. No conversion to an intermediate wrapper type is required. See the image input guide and stream input guide for the full set of accepted input types.

Issue 3: ResultIterator Loop Pattern Has No Direct Equivalent

Tesseract: Code that iterates ResultIterator with iter.Begin(), iter.Next(PageIteratorLevel.Word), and iter.TryGetBoundingBox() is the standard pattern for word-level or character-level extraction. This pattern requires tracking iterator state and level transitions manually.

Solution: Replace the iterator loop with LINQ over result.Words, result.Pages, or the appropriate collection level:

// Before: iterator loop
using var iter = page.GetIterator();
iter.Begin();
do
{
    if (iter.TryGetBoundingBox(PageIteratorLevel.Word, out var bounds))
    {
        string text = iter.GetText(PageIteratorLevel.Word);
        // process text and bounds
    }
}
while (iter.Next(PageIteratorLevel.Word));

// After: enumerable collection
var result = new IronTesseract().Read(imagePath);
foreach (var word in result.Words)
{
    // word.Text, word.X, word.Y, word.Width, word.Height, word.Confidence
}

// Before: iterator loop
using var iter = page.GetIterator();
iter.Begin();
do
{
    if (iter.TryGetBoundingBox(PageIteratorLevel.Word, out var bounds))
    {
        string text = iter.GetText(PageIteratorLevel.Word);
        // process text and bounds
    }
}
while (iter.Next(PageIteratorLevel.Word));

// After: enumerable collection
var result = new IronTesseract().Read(imagePath);
foreach (var word in result.Words)
{
    // word.Text, word.X, word.Y, word.Width, word.Height, word.Confidence
}

Imports IronOcr

' Before: iterator loop
Using iter = page.GetIterator()
    iter.Begin()
    Do
        Dim bounds As Rectangle
        If iter.TryGetBoundingBox(PageIteratorLevel.Word, bounds) Then
            Dim text As String = iter.GetText(PageIteratorLevel.Word)
            ' process text and bounds
        End If
    Loop While iter.Next(PageIteratorLevel.Word)
End Using

' After: enumerable collection
Dim result = New IronTesseract().Read(imagePath)
For Each word In result.Words
    ' word.Text, word.X, word.Y, word.Width, word.Height, word.Confidence
Next

For paragraph-level access — which has no clean analog in the Tesseract iterator — use result.Pages[i].Paragraphs. The read results guide documents all available levels.

Issue 4: PDF Library Code Must Be Removed Entirely

Tesseract: Any code that converts PDF pages to images before passing them to Tesseract — PdfiumViewer document.Render() loops, PDFtoImage Conversion.ToImage() calls, Docnet.Core GetPageReader() patterns, or GhostScript process invocations — exists only to work around Tesseract's inability to open PDFs. These classes, loops, temp file patterns, and native binary deployments are all scaffolding around the real requirement.

Solution: Delete the PDF rendering code entirely. Replace the entire render-then-OCR block with input.LoadPdf(path):

// Before: ~50-150 lines of PdfiumViewer + Tesseract + temp file management
// After:
using var input = new OcrInput();
input.LoadPdf("document.pdf");
input.Deskew();
input.DeNoise();
var result = new IronTesseract().Read(input);

// Before: ~50-150 lines of PdfiumViewer + Tesseract + temp file management
// After:
using var input = new OcrInput();
input.LoadPdf("document.pdf");
input.Deskew();
input.DeNoise();
var result = new IronTesseract().Read(input);

Imports IronOcr

Dim input As New OcrInput()
input.LoadPdf("document.pdf")
input.Deskew()
input.DeNoise()

Dim result = (New IronTesseract()).Read(input)

Remove PdfiumViewer, PDFtoImage, and Docnet.Core package references from the .csproj. Remove native binary deployments (pdfium.dll, GhostScript executables) from build scripts and Dockerfiles. The PDF input guide covers page range selection and password-protected PDFs.

Issue 5: Parallel Processing Engine-Per-Thread Pattern

Tesseract: The standard pattern for safe parallel OCR creates a new TesseractEngine inside the Parallel.ForEach body because a single engine is not thread-safe. This loads the full language model per thread.

Solution: Create IronTesseract once before the loop and reference it inside:

// Before: engine per thread, 40-100 MB per language model, times thread count
Parallel.ForEach(files, file =>
{
    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var img    = Pix.LoadFromFile(file);
    using var page   = engine.Process(img);
    results[file] = page.GetText();
});

// After: single engine, thread-safe, shared pool
var ocr = new IronTesseract();
Parallel.ForEach(files, file =>
{
    var result = ocr.Read(file);
    results[file] = result.Text;
});

// Before: engine per thread, 40-100 MB per language model, times thread count
Parallel.ForEach(files, file =>
{
    using var engine = new TesseractEngine(TessDataPath, "eng", EngineMode.Default);
    using var img    = Pix.LoadFromFile(file);
    using var page   = engine.Process(img);
    results[file] = page.GetText();
});

// After: single engine, thread-safe, shared pool
var ocr = new IronTesseract();
Parallel.ForEach(files, file =>
{
    var result = ocr.Read(file);
    results[file] = result.Text;
});

Imports System.Threading.Tasks
Imports Tesseract
Imports IronOcr

' Before: engine per thread, 40-100 MB per language model, times thread count
Parallel.ForEach(files, Sub(file)
    Using engine As New TesseractEngine(TessDataPath, "eng", EngineMode.Default)
        Using img As Pix = Pix.LoadFromFile(file)
            Using page As Page = engine.Process(img)
                results(file) = page.GetText()
            End Using
        End Using
    End Using
End Sub)

' After: single engine, thread-safe, shared pool
Dim ocr As New IronTesseract()
Parallel.ForEach(files, Sub(file)
    Dim result = ocr.Read(file)
    results(file) = result.Text
End Sub)

The thread-safety change also eliminates the using disposal pattern from inside the loop body, which was necessary only to ensure each per-thread engine was released promptly.

Issue 6: EngineMode Enum Has No Direct Mapping

Tesseract: EngineMode.Default, EngineMode.TesseractOnly, and EngineMode.LstmOnly appear in TesseractEngine constructors to select whether Tesseract uses the legacy engine, LSTM, or both. The charlesw wrapper exposes these modes because Tesseract 4.x retained both engines.

Solution: IronOCR uses the Tesseract 5 LSTM engine exclusively, which is the high-accuracy configuration. No EngineMode parameter exists because there is no legacy engine to fall back to. Remove the EngineMode argument when translating the constructor call. For throughput-versus-accuracy tuning, use ocr.Configuration.PageSegmentationMode and consult the speed optimization guide.

Pre-Migration

Audit the codebase for all Tesseract and tessdata references:

# Find all using directives for the Tesseract namespace
grep -rn "using Tesseract" --include="*.cs" .

# Find TesseractEngine constructors
grep -rn "TesseractEngine\|TessDataPath\|tessdata" --include="*.cs" .

# Find Pix object usage
grep -rn "Pix\." --include="*.cs" .

# Find ResultIterator usage
grep -rn "GetIterator\|ResultIterator\|PageIteratorLevel" --include="*.cs" .

# Find PDF rendering libraries added for Tesseract
grep -rn "PdfiumViewer\|PDFtoImage\|Docnet\|GhostScript" --include="*.cs" .

# Find tessdata references in project files
grep -rn "tessdata\|traineddata" --include="*.csproj" .

# Find tessdata references in Dockerfiles
grep -rn "tessdata\|TESSDATA_PREFIX\|libtesseract" Dockerfile* .

# Find all using directives for the Tesseract namespace
grep -rn "using Tesseract" --include="*.cs" .

# Find TesseractEngine constructors
grep -rn "TesseractEngine\|TessDataPath\|tessdata" --include="*.cs" .

# Find Pix object usage
grep -rn "Pix\." --include="*.cs" .

# Find ResultIterator usage
grep -rn "GetIterator\|ResultIterator\|PageIteratorLevel" --include="*.cs" .

# Find PDF rendering libraries added for Tesseract
grep -rn "PdfiumViewer\|PDFtoImage\|Docnet\|GhostScript" --include="*.cs" .

# Find tessdata references in project files
grep -rn "tessdata\|traineddata" --include="*.csproj" .

# Find tessdata references in Dockerfiles
grep -rn "tessdata\|TESSDATA_PREFIX\|libtesseract" Dockerfile* .

Inventory results to estimate migration scope:

• Count files with using Tesseract to determine how many classes require changes
• Identify which PDF rendering library is in use (PdfiumViewer, PDFtoImage, Docnet.Core, GhostScript)
• Note which languages are referenced in TesseractEngine constructor strings to determine which IronOCR language NuGet packages to add

Code Migration

1. Remove Tesseract NuGet package reference from all .csproj files
2. Remove PDF rendering library NuGet references added solely for Tesseract support (PdfiumViewer, PDFtoImage, Docnet.Core)
3. Install IronOcr NuGet package
4. Install required language NuGet packages (IronOcr.Languages.French, etc.)
5. Add IronOcr.License.LicenseKey = "YOUR-KEY"; at application startup
6. Replace using Tesseract; with using IronOcr; in all affected files
7. Remove TessDataPath constants and all Directory.Exists / File.Exists tessdata guards
8. Replace new TesseractEngine(...) with new IronTesseract()
9. Replace Pix.LoadFromFile(path) with input.LoadImage(path) on an OcrInput instance
10. Replace Pix.LoadFromMemory(bytes) with input.LoadImage(bytes)
11. Replace engine.Process(img) with ocr.Read(input)
12. Replace page.GetText() with result.Text
13. Replace page.GetMeanConfidence() with result.Confidence
14. Replace ResultIterator loops with enumeration over result.Words or result.Pages[i].Paragraphs
15. Replace PDF rendering loops with input.LoadPdf(path) — delete the rendering library code entirely
16. Replace "eng+fra+deu" language strings with ocr.AddSecondaryLanguage(OcrLanguage.X) calls
17. Delete the tessdata folder and its build <Content Include="..."> project items
18. Remove native binary deployment steps from build scripts and Dockerfiles (tessdata COPY, TESSDATA_PREFIX ENV, apt-get libtesseract-dev)

Post-Migration

• Verify basic text extraction on the same sample images used during development with the Tesseract wrapper
• Confirm confidence scores are reasonable (70%+ for clean documents, 85%+ for high-quality scans)
• Test multi-page TIFF input produces the correct number of pages in result.Pages
• Verify PDF input reads scanned PDFs without requiring PdfiumViewer or any external library
• Test password-protected PDF reading with input.LoadPdf(path, Password: "...") against a known encrypted file
• Confirm searchable PDF output opens in Adobe Reader and supports text search
• Test parallel processing: create one IronTesseract instance before a Parallel.ForEach loop and confirm no thread-safety exceptions
• Verify each language pack produces correct output for the target language document set
• Run Docker build without COPY tessdata and apt-get libtesseract-dev — confirm the container starts and processes documents
• Confirm the tessdata folder and native DLL files are absent from the published output directory
• Check that no TesseractException or System.DllNotFoundException appears in logs after removing native binary references